Elep

Network Working Group                                   M. Smith, Editor
INTERNET-DRAFT                             Netscape Communications Corp.
Intended Category: Standards Track                              T. Howes
Obsoletes: RFC 1823                                      Loudcloud, Inc.
Expires: May 2001                                              A. Herron
                                                         Microsoft Corp.
                                                                 M. Wahl
                                                  Sun Microsystems, Inc.
                                                              A. Anantha
                                                         Microsoft Corp.

17 November 2000

The C LDAP Application Program Interface
                 <draft-ietf-ldapext-ldap-c-api-05.txt>

1.  Status of this Memo

This document is an Internet-Draft and is in full conformance with all
provisions of Section 10 of RFC2026.  Internet-Drafts are working docu-
ments of the Internet Engineering Task Force (IETF), its areas, and its
working groups.  Note that other groups may also distribute working
documents as Internet-Drafts.

Internet-Drafts are draft documents valid for a maximum of six months
and may be updated, replaced, or obsoleted by other documents at any
time.  It is inappropriate to use Internet-Drafts as reference material
or to cite them other than as "work in progress."

The list of current Internet-Drafts can be accessed at
http://www.ietf.org/ietf/1id-abstracts.txt.

The list of Internet-Draft Shadow Directories can be accessed at
http://www.ietf.org/shadow.html.

This draft document will be submitted to the RFC Editor as a Standards
Track document. Distribution of this memo is unlimited.  Technical dis-
cussion of this document will take place on the IETF LDAP Extension
Working Group mailing list <ietf-ldapext@netscape.com>.  Please send
editorial comments directly to the authors.

Please see the Copyright section near the end of this document for more
information.

Expires: May 2001                                               [Page 1]

C LDAP API        C LDAP Application Program Interface  17 November 2000

2.  Introduction

This document defines a C language application program interface (API)
to the Lightweight Directory Access Protocol (LDAP). This document
replaces the previous definition of this API, defined in RFC 1823,
updating it to include support for features found in version 3 of the
LDAP protocol.  New extended operation functions were added to support
LDAPv3 features such as controls.  In addition, other LDAP API changes
were made to support information hiding and thread safety.

The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
"SHOULD", "SHOULD NOT", "RECOMMENDED",  "MAY", and "OPTIONAL" in this
document are to be interpreted as described in RFC 2119[1].

The C LDAP API is designed to be powerful, yet simple to use. It defines
compatible synchronous and asynchronous interfaces to LDAP to suit a
wide variety of applications. This document gives a brief overview of
the LDAP model, then an overview of how the API is used by an applica-
tion program to obtain LDAP information.  The API calls are described in
detail, followed by appendices that provide example code demonstrating
use of the API, the namespace consumed by the API, a summary of require-
ments for API extensions, known incompatibilities with RFC 1823, and a
list of changes made since the last revision of this document.

3.  Table of Contents

1.     Status of this Memo............................................1
2.     Introduction...................................................2
3.     Table of Contents..............................................2
4.     Overview of the LDAP Model.....................................4
5.     Overview of LDAP API Use and General Requirements..............4
6.     Header Requirements............................................6
7.     Common Data Structures and Types...............................7
8.     Memory Handling Overview.......................................9
9.     Retrieving Information About the API Implementation............9
9.1.      Retrieving Information at Compile Time......................10
9.2.      Retrieving Information During Execution.....................11
10.    Result Codes...................................................14
11.    Performing LDAP Operations.....................................16
11.1.     Initializing an LDAP Session................................16
11.2.     LDAP Session Handle Options.................................17
11.3.     Working With Controls.......................................23
11.3.1.      A Client Control That Governs Referral Processing........24
11.4.     Authenticating to the directory.............................25
11.5.     Closing the session.........................................27
11.6.     Searching...................................................28
11.7.     Reading an Entry............................................32

Expires: May 2001                                               [Page 2]

C LDAP API        C LDAP Application Program Interface  17 November 2000

11.8.     Listing the Children of an Entry............................32
11.9.     Comparing a Value Against an Entry..........................33
11.10.    Modifying an entry..........................................35
11.11.    Modifying the Name of an Entry..............................37
11.12.    Adding an entry.............................................39
11.13.    Deleting an entry...........................................41
11.14.    Extended Operations.........................................43
12.    Abandoning An Operation........................................44
13.    Obtaining Results and Peeking Inside LDAP Messages.............45
14.    Handling Errors and Parsing Results............................47
15.    Stepping Through a List of Results.............................51
16.    Parsing Search Results.........................................51
16.1.     Stepping Through a List of Entries or References............52
16.2.     Stepping Through the Attributes of an Entry.................53
16.3.     Retrieving the Values of an Attribute.......................54
16.4.     Retrieving the name of an entry.............................55
16.5.     Retrieving controls from an entry...........................56
16.6.     Parsing References..........................................57
17.    Encoded ASN.1 Value Manipulation...............................58
17.1.     BER Data Structures and Types...............................58
17.2.     Memory Disposal and Utility Functions.......................60
17.3.     Encoding....................................................60
17.4.     Encoding Example............................................63
17.5.     Decoding....................................................64
17.6.     Decoding Example............................................67
18.    Security Considerations........................................70
19.    Acknowledgements...............................................70
20.    Copyright......................................................70
21.    Bibliography...................................................71
22.    Authors' Addresses.............................................72
23.    Appendix A - Sample C LDAP API Code............................73
24.    Appendix B - Namespace Consumed By This Specification..........74
25.    Appendix C - Summary of Requirements for API Extensions........75
25.1.     Compatibility...............................................75
25.2.     Style.......................................................75
25.3.     Dependence on Externally Defined Types......................75
25.4.     Compile Time Information....................................76
25.5.     Runtime Information.........................................76
25.6.     Values Used for Session Handle Options......................76
26.    Appendix D - Known Incompatibilities with RFC 1823.............76
26.1.     Opaque LDAP Structure.......................................76
26.2.     Additional Result Codes.....................................77
26.3.     Freeing of String Data with ldap_memfree()..................77
26.4.     Changes to ldap_result()....................................77
26.5.     Changes to ldap_first_attribute() and ldap_next_attribute...77
26.6.     Changes to ldap_modrdn() and ldap_modrdn_s() Functions......78
26.7.     Changes to the berval structure.............................78
26.8.     API Specification Clarified.................................78

Expires: May 2001                                               [Page 3]

C LDAP API        C LDAP Application Program Interface  17 November 2000

26.9.     Deprecated Functions........................................78
27.    Appendix E - Data Types and Legacy Implementations.............79
28.    Appendix F - Changes Made Since Last Document Revision.........80
28.1.     API Changes.................................................80
28.2.     Editorial Changes and Clarifications........................81

4.  Overview of the LDAP Model

LDAP is the lightweight directory access protocol, described in [2] and
[3]. It can provide a lightweight frontend to the X.500 directory [4],
or a stand-alone service. In either mode, LDAP is based on a client-
server model in which a client makes a TCP connection to an LDAP server,
over which it sends requests and receives responses.

The LDAP information model is based on the entry, which contains infor-
mation about some object (e.g., a person).  Entries are composed of
attributes, which have a type and one or more values. Each attribute has
a syntax that determines what kinds of values are allowed in the attri-
bute (e.g., ASCII characters, a jpeg photograph, etc.) and how those
values behave during directory operations (e.g., is case significant
during comparisons).

Entries may be organized in a tree structure, usually based on politi-
cal, geographical, and organizational boundaries. Each entry is uniquely
named relative to its sibling entries by its relative distinguished name
(RDN) consisting of one or more distinguished attribute values from the
entry.  At most one value from each attribute may be used in the RDN.
For example, the entry for the person Babs Jensen might be named with
the "Barbara Jensen" value from the commonName attribute.

A globally unique name for an entry, called a distinguished name or DN,
is constructed by concatenating the sequence of RDNs from the entry up
to the root of the tree. For example, if Babs worked for the University
of Michigan, the DN of her U-M entry might be "cn=Barbara Jensen,
o=University of Michigan, c=US". The DN format used by LDAP is defined
in [5].

Operations are provided to authenticate, search for and retrieve infor-
mation, modify information, and add and delete entries from the tree.
The next sections give an overview of how the API is used and detailed
descriptions of the LDAP API calls that implement all of these func-
tions.

5.  Overview of LDAP API Use and General Requirements

An application generally uses the C LDAP API in four simple steps.

Expires: May 2001                                               [Page 4]

C LDAP API        C LDAP Application Program Interface  17 November 2000

1.   Initialize an LDAP session with a primary LDAP server. The
        ldap_init() function returns a handle to the session, allowing
        multiple connections to be open at once.

2.   Authenticate to the LDAP server. The ldap_sasl_bind() function
        and friends support a variety of authentication methods.

3.   Perform some LDAP operations and obtain some results.
        ldap_search() and friends return results which can be parsed by
        ldap_parse_result(), ldap_first_entry(), ldap_next_entry(), etc.

4.   Close the session. The ldap_unbind() function closes the connec-
        tion.

Operations can be performed either synchronously or asynchronously.  The
names of the synchronous functions end in _s. For example, a synchronous
search can be completed by calling ldap_search_s(). An asynchronous
search can be initiated by calling ldap_search(). All synchronous rou-
tines return an indication of the outcome of the operation (e.g, the
constant LDAP_SUCCESS or some other result code).  The asynchronous rou-
tines make available to the caller the message id of the operation ini-
tiated. This id can be used in subsequent calls to ldap_result() to
obtain the result(s) of the operation. An asynchronous operation can be
abandoned by calling ldap_abandon() or ldap_abandon_ext().  Note that
there is no requirement that an LDAP API implementation not block when
handling asynchronous API functions; the term "asynchronous" as used in
this document refers to the fact that the sending of LDAP requests can
be separated from the receiving of LDAP responses.

Results and errors are returned in an opaque structure called LDAPMes-
sage.  Routines are provided to parse this structure, step through
entries and attributes returned, etc. Routines are also provided to
interpret errors. Later sections of this document describe these rou-
tines in more detail.

LDAP version 3 servers can return referrals and references to other
servers.  By default, implementations of this API will attempt to follow
referrals automatically for the application.  This behavior can be dis-
abled globally (using the ldap_set_option() call) or on a per-request
basis through the use of a client control.

All DN and string attribute values passed into or produced by this C
LDAP API are represented using the character set of the underlying LDAP
protocol version in use.  When this API is used with LDAPv3, DN and
string values are represented as UTF-8[6] characters.  When this API is
used with LDAPv2, the US-ASCII[7] or T.61[7] character set are used.
Future documents MAY specify additional APIs supporting other character
sets.

Expires: May 2001                                               [Page 5]

C LDAP API        C LDAP Application Program Interface  17 November 2000

For compatibility with existing applications, implementations of this
API will by default use version 2 of the LDAP protocol.  Applications
that intend to take advantage of LDAP version 3 features will need to
use the ldap_set_option() call with a LDAP_OPT_PROTOCOL_VERSION to
switch to version 3.

Unless otherwise indicated, conformant implementations of this specifi-
cation MUST implement all of the C LDAP API functions as described in
this document, and they MUST use the function prototypes, macro defini-
tions, and types defined in this document.

Note that this API is designed for use in environments where the 'int'
type is at least 32 bits in size.

6.  Header Requirements

To promote portability of applications, the following requirements are
imposed on the headers used by applications to access the services of
this API:

Name and Inclusion
        Applications only need to include a single header named ldap.h
        to access all of the API services described in this document.
        Therefore, the following C source program MUST compile and exe-
        cute without errors:

#include <ldap.h>

int
           main()
           {
               return 0;
           }

The ldap.h header MAY include other implementation-specific
        headers.

Implementations SHOULD also provide a header named lber.h to facilitate
development of applications desiring compatibility with older LDAP
implementations.  The lber.h header MAY be empty.  Old applications that
include lber.h in order to use BER facilities will need to include
ldap.h.

Idempotence
        All headers SHOULD be idempotent; that is, if they are included
        more than once the effect is as if they had only been included

Expires: May 2001                                               [Page 6]

C LDAP API        C LDAP Application Program Interface  17 November 2000

once.

Must Be Included Before API Is Used
        An application MUST include the ldap.h header before referencing
        any of the function or type definitions described in this API
        specification.

Mutual Independence
        Headers SHOULD be mutually independent with minimal dependence
        on system or any other headers.

Use of the 'const' Keyword
        This API specification is defined in terms of ISO C[8].  It
        makes use of function prototypes and the 'const' keyword.  The
        use of 'const' in this specification is limited to simple, non-
        array function parameters to avoid forcing applications to
        declare parameters and variables that accept return values from
        LDAP API functions as 'const.'  Implementations specifically
        designed to be used with non-ISO C translators SHOULD provide
        function declarations without prototypes or function prototypes
        without specification of 'const' arguments.

Definition of 'struct timeval'
        This API specification uses the 'struct timeval' type.  Imple-
        mentations of this API MUST ensure that the struct timeval type
        is by default defined as a consequence of including the ldap.h
        header.  Because struct timeval is usually defined in one or
        more system headers, it is possible for header conflicts to
        occur if ldap.h also defines it or arranges for it to be defined
        by including another header.  Therefore, applications MAY want
        to arrange for struct timeval to be defined before they include
        ldap.h.  To support this, the ldap.h header MUST NOT itself
        define struct timeval if the preprocessor symbol
        LDAP_TYPE_TIMEVAL_DEFINED is defined before ldap.h is included.

7.  Common Data Structures and Types

Data structures and types that are common to several LDAP API functions
are defined here:

typedef struct ldap LDAP;

typedef struct ldapmsg LDAPMessage;

typedef struct berelement BerElement;

typedef <impl_len_t> ber_len_t;

Expires: May 2001                                               [Page 7]

C LDAP API        C LDAP Application Program Interface  17 November 2000

typedef struct berval {
           ber_len_t       bv_len;
           char            *bv_val;
       } BerValue;

struct timeval {
           <impl_sec_t>    tv_sec;
           <impl_usec_t>   tv_usec;
       };

The LDAP structure is an opaque data type that represents an LDAP ses-
sion Typically this corresponds to a connection to a single server, but
it MAY encompass several server connections in the face of LDAPv3 refer-
rals.

The LDAPMessage structure is an opaque data type that is used to return
entry, reference, result, and error information.  An LDAPMessage struc-
ture can represent the beginning of a list, or chain of messages that
consists of a series of entries, references, and result messages as
returned by LDAP operations such as search.  LDAP API functions such as
ldap_parse_result() that operate on message chains that can contain more
than one result message always operate on the first result message in
the chain.  See the "Obtaining Results and Peeking Inside LDAP Messages"
section of this document for more information.

The BerElement structure is an opaque data type that is used to hold
data and state information about encoded data.  It is described in more
detail in the section "Encoded ASN.1 Value Manipulation" later in this
document.

The `ber_len_t' type is an unsigned integral data type that is large
enough to contain the length of the largest piece of data supported by
the API implementation.  The `<impl_len_t>' in the `ber_len_t' typedef
MUST be replaced with an appropriate type.  The width (number of signi-
ficant bits) of `ber_len_t' MUST be at least 32 and no larger than that
of `unsigned long'.  See the appendix "Data Types and Legacy Implementa-
tions" for additional considerations.

The BerValue structure is used to represent arbitrary binary data and
its fields have the following meanings:

bv_len   Length of data in bytes.

bv_val   A pointer to the data itself.

The timeval structure is used to represent an interval of time and its
fields have the following meanings:

Expires: May 2001                                               [Page 8]

C LDAP API        C LDAP Application Program Interface  17 November 2000

tv_sec   Seconds component of time interval.

tv_usec  Microseconds component of time interval.

Note that because the struct timeval definition typically is derived
from a system header, the types used for the tv_sec and tv_usec com-
ponents are implementation-specific integral types.  Therefore,
`<impl_sec_t>' and `<impl_usec_t>' in the struct timeval definition MUST
be replaced with appropriate types.  See the earlier section "Header
Requirements" for more information on struct timeval.

8.  Memory Handling Overview

All memory that is allocated by a function in this C LDAP API and
returned to the caller SHOULD be disposed of by calling the appropriate
"free" function provided by this API.  The correct "free" function to
call is documented in each section of this document where a function
that allocates memory is described.

Memory that is allocated through means outside of the C LDAP API MUST
NOT be disposed of using a function provided by this API.

If a pointer value passed to one of the C LDAP API "free" functions is
NULL, graceful failure (i.e, ignoring of the NULL pointer) MUST occur.

The complete list of "free" functions that are used to dispose of allo-
cated memory is:

ber_bvecfree()
   ber_bvfree()
   ber_free()
   ldap_control_free()
   ldap_controls_free()
   ldap_memfree()
   ldap_msgfree()
   ldap_value_free()
   ldap_value_free_len()

9.  Retrieving Information About the API Implementation

Applications developed to this specification need to be able to deter-
mine information about the particular API implementation they are using
both at compile time and during execution.

Expires: May 2001                                               [Page 9]

C LDAP API        C LDAP Application Program Interface  17 November 2000

9.1.  Retrieving Information at Compile Time

All conformant implementations MUST include the following five defini-
tions in a header so compile time tests can be done by LDAP software
developers:

#define LDAP_API_VERSION     level
   #define LDAP_VERSION_MIN     min-version
   #define LDAP_VERSION_MAX     max-version
   #define LDAP_VENDOR_NAME     "vend-name"
   #define LDAP_VENDOR_VERSION  vend-version

where:

"level" is replaced with the RFC number given to this C LDAP API
     specification when it is published as a standards track RFC.

min-version is replaced with the lowest LDAP protocol version sup-
     ported by the implementation.

max-version is replaced with the highest LDAP protocol version sup-
     ported by the implementation.  This SHOULD be 3.

"vend-name" is replaced with a text string that identifies the
     party that supplies the API implementation.

"vend-version" is a supplier-specific version number multiplied
     times 100.

Note that the LDAP_VENDOR_NAME macro SHOULD be defined as "" if no ven-
dor name is available and the LDAP_VENDOR_VERSION macro SHOULD be
defined as 0 if no vendor-specific version information is available.

For example, if this specification is published as RFC 88888, Netscape
Communication's version 4.0 implementation that supports LDAPv2 and v3
might include macro definitions like these:

#define LDAP_API_VERSION     88888      /* RFC 88888 compliant */
   #define LDAP_VERSION_MIN     2
   #define LDAP_VERSION_MAX     3
   #define LDAP_VENDOR_NAME     "Netscape Communications Corp."
   #define LDAP_VENDOR_VERSION  400        /* version 4.0 */

and application code can test the C LDAP API version level using a
construct such as this one:

#if (LDAP_API_VERSION >= 88888)
           /* use features supported in RFC 88888 or later */

Expires: May 2001                                              [Page 10]

C LDAP API        C LDAP Application Program Interface  17 November 2000

#endif

Until such time as this document is published as an RFC, implementations
SHOULD use the value 2000 plus the revision number of this draft for
LDAP_API_VERSION.  For example, the correct value for LDAP_API_VERSION
for revision 05 of this draft is 2005.

Documents that extend this specification SHOULD define a macro of the
form:

#define LDAP_API_FEATURE_x level

where "x" is replaced with a name (textual identifier) for the feature
and "level" is replaced with the number of the RFC that specifies the
API extension.  The name SHOULD NOT begin with the string "X_".

For example, if C LDAP API extensions for Transport Layer Security [9]
were published in RFC 99999, that RFC might require conformant implemen-
tations to define a macro like this:

#define LDAP_API_FEATURE_TLS 99999

Private or experimental API extensions SHOULD be indicated by defining a
macro of this same form where "x" (the extension's name) begins with the
string "X_" and "level" is replaced with a integer number that is
specific to the extension.

It is RECOMMENDED that private or experimental API extensions use only
the following prefixes for macros, types, and function names:
       LDAP_X_
       LBER_X_
       ldap_x_
       ber_x_
and that these prefixes not be used by standard extensions.

9.2.  Retrieving Information During Execution

The ldap_get_option() call (described in greater detail later in this
document) can be used during execution in conjunction with an option
parameter value of LDAP_OPT_API_INFO (0x00) to retrieve some basic
information about the API and about the specific implementation being
used.  The ld parameter to ldap_get_option() can be either NULL or a
valid LDAP session handle which was obtained by calling ldap_init().
The optdata parameter to ldap_get_option() SHOULD be the address of an
LDAPAPIInfo structure which is defined as follows:

Expires: May 2001                                              [Page 11]

C LDAP API        C LDAP Application Program Interface  17 November 2000

typedef struct ldapapiinfo {
       int  ldapai_info_version;     /* version of this struct (1) */
       int  ldapai_api_version;      /* revision of API supported */
       int  ldapai_protocol_version; /* highest LDAP version supported */
       char **ldapai_extensions;     /* names of API extensions */
       char *ldapai_vendor_name;     /* name of supplier */
       int  ldapai_vendor_version;   /* supplier-specific version times 100 */
   } LDAPAPIInfo;

In addition, API implementations MUST include the following macro defin-
ition:

#define LDAP_API_INFO_VERSION    1

Note that the ldapai_info_version field of the LDAPAPIInfo structure
SHOULD be set to the value LDAP_API_INFO_VERSION (1) before calling
ldap_get_option() so that it can be checked for consistency.  All other
fields are set by the ldap_get_option() function.

The members of the LDAPAPIInfo structure are:

ldapai_info_version
          A number that identifies the version of the LDAPAPIInfo struc-
          ture.  This SHOULD be set to the value LDAP_API_INFO_VERSION
          (1) before calling ldap_get_option().  If the value received
          is not recognized by the API implementation, the
          ldap_get_option() function sets ldapai_info_version to a valid
          value that would be recognized, sets the ldapai_api_version to
          the correct value, and returns an error without filling in any
          of the other fields in the LDAPAPIInfo structure.

ldapai_api_version
          A number that matches that assigned to the C LDAP API RFC sup-
          ported by the API implementation.  This SHOULD match the value
          of the LDAP_API_VERSION macro defined earlier.

ldapai_protocol_version
          The highest LDAP protocol version supported by the implementa-
          tion.  For example, if LDAPv3 is the highest version supported
          then this field will be set to 3.

ldapai_vendor_name
          A zero-terminated string that contains the name of the party
          that produced the LDAP API implementation.  This field may be
          set to NULL if no name is available.  If non-NULL, the caller
          is responsible for disposing of the memory occupied by passing
          this pointer to ldap_memfree() which is described later in
          this document.  This value SHOULD match the value of the

Expires: May 2001                                              [Page 12]

C LDAP API        C LDAP Application Program Interface  17 November 2000

LDAP_VENDOR_NAME macro described earlier in this document.

ldapai_vendor_version
          An implementation-specific version number multiplied by 100.
          For example, if the implementation version is 4.0 then this
          field will be set to 400.  If no version information is avail-
          able, this field will be set to 0.  This value SHOULD match
          the value of the LDAP_VENDOR_VERSION macro described earlier
          in this document.

ldapai_extensions
          A NULL-terminated array of character strings that lists the
          names of the API extensions supported by the LDAP API imple-
          mentation.  These names will typically match the textual iden-
          tifiers that appear in the "x" portion of the
          LDAP_API_FEATURE_x macros described above, although the pre-
          cise value MUST be defined by documents that specify C LDAP
          API extensions.  If no API extensions are supported, this
          field will be set to NULL.  The caller is responsible for
          disposing of the memory occupied by this array by passing it
          to ldap_value_free() which is described later in this docu-
          ment.  To retrieve more information about a particular exten-
          sion, the ldap_get_option() call can be used with an option
          parameter value of LDAP_OPT_API_FEATURE_INFO (0x15).  The opt-
          data parameter to the ldap_get_option() SHOULD be the address
          of an LDAPAPIFeatureInfo structure which is defined as fol-
          lows:

typedef struct ldap_apifeature_info {
                 int   ldapaif_info_version; /* version of this struct (1) */
                 char  *ldapaif_name;        /* name of supported feature */
                 int   ldapaif_version;      /* revision of supported feature */
             } LDAPAPIFeatureInfo;

In addition, API implementations MUST include the following
          macro definition:

#define LDAP_FEATURE_INFO_VERSION    1

Note that the ldapaif_info_version field of the LDAPAPI-
          FeatureInfo structure SHOULD be set to the value
          LDAP_FEATURE_INFO_VERSION (1) and the ldapaif_name field
          SHOULD be set to the extension name string as described below
          before ldap_get_option() is called.  The call will fill in the
          ldapaif_version field of the LDAPAPIFeatureInfo structure.

The members of the LDAPAPIFeatureInfo structure are:

Expires: May 2001                                              [Page 13]

C LDAP API        C LDAP Application Program Interface  17 November 2000

ldapaif_info_version
             A number that identifies the version of the LDAPAPI-
             FeatureInfo structure.  This SHOULD be set to the value
             LDAP_FEATURE_INFO_VERSION (1) before calling
             ldap_get_option().  If the value received is not recognized
             by the API implementation, the ldap_get_option() function
             sets ldapaif_info_version to a valid value that would be
             recognized and returns an error without filling in the
             ldapaif_version field in the LDAPAPIFeatureInfo structure.

ldapaif_name
             The name of an extension, as returned in the
             ldapai_extensions array of the LDAPAPIInfo structure and as
             specified in the document that describes the extension.

ldapaif_version
             This field will be set as a result of calling
             ldap_get_option().  It is a number that matches that
             assigned to the C LDAP API extension RFC supported for this
             extension.  For private or experimental API extensions, the
             value is extension-specific.  In either case, the value of
             ldapaxi_ext_version SHOULD be identical to the value of the
             LDAP_API_FEATURE_x macro defined for the extension
             (described above).

10.  Result Codes

Many of the LDAP API routines return result codes, some of which indi-
cate local API errors and some of which are LDAP resultCodes that are
returned by servers.  All of the result codes are non-negative integers.
Supported result codes are as follows (hexadecimal values are given in
parentheses after the constant):

LDAP_SUCCESS (0x00)
           LDAP_OPERATIONS_ERROR (0x01)
           LDAP_PROTOCOL_ERROR (0x02)
           LDAP_TIMELIMIT_EXCEEDED (0x03)
           LDAP_SIZELIMIT_EXCEEDED (0x04)
           LDAP_COMPARE_FALSE (0x05)
           LDAP_COMPARE_TRUE (0x06)
           LDAP_STRONG_AUTH_NOT_SUPPORTED (0x07)
           LDAP_STRONG_AUTH_REQUIRED (0x08)
           LDAP_REFERRAL (0x0a)                            -- new in LDAPv3
           LDAP_ADMINLIMIT_EXCEEDED (0x0b)                 -- new in LDAPv3
           LDAP_UNAVAILABLE_CRITICAL_EXTENSION (0x0c)      -- new in LDAPv3
           LDAP_CONFIDENTIALITY_REQUIRED (0x0d)            -- new in LDAPv3
           LDAP_SASL_BIND_IN_PROGRESS (0x0e)               -- new in LDAPv3

Expires: May 2001                                              [Page 14]

C LDAP API        C LDAP Application Program Interface  17 November 2000

LDAP_NO_SUCH_ATTRIBUTE (0x10)
           LDAP_UNDEFINED_TYPE (0x11)
           LDAP_INAPPROPRIATE_MATCHING (0x12)
           LDAP_CONSTRAINT_VIOLATION (0x13)
           LDAP_TYPE_OR_VALUE_EXISTS (0x14)
           LDAP_INVALID_SYNTAX (0x15)
           LDAP_NO_SUCH_OBJECT (0x20)
           LDAP_ALIAS_PROBLEM (0x21)
           LDAP_INVALID_DN_SYNTAX (0x22)
           LDAP_IS_LEAF (0x23)                             -- not used in LDAPv3
           LDAP_ALIAS_DEREF_PROBLEM (0x24)
           LDAP_INAPPROPRIATE_AUTH (0x30)
           LDAP_INVALID_CREDENTIALS (0x31)
           LDAP_INSUFFICIENT_ACCESS (0x32)
           LDAP_BUSY (0x33)
           LDAP_UNAVAILABLE (0x34)
           LDAP_UNWILLING_TO_PERFORM (0x35)
           LDAP_LOOP_DETECT (0x36)
           LDAP_NAMING_VIOLATION (0x40)
           LDAP_OBJECT_CLASS_VIOLATION (0x41)
           LDAP_NOT_ALLOWED_ON_NONLEAF (0x42)
           LDAP_NOT_ALLOWED_ON_RDN (0x43)
           LDAP_ALREADY_EXISTS (0x44)
           LDAP_NO_OBJECT_CLASS_MODS (0x45)
           LDAP_RESULTS_TOO_LARGE (0x46)                   -- reserved for CLDAP
           LDAP_AFFECTS_MULTIPLE_DSAS (0x47)               -- new in LDAPv3
           LDAP_OTHER (0x50)
           LDAP_SERVER_DOWN (0x51)
           LDAP_LOCAL_ERROR (0x52)
           LDAP_ENCODING_ERROR (0x53)
           LDAP_DECODING_ERROR (0x54)
           LDAP_TIMEOUT (0x55)
           LDAP_AUTH_UNKNOWN (0x56)
           LDAP_FILTER_ERROR (0x57)
           LDAP_USER_CANCELLED (0x58)
           LDAP_PARAM_ERROR (0x59)
           LDAP_NO_MEMORY (0x5a)
           LDAP_CONNECT_ERROR (0x5b)
           LDAP_NOT_SUPPORTED (0x5c)
           LDAP_CONTROL_NOT_FOUND (0x5d)
           LDAP_NO_RESULTS_RETURNED (0x5e)
           LDAP_MORE_RESULTS_TO_RETURN (0x5f)
           LDAP_CLIENT_LOOP (0x60)
           LDAP_REFERRAL_LIMIT_EXCEEDED (0x61)

Expires: May 2001                                              [Page 15]

C LDAP API        C LDAP Application Program Interface  17 November 2000

11.  Performing LDAP Operations

This section describes each LDAP operation API call in detail. Most
functions take a "session handle," a pointer to an LDAP structure con-
taining per-connection information.  Many routines return results in an
LDAPMessage structure. These structures and others are described as
needed below.

11.1.  Initializing an LDAP Session

ldap_init() initializes a session with an LDAP server. The server is not
actually contacted until an operation is performed that requires it,
allowing various options to be set after initialization.

LDAP *ldap_init(
                   const char      *hostname,
                   int             portno
           );

Use of the following routine is deprecated:

LDAP *ldap_open(
                   const char      *hostname,
                   int             portno
           );

Unlike ldap_init(), ldap_open() attempts to make a server connection
before returning to the caller.  A more complete description can be
found in RFC 1823.

Parameters are:

hostname Contains a space-separated list of hostnames or dotted strings
         representing the IP address of hosts running an LDAP server to
         connect to. Each hostname in the list MAY include a port number
         which is separated from the host itself with a colon (:) char-
         acter.  The hosts will be tried in the order listed, stopping
         with the first one to which a successful connection is made.

Note: A suitable representation for including a literal IPv6[10]
   address in the hostname parameter is desired, but has not yet been
   determined or implemented in practice.

portno   Contains the TCP port number to connect to. The default LDAP
         port of 389 can be obtained by supplying the value zero (0) or
         the macro LDAP_PORT (389).  If a host includes a port number
         then this parameter is ignored.

Expires: May 2001                                              [Page 16]

C LDAP API        C LDAP Application Program Interface  17 November 2000

ldap_init() and ldap_open() both return a "session handle," a pointer to
an opaque structure that MUST be passed to subsequent calls pertaining
to the session. These routines return NULL if the session cannot be ini-
tialized in which case the operating system error reporting mechanism
can be checked to see why the call failed.

Note that if you connect to an LDAPv2 server, one of the LDAP bind calls
described below SHOULD be completed before other operations can be per-
formed on the session.  LDAPv3 does not require that a bind operation be
completed before other operations can be performed.

The calling program can set various attributes of the session by calling
the routines described in the next section.

11.2.  LDAP Session Handle Options

The LDAP session handle returned by ldap_init() is a pointer to an
opaque data type representing an LDAP session. In RFC 1823 this data
type was a structure exposed to the caller, and various fields in the
structure could be set to control aspects of the session, such as size
and time limits on searches.

In the interest of insulating callers from inevitable changes to this
structure, these aspects of the session are now accessed through a pair
of accessor functions, described below.

ldap_get_option() is used to access the current value of various
session-wide parameters. ldap_set_option() is used to set the value of
these parameters.  Note that some options are READ-ONLY and cannot be
set; it is an error to call ldap_set_option() and attempt to set a
READ-ONLY option.

Note that if automatic referral following is enabled (the default), any
connections created during the course of following referrals will
inherit the options associated with the session that sent the original
request that caused the referrals to be returned.

int ldap_get_option(
                   LDAP            *ld,
                   int             option,
                   void            *outvalue
           );

int ldap_set_option(
                   LDAP            *ld,
                   int             option,
                   const void      *invalue

Expires: May 2001                                              [Page 17]

C LDAP API        C LDAP Application Program Interface  17 November 2000

);

#define LDAP_OPT_ON     (<impl_void_star_value>)
           #define LDAP_OPT_OFF    ((void *)0)

LDAP_OPT_ON MUST be defined as a non-null pointer to void;  that is,
   <impl_void_star_value> MUST be replaced with a non-null pointer to
   void, e.g., one could use:
           #define LDAP_OPT_ON     ((void *)1)
   if that value is safe to use on the architecture where the API is
   implemented.

Parameters are:

ld     The session handle.  If this is NULL, a set of global defaults is
       accessed.  New LDAP session handles created with ldap_init() or
       ldap_open() inherit their characteristics from these global
       defaults.

option The name of the option being accessed or set. This parameter
       SHOULD be one of the following constants, which have the indi-
       cated meanings.  After the constant the actual hexadecimal value
       of the constant is listed in parentheses.

LDAP_OPT_API_INFO (0x00)
      Type for invalue parameter: not applicable (option is READ-ONLY)

Type for outvalue parameter: LDAPAPIInfo *

Description:
           Used to retrieve some basic information about the LDAP API
           implementation at execution time.  See the section "Retriev-
           ing Information About the API Implementation" above for more
           information.  This option is READ-ONLY and cannot be set.

LDAP_OPT_DEREF (0x02)
      Type for invalue parameter: int *

Type for outvalue parameter: int *

Description:
           Determines how aliases are handled during search. It SHOULD
           have one of the following values: LDAP_DEREF_NEVER (0x00),
           LDAP_DEREF_SEARCHING (0x01), LDAP_DEREF_FINDING (0x02), or
           LDAP_DEREF_ALWAYS (0x03).  The LDAP_DEREF_SEARCHING value
           means aliases are dereferenced during the search but not when
           locating the base object of the search. The

Expires: May 2001                                              [Page 18]

C LDAP API        C LDAP Application Program Interface  17 November 2000

LDAP_DEREF_FINDING value means aliases are dereferenced when
           locating the base object but not during the search.  The
           default value for this option is LDAP_DEREF_NEVER.

LDAP_OPT_SIZELIMIT (0x03)
      Type for invalue parameter: int *

Type for outvalue parameter: int *

Description:
           A limit on the number of entries to return from a search. A
           value of LDAP_NO_LIMIT (0) means no limit.  The default value
           for this option is LDAP_NO_LIMIT.

LDAP_OPT_TIMELIMIT (0x04)
      Type for invalue parameter: int *

Type for outvalue parameter: int *

Description:
           A limit on the number of seconds to spend on a search. A
           value of LDAP_NO_LIMIT (0) means no limit.  The default value
           for this option is LDAP_NO_LIMIT.  This value is passed to
           the server in the search request only; it does not affect how
           long the C LDAP API implementation itself will wait locally
           for search results.  Note that the timeout parameter passed
           to the ldap_search_ext_s() or ldap_result() functions can be
           used to specify a limit on how long the API implementation
           will wait for results.

LDAP_OPT_REFERRALS (0x08)
      Type for invalue parameter: void * (LDAP_OPT_ON or LDAP_OPT_OFF)

Type for outvalue parameter: int *

Description:
           Determines whether the LDAP library automatically follows
           referrals returned by LDAP servers or not. It MAY be set to
           one of the constants LDAP_OPT_ON or LDAP_OPT_OFF; any non-
           NULL pointer value passed to ldap_set_option() enables this
           option.  When reading the current setting using
           ldap_get_option(), a zero value means OFF and any non-zero
           value means ON.  By default, this option is ON.

LDAP_OPT_RESTART (0x09)
      Type for invalue parameter: void * (LDAP_OPT_ON or LDAP_OPT_OFF)

Type for outvalue parameter: int *

Expires: May 2001                                              [Page 19]

C LDAP API        C LDAP Application Program Interface  17 November 2000

Description:
           Determines whether LDAP I/O operations are automatically res-
           tarted if they abort prematurely. It MAY be set to one of the
           constants LDAP_OPT_ON or LDAP_OPT_OFF; any non-NULL pointer
           value passed to ldap_set_option() enables this option.  When
           reading the current setting using ldap_get_option(), a zero
           value means OFF and any non-zero value means ON. This option
           is useful if an LDAP I/O operation can be interrupted prema-
           turely, for example by a timer going off, or other interrupt.
           By default, this option is OFF.

LDAP_OPT_PROTOCOL_VERSION (0x11)
      Type for invalue parameter: int *

Type for outvalue parameter: int *

Description:
           This option indicates the version of the LDAP protocol used
           when communicating with the primary LDAP server. It SHOULD be
           one of the constants LDAP_VERSION2 (2) or LDAP_VERSION3 (3).
           If no version is set the default is LDAP_VERSION2 (2).

LDAP_OPT_SERVER_CONTROLS (0x12)
      Type for invalue parameter: LDAPControl **

Type for outvalue parameter: LDAPControl ***

Description:
           A default list of LDAP server controls to be sent with each
           request.  See the Working With Controls section below.

LDAP_OPT_CLIENT_CONTROLS (0x13)
      Type for invalue parameter: LDAPControl **

Type for outvalue parameter: LDAPControl ***

Description:
           A default list of client controls that affect the LDAP ses-
           sion.  See the Working With Controls section below.

LDAP_OPT_API_FEATURE_INFO (0x15)
      Type for invalue parameter: not applicable (option is READ-ONLY)

Type for outvalue parameter: LDAPAPIFeatureInfo *

Description:
           Used to retrieve version information about LDAP API extended
           features at execution time.  See the section "Retrieving

Expires: May 2001                                              [Page 20]

C LDAP API        C LDAP Application Program Interface  17 November 2000

Information About the API Implementation" above for more
           information.  This option is READ-ONLY and cannot be set.

LDAP_OPT_HOST_NAME (0x30)
      Type for invalue parameter: char *

Type for outvalue parameter: char **

Description:
           The host name (or list of hosts) for the primary LDAP server.
           See the definition of the hostname parameter to ldap_init()
           for the allowed syntax.  Note that if the portno parameter
           passed to ldap_init() is a value other than 0 or 389
           (LDAP_PORT), this value SHOULD include a string like
           ":portno" after each hostname or IP address that did not have
           one in the original hostname parameter that was passed to
           ldap_init().  For example, if this hostname value was passed
           to ldap_init():

"ldap.example.com:389 ldap2.example.com"

and the portno parameter passed to ldap_init() was 6389, then
           the value returned for the LDAP_OPT_HOST_NAME option SHOULD
           be:

"ldap.example.com:389 ldap2.example.com:6389"

LDAP_OPT_RESULT_CODE (0x31)
      Type for invalue parameter: int *

Type for outvalue parameter: int *

Description:
           The most recent local (API generated) or server returned LDAP
           result code that occurred for this session.

LDAP_OPT_ERROR_STRING (0x32)
      Type for invalue parameter: char *

Type for outvalue parameter: char **

Description:
           The message returned with the most recent LDAP error that
           occurred for this session.

LDAP_OPT_MATCHED_DN (0x33)
      Type for invalue parameter: char *

Expires: May 2001                                              [Page 21]

C LDAP API        C LDAP Application Program Interface  17 November 2000

Type for outvalue parameter: char **

Description:
           The matched DN value returned with the most recent LDAP error
           that occurred for this session.

outvalue The address of a place to put the value of the option. The
         actual type of this parameter depends on the setting of the
         option parameter.  For outvalues of type char ** and LDAPCon-
         trol **, a copy of the data that is associated with the LDAP
         session ld is returned; callers should dispose of the memory by
         calling ldap_memfree() or ldap_controls_free(), depending on
         the type of data returned.

invalue  A pointer to the value the option is to be given. The actual
         type of this parameter depends on the setting of the option
         parameter. The data associated with invalue is copied by the
         API implementation to allow callers of the API to dispose of or
         otherwise change their copy of the data after a successful call
         to ldap_set_option().  If a value passed for invalue is invalid
         or cannot be accepted by the implementation, ldap_set_option()
         should return -1 to indicate an error.

Both ldap_get_option() and ldap_set_option() return 0 if successful and
-1 if an error occurs.  If -1 is returned by either function, a specific
result code MAY be retrieved by calling ldap_get_option() with an option
value of LDAP_OPT_RESULT_CODE.  Note that there is no way to retrieve a
more specific result code if a call to ldap_get_option() with an option
value of LDAP_OPT_RESULT_CODE fails.

When a call to ldap_get_option() succeeds, the API implementation MUST
NOT change the state of the LDAP session handle or the state of the
underlying implementation in a way that affects the behavior of future
LDAP API calls.  When a call to ldap_get_option() fails, the only ses-
sion handle change permitted is setting the LDAP result code (as
returned by the LDAP_OPT_RESULT_CODE option).

When a call to ldap_set_option() fails, it MUST NOT change the state of
the LDAP session handle or the state of the underlying implementation in
a way that affects the behavior of future LDAP API calls.

Standards track documents that extend this specification and specify new
options SHOULD use values for option macros that are between 0x1000 and
0x3FFF inclusive.  Private and experimental extensions SHOULD use values
for the option macros that are between 0x4000 and 0x7FFF inclusive.  All
values below 0x1000 and above 0x7FFF that are not defined in this docu-
ment are reserved and SHOULD NOT be used.  The following macro MUST be

Expires: May 2001                                              [Page 22]

C LDAP API        C LDAP Application Program Interface  17 November 2000

defined by C LDAP API implementations to aid extension implementors:
   #define LDAP_OPT_PRIVATE_EXTENSION_BASE 0x4000  /* to 0x7FFF inclusive */

11.3.  Working With Controls

LDAPv3 operations can be extended through the use of controls.  Controls
can be sent to a server or returned to the client with any LDAP message.
These controls are referred to as server controls.

The LDAP API also supports a client-side extension mechanism through the
use of client controls. These controls affect the behavior of the LDAP
API only and are never sent to a server.  A common data structure is
used to represent both types of controls:

typedef struct ldapcontrol {
                   char            *ldctl_oid;
                   struct berval   ldctl_value;
                   char            ldctl_iscritical;
           } LDAPControl;

The fields in the ldapcontrol structure have the following meanings:

ldctl_oid        The control type, represented as a string.

ldctl_value      The data associated with the control (if any).  To
                 specify a zero-length value, set ldctl_value.bv_len to
                 zero and ldctl_value.bv_val to a zero-length string.
                 To indicate that no data is associated with the con-
                 trol, set ldctl_value.bv_val to NULL.

ldctl_iscritical Indicates whether the control is critical of not. If
                 this field is non-zero, the operation will only be car-
                 ried out if the control is recognized by the server
                 and/or client.  Note that the LDAP unbind and abandon
                 operations have no server response, so clients SHOULD
                 NOT mark server controls critical when used with these
                 two operations.

Some LDAP API calls allocate an ldapcontrol structure or a NULL-
terminated array of ldapcontrol structures.  The following routines can
be used to dispose of a single control or an array of controls:

void ldap_control_free( LDAPControl *ctrl );
           void ldap_controls_free( LDAPControl **ctrls );
If the ctrl or ctrls parameter is NULL, these calls do nothing.

Expires: May 2001                                              [Page 23]

C LDAP API        C LDAP Application Program Interface  17 November 2000

A set of controls that affect the entire session can be set using the
ldap_set_option() function (see above).  A list of controls can also be
passed directly to some LDAP API calls such as ldap_search_ext(), in
which case any controls set for the session through the use of
ldap_set_option() are ignored. Control lists are represented as a NULL-
terminated array of pointers to ldapcontrol structures.

Server controls are defined by LDAPv3 protocol extension documents; for
example, a control has been proposed to support server-side sorting of
search results [11].

One client control is defined in this document (described in the follow-
ing section).  Other client controls MAY be defined in future revisions
of this document or in documents that extend this API.

11.3.1.  A Client Control That Governs Referral Processing

As described previously in the section "LDAP Session Handle Options,"
applications can enable and disable automatic chasing of referrals on a
session-wide basic by using the ldap_set_option() function with the
LDAP_OPT_REFERRALS option.  It is also useful to govern automatic refer-
ral chasing on per-request basis.  A client control with an OID of
1.2.840.113556.1.4.616 exists to provide this functionality.

/* OID for referrals client control */
   #define LDAP_CONTROL_REFERRALS              "1.2.840.113556.1.4.616"

/* Flags for referrals client control value */
   #define LDAP_CHASE_SUBORDINATE_REFERRALS    0x00000020U
   #define LDAP_CHASE_EXTERNAL_REFERRALS       0x00000040U

To create a referrals client control, the ldctl_oid field of an LDAPCon-
trol structure MUST be set to LDAP_CONTROL_REFERRALS
("1.2.840.113556.1.4.616") and the ldctl_value field MUST be set to a
value that contains a set of flags.  The ldctl_value.bv_len field MUST
be set to sizeof(ber_uint_t), and the ldctl_value.bv_val field MUST
point to a ber_uint_t which contains the flags value." The ber_uint_t
type is define in the section "BER Data Structures and Types" below.

The flags value can be set to zero to disable automatic chasing of
referrals and LDAPv3 references altogether.  Alternatively, the flags
value can be set to the value LDAP_CHASE_SUBORDINATE_REFERRALS
(0x00000020U) to indicate that only LDAPv3 search continuation refer-
ences are to be automatically chased by the API implementation, to the
value LDAP_CHASE_EXTERNAL_REFERRALS (0x00000040U) to indicate that only
LDAPv3 referrals are to be automatically chased, or the logical OR of
the two flag values (0x00000060U) to indicate that both referrals and

Expires: May 2001                                              [Page 24]

C LDAP API        C LDAP Application Program Interface  17 November 2000

references are to be automatically chased.

11.4.  Authenticating to the directory

The following functions are used to authenticate an LDAP client to an
LDAP directory server.

The ldap_sasl_bind() and ldap_sasl_bind_s() functions can be used to do
general and extensible authentication over LDAP through the use of the
Simple Authentication Security Layer [12].  The routines both take the
dn to bind as, the method to use, as a dotted-string representation of
an OID identifying the method, and a struct berval holding the creden-
tials. The special constant value LDAP_SASL_SIMPLE (NULL) can be passed
to request simple authentication, or the simplified routines
ldap_simple_bind() or ldap_simple_bind_s() can be used.

int ldap_sasl_bind(
                   LDAP                    *ld,
                   const char              *dn,
                   const char              *mechanism,
                   const struct berval     *cred,
                   LDAPControl             **serverctrls,
                   LDAPControl             **clientctrls,
                   int                     *msgidp
           );

int ldap_sasl_bind_s(
                   LDAP                    *ld,
                   const char              *dn,
                   const char              *mechanism,
                   const struct berval     *cred,
                   LDAPControl             **serverctrls,
                   LDAPControl             **clientctrls,
                   struct berval           **servercredp
           );

int ldap_simple_bind(
                   LDAP                    *ld,
                   const char              *dn,
                   const char              *passwd
           );

int ldap_simple_bind_s(
                   LDAP                    *ld,
                   const char              *dn,
                   const char              *passwd
           );

Expires: May 2001                                              [Page 25]

C LDAP API        C LDAP Application Program Interface  17 November 2000

The use of the following routines is deprecated and more complete
   descriptions can be found in RFC 1823:

int ldap_bind( LDAP *ld, const char *dn, const char *cred,
                   int method );

int ldap_bind_s( LDAP *ld, const char *dn, const char *cred,
                   int method );

int ldap_kerberos_bind( LDAP *ld, const char *dn );

int ldap_kerberos_bind_s( LDAP *ld, const char *dn );

Parameters are:

ld           The session handle.

dn           The name of the entry to bind as.  If NULL, a zero length
             DN is sent to the server.

mechanism    Either LDAP_SASL_SIMPLE (NULL) to get simple authentica-
             tion, or a text string identifying the SASL method.

cred         The credentials with which to authenticate. Arbitrary
             credentials can be passed using this parameter. The format
             and content of the credentials depends on the setting of
             the mechanism parameter.  If the cred parameter is NULL and
             the mechanism is LDAP_SASL_SIMPLE, a zero-length octet
             string is sent to the server in the simple credentials
             field of the bind request.  If the cred parameter is NULL
             and the mechanism is anything else, no credentials are sent
             to the server in the bind request.

passwd       For ldap_simple_bind(), the password that is sent to the
             server in the simple credentials field of the bind request.
             If NULL, a zero length password is sent to the server.

serverctrls  List of LDAP server controls, or NULL if no server controls
             are used.

clientctrls  List of client controls, or NULL if no client controls are
             used.

msgidp       This result parameter will be set to the message id of the
             request if the ldap_sasl_bind() call succeeds.  The value
             is undefined if a value other than LDAP_SUCCESS is
             returned.

Expires: May 2001                                              [Page 26]

C LDAP API        C LDAP Application Program Interface  17 November 2000

servercredp  This result parameter will be filled in with the creden-
             tials passed back by the server for mutual authentication,
             if given. An allocated berval structure is returned that
             SHOULD be disposed of by calling ber_bvfree().  NULL SHOULD
             be passed to ignore this field.  If an API error occurs or
             the server did not return any credentials, *servercredp is
             set to NULL.

Additional parameters for the deprecated routines are not described.
Interested readers are referred to RFC 1823.

The ldap_sasl_bind() function initiates an asynchronous bind operation
and returns the constant LDAP_SUCCESS if the request was successfully
sent, or another LDAP result code if not.  See the section below on
error handling for more information about possible errors and how to
interpret them.  If successful, ldap_sasl_bind() places the message id
of the request in *msgidp. A subsequent call to ldap_result(), described
below, can be used to obtain the result of the bind.

The ldap_simple_bind() function initiates a simple asynchronous bind
operation and returns the message id of the operation initiated.  A sub-
sequent call to ldap_result(), described below, can be used to obtain
the result of the bind. In case of error, ldap_simple_bind() will return
-1, setting the session error parameters in the LDAP structure appropri-
ately.

The synchronous ldap_sasl_bind_s() and ldap_simple_bind_s() functions
both return the result of the operation, either the constant
LDAP_SUCCESS if the operation was successful, or another LDAP result
code if it was not. See the section below on error handling for more
information about possible errors and how to interpret them.

Note that if an LDAPv2 server is contacted, no other operations over the
connection can be attempted before a bind call has successfully com-
pleted.

Subsequent bind calls can be used to re-authenticate over the same con-
nection, and multistep SASL sequences can be accomplished through a
sequence of calls to ldap_sasl_bind() or ldap_sasl_bind_s().

11.5.  Closing the session

The following functions are used to unbind from the directory, close
open connections, and dispose of the session handle.

int ldap_unbind_ext( LDAP *ld, LDAPControl **serverctrls,
                   LDAPControl **clientctrls );

Expires: May 2001                                              [Page 27]

C LDAP API        C LDAP Application Program Interface  17 November 2000

int ldap_unbind( LDAP *ld );

int ldap_unbind_s( LDAP *ld );

Parameters are:

ld           The session handle.

serverctrls  List of LDAP server controls, or NULL if no server controls
             are to be used.

clientctrls  List of client controls, or NULL if no client controls are
             to be used.

The ldap_unbind_ext(), ldap_unbind() and ldap_unbind_s() all work syn-
chronously in the sense that they send an unbind request to the server,
close all open connections associated with the LDAP session handle, and
dispose of all resources associated with the session handle before
returning.  Note, however, that there is no server response to an LDAP
unbind operation.  All three of the unbind functions return LDAP_SUCCESS
(or another LDAP result code if the request cannot be sent to the LDAP
server).  After a call to one of the unbind functions, the session han-
dle ld is invalid and it is illegal to make any further LDAP API calls
using ld.

The ldap_unbind() and ldap_unbind_s() functions behave identically.  The
ldap_unbind_ext() function allows server and client controls to be
included explicitly, but note that since there is no server response to
an unbind request there is no way to receive a response to a server con-
trol sent with an unbind request.

11.6.  Searching

The following functions are used to search the LDAP directory, returning
a requested set of attributes for each entry matched.  There are five
variations.

int ldap_search_ext(
                   LDAP            *ld,
                   const char      *base,
                   int             scope,
                   const char      *filter,
                   char            **attrs,
                   int             attrsonly,
                   LDAPControl     **serverctrls,
                   LDAPControl     **clientctrls,

Expires: May 2001                                              [Page 28]

C LDAP API        C LDAP Application Program Interface  17 November 2000

struct timeval  *timeout,
                   int             sizelimit,
                   int             *msgidp
           );

int ldap_search_ext_s(
                   LDAP            *ld,
                   const char      *base,
                   int             scope,
                   const char      *filter,
                   char            **attrs,
                   int             attrsonly,
                   LDAPControl     **serverctrls,
                   LDAPControl     **clientctrls,
                   struct timeval  *timeout,
                   int             sizelimit,
                   LDAPMessage     **res
           );

int ldap_search(
                   LDAP            *ld,
                   const char      *base,
                   int             scope,
                   const char      *filter,
                   char            **attrs,
                   int             attrsonly
           );

int ldap_search_s(
                   LDAP            *ld,
                   const char      *base,
                   int             scope,
                   const char      *filter,
                   char            **attrs,
                   int             attrsonly,
                   LDAPMessage     **res
           );

int ldap_search_st(
                   LDAP            *ld,
                   const char      *base,
                   int             scope,
                   const char      *filter,
                   char            **attrs,
                   int             attrsonly,
                   struct timeval  *timeout,
                   LDAPMessage     **res
           );

Expires: May 2001                                              [Page 29]

C LDAP API        C LDAP Application Program Interface  17 November 2000

Parameters are:

ld           The session handle.

base         The dn of the entry at which to start the search.  If NULL,
             a zero length DN is sent to the server.

scope        One of LDAP_SCOPE_BASE (0x00), LDAP_SCOPE_ONELEVEL (0x01),
             or LDAP_SCOPE_SUBTREE (0x02), indicating the scope of the
             search.

filter       A character string as described in [13], representing the
             search filter.  The value NULL can be passed to indicate
             that the filter "(objectclass=*)" which matches all entries
             is to be used.  Note that if the caller of the API is using
             LDAPv2, only a subset of the filter functionality described
             in [13] can be successfully used.

attrs        A NULL-terminated array of strings indicating which attri-
             butes to return for each matching entry. Passing NULL for
             this parameter causes all available user attributes to be
             retrieved.  The special constant string LDAP_NO_ATTRS
             ("1.1") MAY be used as the only string in the array to
             indicate that no attribute types are to be returned by the
             server.  The special constant string LDAP_ALL_USER_ATTRS
             ("*") can be used in the attrs array along with the names
             of some operational attributes to indicate that all user
             attributes plus the listed operational attributes are to be
             returned.

attrsonly    A boolean value that MUST be zero if both attribute types
             and values are to be returned, and non-zero if only types
             are wanted.

timeout      For the ldap_search_st() function, this specifies the local
             search timeout value (if it is NULL, the timeout is infin-
             ite).  If a zero timeout (where tv_sec and tv_usec are both
             zero) is passed, API implementations SHOULD return
             LDAP_PARAM_ERROR.

For the ldap_search_ext() and ldap_search_ext_s() func-
             tions, the timeout parameter specifies both the local
             search timeout value and the operation time limit that is
             sent to the server within the search request.  Passing a
             NULL value for timeout causes the default timeout stored in
             the LDAP session handle (set by using ldap_set_option()
             with the LDAP_OPT_TIMELIMIT parameter) to be sent to the
             server with the request but an infinite local search

Expires: May 2001                                              [Page 30]

C LDAP API        C LDAP Application Program Interface  17 November 2000

timeout to be used.  If a zero timeout (where tv_sec and
             tv_usec are both zero) is passed in, API implementations
             SHOULD return LDAP_PARAM_ERROR.  If a zero value for tv_sec
             is used but tv_usec is non-zero, an operation time limit of
             1 SHOULD be passed to the LDAP server as the operation time
             limit.  For other values of tv_sec, the tv_sec value itself
             SHOULD be passed to the LDAP server.

sizelimit    For the ldap_search_ext() and ldap_search_ext_s() calls,
             this is a limit on the number of entries to return from the
             search.  A value of LDAP_NO_LIMIT (0) means no limit.  A
             value of LDAP_DEFAULT_SIZELIMIT (-1) means use the default
             timeout from the LDAP session handle (which is set by cal-
             ling ldap_set_option() with the LDAP_OPT_SIZELIMIT parame-
             ter).

res          For the synchronous calls, this is a result parameter which
             will contain the results of the search upon completion of
             the call.  If an API error occurs or no results are
             returned, *res is set to NULL.

serverctrls  List of LDAP server controls, or NULL if no server controls
             are to be used.

clientctrls  List of client controls, or NULL if no client controls are
             to be used.

msgidp       This result parameter will be set to the message id of the
             request if the ldap_search_ext() call succeeds. The value
             is undefined if a value other than LDAP_SUCCESS is
             returned.

There are three options in the session handle ld which potentially
affect how the search is performed. They are:

LDAP_OPT_SIZELIMIT
        LDAP_OPT_TIMELIMIT
        LDAP_OPT_DEREF

These options are fully described in the earlier section "LDAP Session
Handle Options."

The ldap_search_ext() function initiates an asynchronous search opera-
tion and returns the constant LDAP_SUCCESS if the request was success-
fully sent, or another LDAP result code if not.  See the section below
on error handling for more information about possible errors and how to
interpret them.  If successful, ldap_search_ext() places the message id
of the request in *msgidp. A subsequent call to ldap_result(), described

Expires: May 2001                                              [Page 31]

C LDAP API        C LDAP Application Program Interface  17 November 2000

below, can be used to obtain the results from the search.  These results
can be parsed using the result parsing routines described in detail
later.

Similar to ldap_search_ext(), the ldap_search() function initiates an
asynchronous search operation and returns the message id of the opera-
tion initiated.  As for ldap_search_ext(), a subsequent call to
ldap_result(), described below, can be used to obtain the result of the
bind. In case of error, ldap_search() will return -1, setting the ses-
sion error parameters in the LDAP structure appropriately.

The synchronous ldap_search_ext_s(), ldap_search_s(), and
ldap_search_st() functions all return the result of the operation,
either the constant LDAP_SUCCESS if the operation was successful, or
another LDAP result code if it was not. See the section below on error
handling for more information about possible errors and how to interpret
them.  Entries returned from the search (if any) are contained in the
res parameter. This parameter is opaque to the caller.  Entries, attri-
butes, values, etc., can be extracted by calling the parsing routines
described below. The results contained in res SHOULD be freed when no
longer in use by calling ldap_msgfree(), described later.

The ldap_search_ext() and ldap_search_ext_s() functions support LDAPv3
server controls, client controls, and allow varying size and time limits
to be easily specified for each search operation.  The ldap_search_st()
function is identical to ldap_search_s() except that it takes an addi-
tional parameter specifying a local timeout for the search.  The local
search timeout is used to limit the amount of time the API implementa-
tion will wait for a search to complete.  After the local search timeout
expires, the API implementation will send an abandon operation to abort
the search operation.

11.7.  Reading an Entry

LDAP does not support a read operation directly. Instead, this operation
is emulated by a search with base set to the DN of the entry to read,
scope set to LDAP_SCOPE_BASE, and filter set to "(objectclass=*)" or
NULL. attrs contains the list of attributes to return.

11.8.  Listing the Children of an Entry

LDAP does not support a list operation directly. Instead, this operation
is emulated by a search with base set to the DN of the entry to list,
scope set to LDAP_SCOPE_ONELEVEL, and filter set to "(objectclass=*)" or
NULL. attrs contains the list of attributes to return for each child
entry.

Expires: May 2001                                              [Page 32]

C LDAP API        C LDAP Application Program Interface  17 November 2000

11.9.  Comparing a Value Against an Entry

The following routines are used to compare a given attribute value
assertion against an LDAP entry.  There are four variations:

int ldap_compare_ext(
                   LDAP                    *ld,
                   const char              *dn,
                   const char              *attr,
                   const struct berval     *bvalue,
                   LDAPControl             **serverctrls,
                   LDAPControl             **clientctrls,
                   int                     *msgidp
           );

int ldap_compare_ext_s(
                   LDAP                    *ld,
                   const char              *dn,
                   const char              *attr,
                   const struct berval     *bvalue,
                   LDAPControl             **serverctrls,
                   LDAPControl             **clientctrls
           );

int ldap_compare(
                   LDAP                    *ld,
                   const char              *dn,
                   const char              *attr,
                   const char              *value
           );

int ldap_compare_s(
                   LDAP                    *ld,
                   const char              *dn,
                   const char              *attr,
                   const char              *value
           );

Parameters are:

ld           The session handle.

dn           The name of the entry to compare against.  If NULL, a zero
             length DN is sent to the server.

attr         The attribute to compare against.

bvalue       The attribute value to compare against those found in the

Expires: May 2001                                              [Page 33]

C LDAP API        C LDAP Application Program Interface  17 November 2000

given entry. This parameter is used in the extended rou-
             tines and is a pointer to a struct berval so it is possible
             to compare binary values.

value        A string attribute value to compare against, used by the
             ldap_compare() and ldap_compare_s() functions.  Use
             ldap_compare_ext() or ldap_compare_ext_s() if you need to
             compare binary values.

serverctrls  List of LDAP server controls, or NULL if no server controls
             are to be used.

clientctrls  List of client controls, or NULL if no client controls are
             to be used.

msgidp       This result parameter will be set to the message id of the
             request if the ldap_compare_ext() call succeeds. The value
             is undefined if a value other than LDAP_SUCCESS is
             returned.

The ldap_compare_ext() function initiates an asynchronous compare opera-
tion and returns the constant LDAP_SUCCESS if the request was success-
fully sent, or another LDAP result code if not.  See the section below
on error handling for more information about possible errors and how to
interpret them.  If successful, ldap_compare_ext() places the message id
of the request in *msgidp. A subsequent call to ldap_result(), described
below, can be used to obtain the result of the compare.

Similar to ldap_compare_ext(), the ldap_compare() function initiates an
asynchronous compare operation and returns the message id of the opera-
tion initiated.  As for ldap_compare_ext(), a subsequent call to
ldap_result(), described below, can be used to obtain the result of the
bind. In case of error, ldap_compare() will return -1, setting the ses-
sion error parameters in the LDAP structure appropriately.

The synchronous ldap_compare_ext_s() and ldap_compare_s() functions both
return the result of the operation: one of the constants
LDAP_COMPARE_TRUE or LDAP_COMPARE_FALSE if the operation was successful,
or another LDAP result code if it was not.  See the section below on
error handling for more information about possible errors and how to
interpret them.

The ldap_compare_ext() and ldap_compare_ext_s() functions support LDAPv3
server controls and client controls.

Expires: May 2001                                              [Page 34]

C LDAP API        C LDAP Application Program Interface  17 November 2000

11.10.  Modifying an entry

The following routines are used to modify an existing LDAP entry.  There
are four variations:

typedef union mod_vals_u {
                   char            **modv_strvals;
                   struct berval   **modv_bvals;
           } mod_vals_u_t;

typedef struct ldapmod {
                   int             mod_op;
                   char            *mod_type;
                   mod_vals_u_t    mod_vals;
           } LDAPMod;
           #define mod_values      mod_vals.modv_strvals
           #define mod_bvalues     mod_vals.modv_bvals

int ldap_modify_ext(
                   LDAP            *ld,
                   const char      *dn,
                   LDAPMod         **mods,
                   LDAPControl     **serverctrls,
                   LDAPControl     **clientctrls,
                   int             *msgidp
           );

int ldap_modify_ext_s(
                   LDAP            *ld,
                   const char      *dn,
                   LDAPMod         **mods,
                   LDAPControl     **serverctrls,
                   LDAPControl     **clientctrls
           );

int ldap_modify(
                   LDAP            *ld,
                   const char      *dn,
                   LDAPMod         **mods
           );

int ldap_modify_s(
                   LDAP            *ld,
                   const char      *dn,
                   LDAPMod         **mods
           );

Parameters are:

Expires: May 2001                                              [Page 35]

C LDAP API        C LDAP Application Program Interface  17 November 2000

ld           The session handle.

dn           The name of the entry to modify.  If NULL, a zero length DN
             is sent to the server.

mods         A NULL-terminated array of modifications to make to the
             entry.

serverctrls  List of LDAP server controls, or NULL if no server controls
             are to be used.

clientctrls  List of client controls, or NULL if no client controls are
             to be used.

msgidp       This result parameter will be set to the message id of the
             request if the ldap_modify_ext() call succeeds. The value
             is undefined if a value other than LDAP_SUCCESS is
             returned.

The fields in the LDAPMod structure have the following meanings:

mod_op       The modification operation to perform. It MUST be one of
             LDAP_MOD_ADD (0x00), LDAP_MOD_DELETE (0x01), or
             LDAP_MOD_REPLACE (0x02).  This field also indicates the
             type of values included in the mod_vals union. It is logi-
             cally ORed with LDAP_MOD_BVALUES (0x80) to select the
             mod_bvalues form. Otherwise, the mod_values form is used.

mod_type     The type of the attribute to modify.

mod_vals     The values (if any) to add, delete, or replace. Only one of
             the mod_values or mod_bvalues variants can be used,
             selected by ORing the mod_op field with the constant
             LDAP_MOD_BVALUES. mod_values is a NULL-terminated array of
             zero-terminated strings and mod_bvalues is a NULL-
             terminated array of berval structures that can be used to
             pass binary values such as images.

For LDAP_MOD_ADD modifications, the given values are added to  the
entry, creating the attribute if necessary.

For LDAP_MOD_DELETE modifications, the given values are deleted from the
entry, removing the attribute if no values remain. If the entire attri-
bute is to be deleted, the mod_vals field can be set to NULL.

For LDAP_MOD_REPLACE modifications, the attribute will have the listed
values after the modification, having been created if necessary, or
removed if the mod_vals field is NULL. All modifications are performed

Expires: May 2001                                              [Page 36]

C LDAP API        C LDAP Application Program Interface  17 November 2000

in the order in which they are listed.

The ldap_modify_ext() function initiates an asynchronous modify opera-
tion and returns the constant LDAP_SUCCESS if the request was success-
fully sent, or another LDAP result code if not.  See the section below
on error handling for more information about possible errors and how to
interpret them.  If successful, ldap_modify_ext() places the message id
of the request in *msgidp. A subsequent call to ldap_result(), described
below, can be used to obtain the result of the modify.

Similar to ldap_modify_ext(), the ldap_modify() function initiates an
asynchronous modify operation and returns the message id of the opera-
tion initiated.  As for ldap_modify_ext(), a subsequent call to
ldap_result(), described below, can be used to obtain the result of the
modify. In case of error, ldap_modify() will return -1, setting the ses-
sion error parameters in the LDAP structure appropriately.

The synchronous ldap_modify_ext_s() and ldap_modify_s() functions both
return the result of the operation, either the constant LDAP_SUCCESS if
the operation was successful, or another LDAP result code if it was not.
See the section below on error handling for more information about pos-
sible errors and how to interpret them.

The ldap_modify_ext() and ldap_modify_ext_s() functions support LDAPv3
server controls and client controls.

11.11.  Modifying the Name of an Entry

In LDAPv2, the ldap_modrdn(), ldap_modrdn_s(), ldap_modrdn2(), and
ldap_modrdn2_s() routines were used to change the name of an LDAP entry.
They could only be used to change the least significant component of a
name (the RDN or relative distinguished name). LDAPv3 provides the
Modify DN protocol operation that allows more general name change
access. The ldap_rename() and ldap_rename_s() routines are used to
change the name of an entry, and the use of the ldap_modrdn(),
ldap_modrdn_s(), ldap_modrdn2(), and ldap_modrdn2_s() routines is depre-
cated.

int ldap_rename(
                   LDAP            *ld,
                   const char      *dn,
                   const char      *newrdn,
                   const char      *newparent,
                   int             deleteoldrdn,
                   LDAPControl     **serverctrls,
                   LDAPControl     **clientctrls,
                   int             *msgidp

Expires: May 2001                                              [Page 37]

C LDAP API        C LDAP Application Program Interface  17 November 2000

);
           int ldap_rename_s(
                   LDAP            *ld,
                   const char      *dn,
                   const char      *newrdn,
                   const char      *newparent,
                   int             deleteoldrdn,
                   LDAPControl     **serverctrls,
                   LDAPControl     **clientctrls
           );

The use of the following routines is deprecated and more complete
   descriptions can be found in RFC 1823:

int ldap_modrdn(
                   LDAP            *ld,
                   const char      *dn,
                   const char      *newrdn
           );
           int ldap_modrdn_s(
                   LDAP            *ld,
                   const char      *dn,
                   const char      *newrdn
           );
           int ldap_modrdn2(
                   LDAP            *ld,
                   const char      *dn,
                   const char      *newrdn,
                   int             deleteoldrdn
           );
           int ldap_modrdn2_s(
                   LDAP            *ld,
                   const char      *dn,
                   const char      *newrdn,
                   int             deleteoldrdn
           );

Parameters are:

ld           The session handle.

dn           The name of the entry whose DN is to be changed.  If NULL,
             a zero length DN is sent to the server.

newrdn       The new RDN to give the entry.

newparent    The new parent, or superior entry.  If this parameter is
             NULL, only the RDN of the entry is changed.  The root DN

Expires: May 2001                                              [Page 38]

C LDAP API        C LDAP Application Program Interface  17 November 2000

SHOULD be specified by passing a zero length string, "".
             The newparent parameter SHOULD always be NULL when using
             version 2 of the LDAP protocol; otherwise the server's
             behavior is undefined.

deleteoldrdn This parameter only has meaning on the rename routines if
             newrdn is different than the old RDN. It is a boolean
             value, if non-zero indicating that the old RDN value(s) is
             to be removed, if zero indicating that the old RDN value(s)
             is to be retained as non-distinguished values of the entry.

serverctrls  List of LDAP server controls, or NULL if no server controls
             are to be used.

clientctrls  List of client controls, or NULL if no client controls are
             to be used.

msgidp       This result parameter will be set to the message id of the
             request if the ldap_rename() call succeeds. The value is
             undefined if a value other than LDAP_SUCCESS is returned.

The ldap_rename() function initiates an asynchronous modify DN operation
and returns the constant LDAP_SUCCESS if the request was successfully
sent, or another LDAP result code if not.  See the section below on
error handling for more information about possible errors and how to
interpret them.  If successful, ldap_rename() places the DN message id
of the request in *msgidp. A subsequent call to ldap_result(), described
below, can be used to obtain the result of the rename.

The synchronous ldap_rename_s() returns the result of the operation,
either the constant LDAP_SUCCESS if the operation was successful, or
another LDAP result code if it was not.  See the section below on error
handling for more information about possible errors and how to interpret
them.

The ldap_rename() and ldap_rename_s() functions both support LDAPv3
server controls and client controls.

11.12.  Adding an entry

The following functions are used to add entries to the LDAP directory.
There are four variations:

int ldap_add_ext(
                   LDAP            *ld,
                   const char      *dn,
                   LDAPMod         **attrs,

Expires: May 2001                                              [Page 39]

C LDAP API        C LDAP Application Program Interface  17 November 2000

LDAPControl     **serverctrls,
                   LDAPControl     **clientctrls,
                   int             *msgidp
           );

int ldap_add_ext_s(
                   LDAP            *ld,
                   const char      *dn,
                   LDAPMod         **attrs,
                   LDAPControl     **serverctrls,
                   LDAPControl     **clientctrls
           );

int ldap_add(
                   LDAP            *ld,
                   const char      *dn,
                   LDAPMod         **attrs
           );

int ldap_add_s(
                   LDAP            *ld,
                   const char      *dn,
                   LDAPMod         **attrs
           );

Parameters are:

ld           The session handle.

dn           The name of the entry to add.  If NULL, a zero length DN is
             sent to the server.

attrs        The entry's attributes, specified using the LDAPMod struc-
             ture defined for ldap_modify(). The mod_type and mod_vals
             fields MUST be filled in.  The mod_op field is ignored
             unless ORed with the constant LDAP_MOD_BVALUES, used to
             select the mod_bvalues case of the mod_vals union.

serverctrls  List of LDAP server controls, or NULL if no server controls
             are to be used.

clientctrls  List of client controls, or NULL if no client controls are
             to be used.

msgidp       This result parameter will be set to the message id of the
             request if the ldap_add_ext() call succeeds. The value is
             undefined if a value other than LDAP_SUCCESS is returned.

Expires: May 2001                                              [Page 40]

C LDAP API        C LDAP Application Program Interface  17 November 2000

Note that the parent of the entry being added must already exist or the
parent must be empty (i.e., equal to the root DN) for an add to succeed.

The ldap_add_ext() function initiates an asynchronous add operation and
returns the constant LDAP_SUCCESS if the request was successfully sent,
or another LDAP result code if not.  See the section below on error han-
dling for more information about possible errors and how to interpret
them.  If successful, ldap_add_ext() places the message id of the
request in *msgidp. A subsequent call to ldap_result(), described below,
can be used to obtain the result of the add.

Similar to ldap_add_ext(), the ldap_add() function initiates an asyn-
chronous add operation and returns the message id of the operation ini-
tiated.  As for ldap_add_ext(), a subsequent call to ldap_result(),
described below, can be used to obtain the result of the add. In case of
error, ldap_add() will return -1, setting the session error parameters
in the LDAP structure appropriately.

The synchronous ldap_add_ext_s() and ldap_add_s() functions both return
the result of the operation, either the constant LDAP_SUCCESS if the
operation was successful, or another LDAP result code if it was not.
See the section below on error handling for more information about pos-
sible errors and how to interpret them.

The ldap_add_ext() and ldap_add_ext_s() functions support LDAPv3 server
controls and client controls.

11.13.  Deleting an entry

The following functions are used to delete a leaf entry from the LDAP
directory.  There are four variations:

int ldap_delete_ext(
                   LDAP            *ld,
                   const char      *dn,
                   LDAPControl     **serverctrls,
                   LDAPControl     **clientctrls,
                   int             *msgidp
           );

int ldap_delete_ext_s(
                   LDAP            *ld,
                   const char      *dn,
                   LDAPControl     **serverctrls,
                   LDAPControl     **clientctrls
           );

Expires: May 2001                                              [Page 41]

C LDAP API        C LDAP Application Program Interface  17 November 2000

int ldap_delete(
                   LDAP            *ld,
                   const char      *dn
           );

int ldap_delete_s(
                   LDAP            *ld,
                   const char      *dn
           );

Parameters are:

ld           The session handle.

dn           The name of the entry to delete.  If NULL, a zero length DN
             is sent to the server.

serverctrls  List of LDAP server controls, or NULL if no server controls
             are to be used.

clientctrls  List of client controls, or NULL if no client controls are
             to be used.

msgidp       This result parameter will be set to the message id of the
             request if the ldap_delete_ext() call succeeds. The value
             is undefined if a value other than LDAP_SUCCESS is
             returned.

Note that the entry to delete must be a leaf entry (i.e., it must have
no children). Deletion of entire subtrees in a single operation is not
supported by LDAP.

The ldap_delete_ext() function initiates an asynchronous delete opera-
tion and returns the constant LDAP_SUCCESS if the request was success-
fully sent, or another LDAP result code if not.  See the section below
on error handling for more information about possible errors and how to
interpret them.  If successful, ldap_delete_ext() places the message id
of the request in *msgidp. A subsequent call to ldap_result(), described
below, can be used to obtain the result of the delete.

Similar to ldap_delete_ext(), the ldap_delete() function initiates an
asynchronous delete operation and returns the message id of the opera-
tion initiated.  As for ldap_delete_ext(), a subsequent call to
ldap_result(), described below, can be used to obtain the result of the
delete. In case of error, ldap_delete() will return -1, setting the ses-
sion error parameters in the LDAP structure appropriately.

Expires: May 2001                                              [Page 42]

C LDAP API        C LDAP Application Program Interface  17 November 2000

The synchronous ldap_delete_ext_s() and ldap_delete_s() functions both
return the result of the operation, either the constant LDAP_SUCCESS if
the operation was successful, or another LDAP result code if it was not.
See the section below on error handling for more information about pos-
sible errors and how to interpret them.

The ldap_delete_ext() and ldap_delete_ext_s() functions support LDAPv3
server controls and client controls.

11.14.  Extended Operations

The ldap_extended_operation() and ldap_extended_operation_s() routines
allow extended LDAP operations to be passed to the server, providing a
general protocol extensibility mechanism.

int ldap_extended_operation(
                   LDAP                    *ld,
                   const char              *requestoid,
                   const struct berval     *requestdata,
                   LDAPControl             **serverctrls,
                   LDAPControl             **clientctrls,
                   int                     *msgidp
           );

int ldap_extended_operation_s(
                   LDAP                    *ld,
                   const char              *requestoid,
                   const struct berval     *requestdata,
                   LDAPControl             **serverctrls,
                   LDAPControl             **clientctrls,
                   char                    **retoidp,
                   struct berval           **retdatap
           );

Parameters are:

ld           The session handle.

requestoid   The dotted-OID text string naming the request.

requestdata  The arbitrary data needed by the operation (if NULL, no
             data is sent to the server).

serverctrls  List of LDAP server controls, or NULL if no server controls
             are to be used.

clientctrls  List of client controls, or NULL if no client controls are

Expires: May 2001                                              [Page 43]

C LDAP API        C LDAP Application Program Interface  17 November 2000

to be used.

msgidp       This result parameter will be set to the message id of the
             request if the ldap_extended_operation() call succeeds. The
             value is undefined if a value other than LDAP_SUCCESS is
             returned.

retoidp      Pointer to a character string that will be set to an allo-
             cated, dotted-OID text string returned by the server.  This
             string SHOULD be disposed of using the ldap_memfree() func-
             tion.  If an API error occurs or no OID is returned by the
             server, *retoidp is set to NULL.

retdatap     Pointer to a berval structure pointer that will be set an
             allocated copy of the data returned by the server.  This
             struct berval SHOULD be disposed of using ber_bvfree().  If
             an API error occurs or no data is returned by the server,
             *retdatap is set to NULL.

The ldap_extended_operation() function initiates an asynchronous
extended operation and returns the constant LDAP_SUCCESS if the request
was successfully sent, or another LDAP result code if not.  See the sec-
tion below on error handling for more information about possible errors
and how to interpret them.  If successful, ldap_extended_operation()
places the message id of the request in *msgidp. A subsequent call to
ldap_result(), described below, can be used to obtain the result of the
extended operation which can be passed to ldap_parse_extended_result()
to obtain the OID and data contained in the response.

The synchronous ldap_extended_operation_s() function returns the result
of the operation, either the constant LDAP_SUCCESS if the operation was
successful, or another LDAP result code if it was not.  See the section
below on error handling for more information about possible errors and
how to interpret them.  The retoid and retdata parameters are filled in
with the OID and data from the response.

The ldap_extended_operation() and ldap_extended_operation_s() functions
both support LDAPv3 server controls and client controls.

12.  Abandoning An Operation

The following calls are used to abandon an operation in progress:

int ldap_abandon_ext(
                   LDAP            *ld,
                   int             msgid,
                   LDAPControl     **serverctrls,

Expires: May 2001                                              [Page 44]

C LDAP API        C LDAP Application Program Interface  17 November 2000

LDAPControl     **clientctrls
           );

int ldap_abandon(
                   LDAP            *ld,
                   int             msgid
           );

ld           The session handle.

msgid        The message id of the request to be abandoned.

serverctrls  List of LDAP server controls, or NULL if no server controls
             are to be used.

clientctrls  List of client controls, or NULL if no client controls are
             to be used.

ldap_abandon_ext() abandons the operation with message id msgid and
returns the constant LDAP_SUCCESS if the abandon was successful or
another LDAP result code if not.  See the section below on error han-
dling for more information about possible errors and how to interpret
them.

ldap_abandon() is identical to ldap_abandon_ext() except that it does
not accept client or server controls and it returns zero if the abandon
was successful, -1 otherwise.

After a successful call to ldap_abandon() or ldap_abandon_ext(), results
with the given message id are never returned from a subsequent call to
ldap_result().  There is no server response to LDAP abandon operations.

13.  Obtaining Results and Peeking Inside LDAP Messages

ldap_result() is used to obtain the result of a previous asynchronously
initiated operation. Note that depending on how it is called,
ldap_result() can actually return a list or "chain" of result messages.
The ldap_result() function only returns messages for a single request,
so for all LDAP operations other than search only one result message is
expected; that is, the only time the "result chain" can contain more
than one message is if results from a search operation are returned.
Once a chain of messages has been returned to the caller, it is no
longer tied in any caller-visible way to the LDAP request that produced
it.  However, it MAY be tied to the session handle.  Therefore, a chain
of messages returned by calling ldap_result() or by calling a synchro-
nous search routine will never be affected by subsequent LDAP API calls

Expires: May 2001                                              [Page 45]

C LDAP API        C LDAP Application Program Interface  17 November 2000

except for ldap_msgfree() (which is used to dispose of a chain of mes-
sages) and the unbind calls (which dispose of a session handle):
ldap_unbind(), ldap_unbind_s(), or ldap_unbind_ext(), or functions
defined by extensions of this API.

ldap_msgfree() frees the result messages (possibly an entire chain of
messages) obtained from a previous call to ldap_result() or from a call
to a synchronous search routine.

ldap_msgtype() returns the type of an LDAP message.

ldap_msgid() returns the message ID of an LDAP message.

int ldap_result(
                   LDAP            *ld,
                   int             msgid,
                   int             all,
                   struct timeval  *timeout,
                   LDAPMessage     **res
           );

int ldap_msgfree( LDAPMessage *res );

int ldap_msgtype( LDAPMessage *res );

int ldap_msgid( LDAPMessage *res );

Parameters are:

ld       The session handle.

msgid    The message id of the operation whose results are to be
         returned, the constant LDAP_RES_UNSOLICITED (0) if an unsoli-
         cited result is desired, or or the constant LDAP_RES_ANY (-1)
         if any result is desired.

all      Specifies how many messages will be retrieved in a single call
         to ldap_result().  This parameter only has meaning for search
         results.  Pass the constant LDAP_MSG_ONE (0x00) to retrieve one
         message at a time.  Pass LDAP_MSG_ALL (0x01) to request that
         all results of a search be received before returning all
         results in a single chain.  Pass LDAP_MSG_RECEIVED (0x02) to
         indicate that all messages retrieved so far are to be returned
         in the result chain.

timeout  A timeout specifying how long to wait for results to be
         returned.  A NULL value causes ldap_result() to block until
         results are available.  A timeout value of zero seconds

Expires: May 2001                                              [Page 46]

C LDAP API        C LDAP Application Program Interface  17 November 2000

specifies a polling behavior.

res      For ldap_result(), a result parameter that will contain the
         result(s) of the operation. If an API error occurs or no
         results are returned, *res is set to NULL.  For ldap_msgfree(),
         the result chain to be freed, obtained from a previous call to
         ldap_result(), ldap_search_s(), or ldap_search_st().  If res is
         NULL, nothing is done and ldap_msgfree() returns zero.

Upon successful completion, ldap_result() returns the type of the first
result returned in the res parameter. This will be one of the following
constants.

LDAP_RES_BIND (0x61)
             LDAP_RES_SEARCH_ENTRY (0x64)
             LDAP_RES_SEARCH_REFERENCE (0x73)      -- new in LDAPv3
             LDAP_RES_SEARCH_RESULT (0x65)
             LDAP_RES_MODIFY (0x67)
             LDAP_RES_ADD (0x69)
             LDAP_RES_DELETE (0x6B)
             LDAP_RES_MODDN (0x6D)
             LDAP_RES_COMPARE (0x6F)
             LDAP_RES_EXTENDED (0x78)              -- new in LDAPv3

ldap_result() returns 0 if the timeout expired and -1 if an error
occurs, in which case the error parameters of the LDAP session handle
will be set accordingly.

ldap_msgfree() frees each message in the result chain pointed to by res
and returns the type of the last message in the chain.  If res is NULL,
nothing is done and the value zero is returned.

ldap_msgtype() returns the type of the LDAP message it is passed as a
parameter. The type will be one of the types listed above, or -1 on
error.

ldap_msgid() returns the message ID associated with the LDAP message
passed as a parameter, or -1 on error.

14.  Handling Errors and Parsing Results

The following calls are used to extract information from results and
handle errors returned by other LDAP API routines.  Note that
ldap_parse_sasl_bind_result() and ldap_parse_extended_result() must typ-
ically be used in addition to ldap_parse_result() to retrieve all the
result information from SASL Bind and Extended Operations respectively.

Expires: May 2001                                              [Page 47]

C LDAP API        C LDAP Application Program Interface  17 November 2000

int ldap_parse_result(
                   LDAP            *ld,
                   LDAPMessage     *res,
                   int             *errcodep,
                   char            **matcheddnp,
                   char            **errmsgp,
                   char            ***referralsp,
                   LDAPControl     ***serverctrlsp,
                   int             freeit
           );

int ldap_parse_sasl_bind_result(
                   LDAP            *ld,
                   LDAPMessage     *res,
                   struct berval   **servercredp,
                   int             freeit
           );

int ldap_parse_extended_result(
                   LDAP            *ld,
                   LDAPMessage     *res,
                   char            **retoidp,
                   struct berval   **retdatap,
                   int             freeit
           );

#define LDAP_NOTICE_OF_DISCONNECTION    "1.3.6.1.4.1.1466.20036"

char *ldap_err2string( int err );

The use of the following routines is deprecated and more complete
   descriptions can be found in RFC 1823:

int ldap_result2error(
                   LDAP            *ld,
                   LDAPMessage     *res,
                   int             freeit
           );

void ldap_perror( LDAP *ld, const char *msg );

Parameters are:

ld           The session handle.

res          The result of an LDAP operation as returned by
             ldap_result() or one of the synchronous API operation
             calls.

Expires: May 2001                                              [Page 48]

C LDAP API        C LDAP Application Program Interface  17 November 2000

errcodep     This result parameter will be filled in with the LDAP
             resultCode field from the LDAPMessage message.  This is the
             indication from the server of the outcome of the operation.
             NULL SHOULD be passed to ignore this field.

matcheddnp   If the server returned a matchedDN string to indicate how
             much of a name passed in a request was recognized, this
             result parameter will be filled in with that matchedDN
             string.  Otherwise, this field will be set to NULL.  NULL
             SHOULD be passed to ignore this field.  The matched DN
             string SHOULD be freed by calling ldap_memfree() which is
             described later in this document.  Note that the server may
             return a zero length matchedDN (in which case *matchednp is
             set to an allocated copy of "") which is different than not
             returning a value at all (in which case *matcheddnp is set
             to NULL).

errmsgp      This result parameter will be filled in with the contents
             of the error message field from the LDAPMessage message.
             The error message string SHOULD be freed by calling
             ldap_memfree() which is described later in this document.
             NULL SHOULD be passed to ignore this field.

referralsp   This result parameter will be filled in with the contents
             of the referrals field from the LDAPMessage message, indi-
             cating zero or more alternate LDAP servers where the
             request is to be retried.  The referrals array SHOULD be
             freed by calling ldap_value_free() which is described later
             in this document.  NULL SHOULD be passed to ignore this
             field.  If no referrals were returned, *referralsp is set
             to NULL.

serverctrlsp This result parameter will be filled in with an allocated
             array of controls copied out of the LDAPMessage message.
             If serverctrlsp is NULL, no controls are returned.  The
             control array SHOULD be freed by calling
             ldap_controls_free() which was described earlier.  If no
             controls were returned, *serverctrlsp is set to NULL.

freeit       A boolean that determines whether the res parameter is
             disposed of or not.  Pass any non-zero value to have these
             routines free res after extracting the requested informa-
             tion.  This is provided as a convenience; you can also use
             ldap_msgfree() to free the result later.  If freeit is
             non-zero, the entire chain of messages represented by res
             is disposed of.

servercredp  For SASL bind results, this result parameter will be filled

Expires: May 2001                                              [Page 49]

C LDAP API        C LDAP Application Program Interface  17 November 2000

in with the credentials passed back by the server for
             mutual authentication, if given. An allocated berval struc-
             ture is returned that SHOULD be disposed of by calling
             ber_bvfree().  NULL SHOULD be passed to ignore this field.

retoidp      For extended results, this result parameter will be filled
             in with the dotted-OID text representation of the name of
             the extended operation response.  This string SHOULD be
             disposed of by calling ldap_memfree().  NULL SHOULD be
             passed to ignore this field.  If no OID was returned,
             *retoidp is set to NULL.  The LDAP_NOTICE_OF_DISCONNECTION
             macro is defined as a convenience for clients that wish to
             check an OID to see if it matches the one used for the
             unsolicited Notice of Disconnection (defined in RFC 2251[2]
             section 4.4.1).

retdatap     For extended results, this result parameter will be filled
             in with a pointer to a struct berval containing the data in
             the extended operation response.  It SHOULD be disposed of
             by calling ber_bvfree(). NULL SHOULD be passed to ignore
             this field.  If no data is returned, *retdatap is set to
             NULL.

err          For ldap_err2string(), an LDAP result code, as returned by
             ldap_parse_result() or another LDAP API call.

Additional parameters for the deprecated routines are not described.
Interested readers are referred to RFC 1823.

The ldap_parse_result(), ldap_parse_sasl_bind_result(), and
ldap_parse_extended_result() functions all skip over messages of type
LDAP_RES_SEARCH_ENTRY and LDAP_RES_SEARCH_REFERENCE when looking for a
result message to parse.  They return the constant LDAP_SUCCESS if the
result was successfully parsed and another LDAP API result code if not.
If a value other than LDAP_SUCCESS is returned, the values of all the
result parameters are undefined.  Note that the LDAP result code that
indicates the outcome of the operation performed by the server is placed
in the errcodep ldap_parse_result() parameter.  If a chain of messages
that contains more than one result message is passed to these routines
they always operate on the first result in the chain.

ldap_err2string() is used to convert a numeric LDAP result code, as
returned by ldap_parse_result(), ldap_parse_sasl_bind_result(),
ldap_parse_extended_result() or one of the synchronous API operation
calls, into an informative zero-terminated character string message
describing the error.  It returns a pointer to static data and it MUST
NOT return NULL; the value returned is always a valid null-terminated
"C" string.

Expires: May 2001                                              [Page 50]

C LDAP API        C LDAP Application Program Interface  17 November 2000

15.  Stepping Through a List of Results

The ldap_first_message() and ldap_next_message() routines are used to
step through the list of messages in a result chain returned by
ldap_result().  For search operations, the result chain can actually
include referral messages, entry messages, and result messages.
ldap_count_messages() is used to count the number of messages returned.
The ldap_msgtype() function, described above, can be used to distinguish
between the different message types.

LDAPMessage *ldap_first_message( LDAP *ld, LDAPMessage *res );

LDAPMessage *ldap_next_message( LDAP *ld, LDAPMessage *msg );

int ldap_count_messages( LDAP *ld, LDAPMessage *res );

Parameters are:

ld     The session handle.

res    The result chain, as obtained by a call to one of the synchronous
       search routines or ldap_result().

msg    The message returned by a previous call to ldap_first_message()
       or ldap_next_message().

ldap_first_message() and ldap_next_message() will return NULL when no
more messages exist in the result set to be returned.  NULL is also
returned if an error occurs while stepping through the entries, in which
case the error parameters in the session handle ld will be set to indi-
cate the error.

If successful, ldap_count_messages() returns the number of messages con-
tained in a chain of results; if an error occurs such as the res parame-
ter being invalid, -1 is returned.  The ldap_count_messages() call can
also be used to count the number of messages that remain in a chain if
called with a message, entry, or reference returned by
ldap_first_message(), ldap_next_message(), ldap_first_entry(),
ldap_next_entry(), ldap_first_reference(), ldap_next_reference().

16.  Parsing Search Results

The following calls are used to parse the entries and references
returned by ldap_search() and friends. These results are returned in an
opaque structure that MAY be accessed by calling the routines described
below. Routines are provided to step through the entries and references
returned, step through the attributes of an entry, retrieve the name of

Expires: May 2001                                              [Page 51]

C LDAP API        C LDAP Application Program Interface  17 November 2000

an entry, and retrieve the values associated with a given attribute in
an entry.

16.1.  Stepping Through a List of Entries or References

The ldap_first_entry() and ldap_next_entry() routines are used to step
through and retrieve the list of entries from a search result chain.
The ldap_first_reference() and ldap_next_reference() routines are used
to step through and retrieve the list of continuation references from a
search result chain.  ldap_count_entries() is used to count the number
of entries returned. ldap_count_references() is used to count the number
of references returned.

LDAPMessage *ldap_first_entry( LDAP *ld, LDAPMessage *res );

LDAPMessage *ldap_next_entry( LDAP *ld, LDAPMessage *entry );

LDAPMessage *ldap_first_reference( LDAP *ld, LDAPMessage *res );

LDAPMessage *ldap_next_reference( LDAP *ld, LDAPMessage *ref );

int ldap_count_entries( LDAP *ld, LDAPMessage *res );

int ldap_count_references( LDAP *ld, LDAPMessage *res );

Parameters are:

ld     The session handle.

res    The search result, as obtained by a call to one of the synchro-
       nous search routines or ldap_result().

entry  The entry returned by a previous call to ldap_first_entry() or
       ldap_next_entry().

ref    The reference returned by a previous call to
       ldap_first_reference() or ldap_next_reference().

ldap_first_entry(), ldap_next_entry(), ldap_first_reference() and
ldap_next_reference() all return NULL when no more entries or references
exist in the result set to be returned.  NULL is also returned if an
error occurs while stepping through the entries or references, in which
case the error parameters in the session handle ld will be set to indi-
cate the error.

ldap_count_entries() returns the number of entries contained in a chain
of entries; if an error occurs such as the res parameter being invalid,

Expires: May 2001                                              [Page 52]

C LDAP API        C LDAP Application Program Interface  17 November 2000

-1 is returned.  The ldap_count_entries() call can also be used to count
the number of entries that remain in a chain if called with a message,
entry or reference returned by ldap_first_message(),
ldap_next_message(), ldap_first_entry(), ldap_next_entry(),
ldap_first_reference(), ldap_next_reference().

ldap_count_references() returns the number of references contained in a
chain of search results; if an error occurs such as the res parameter
being invalid, -1 is returned.  The ldap_count_references() call can
also be used to count the number of references that remain in a chain.

16.2.  Stepping Through the Attributes of an Entry

The ldap_first_attribute() and ldap_next_attribute() calls are used to
step through the list of attribute types returned with an entry.

char *ldap_first_attribute(
                   LDAP            *ld,
                   LDAPMessage     *entry,
                   BerElement      **ptr
           );

char *ldap_next_attribute(
                   LDAP            *ld,
                   LDAPMessage     *entry,
                   BerElement      *ptr
           );

void ldap_memfree( char *mem );

Parameters are:

ld     The session handle.

entry  The entry whose attributes are to be stepped through, as returned
       by ldap_first_entry() or ldap_next_entry().

ptr    In ldap_first_attribute(), the address of a pointer used inter-
       nally to keep track of the current position in the entry. In
       ldap_next_attribute(), the pointer returned by a previous call to
       ldap_first_attribute().  The BerElement type itself is an opaque
       structure that is described in more detail later in this document
       in the section "Encoded ASN.1 Value Manipulation".

mem    A pointer to memory allocated by the LDAP library, such as the
       attribute type names returned by ldap_first_attribute() and
       ldap_next_attribute, or the DN returned by ldap_get_dn().  If mem

Expires: May 2001                                              [Page 53]

C LDAP API        C LDAP Application Program Interface  17 November 2000

is NULL, the ldap_memfree() call does nothing.

ldap_first_attribute() and ldap_next_attribute() will return NULL when
the end of the attributes is reached, or if there is an error, in which
case the error parameters in the session handle ld will be set to indi-
cate the error.

Both routines return a pointer to an allocated buffer containing the
current attribute name. This SHOULD be freed when no longer in use by
calling ldap_memfree().

ldap_first_attribute() will allocate and return in ptr a pointer to a
BerElement used to keep track of the current position. This pointer MAY
be passed in subsequent calls to ldap_next_attribute() to step through
the entry's attributes. After a set of calls to ldap_first_attribute()
and ldap_next_attribute(), if ptr is non-NULL, it SHOULD be freed by
calling ber_free( ptr, 0 ). Note that it is very important to pass the
second parameter as 0 (zero) in this call, since the buffer associated
with the BerElement does not point to separately allocated memory.

The attribute type names returned are suitable for passing in a call to
ldap_get_values() and friends to retrieve the associated values.

16.3.  Retrieving the Values of an Attribute

ldap_get_values() and ldap_get_values_len() are used to retrieve the
values of a given attribute from an entry. ldap_count_values() and
ldap_count_values_len() are used to count the returned values.
ldap_value_free() and ldap_value_free_len() are used to free the values.

char **ldap_get_values(
                   LDAP            *ld,
                   LDAPMessage     *entry,
                   const char      *attr
           );

struct berval **ldap_get_values_len(
                   LDAP            *ld,
                   LDAPMessage     *entry,
                   const char      *attr
           );

int ldap_count_values( char **vals );

int ldap_count_values_len( struct berval **vals );

void ldap_value_free( char **vals );

Expires: May 2001                                              [Page 54]

C LDAP API        C LDAP Application Program Interface  17 November 2000

void ldap_value_free_len( struct berval **vals );

Parameters are:

ld     The session handle.

entry  The entry from which to retrieve values, as returned by
       ldap_first_entry() or ldap_next_entry().

attr   The attribute whose values are to be retrieved, as returned by
       ldap_first_attribute() or ldap_next_attribute(), or a caller-
       supplied string (e.g., "mail").

vals   The values returned by a previous call to ldap_get_values() or
       ldap_get_values_len().

Two forms of the various calls are provided. The first form is only
suitable for use with non-binary character string data. The second _len
form is used with any kind of data.

ldap_get_values() and ldap_get_values_len() return NULL if no values are
found for attr or if an error occurs.

ldap_count_values() and ldap_count_values_len() return -1 if an error
occurs such as the vals parameter being invalid.

If a NULL vals parameter is passed to ldap_value_free() or
ldap_value_free_len(), nothing is done.

Note that the values returned are dynamically allocated and SHOULD be
freed by calling either ldap_value_free() or ldap_value_free_len() when
no longer in use.

16.4.  Retrieving the name of an entry

ldap_get_dn() is used to retrieve the name of an entry.
ldap_explode_dn() and ldap_explode_rdn() are used to break up a name
into its component parts. ldap_dn2ufn() is used to convert the name into
a more "user friendly" format.

char *ldap_get_dn( LDAP *ld, LDAPMessage *entry );

char **ldap_explode_dn( const char *dn, int notypes );

char **ldap_explode_rdn( const char *rdn, int notypes );

char *ldap_dn2ufn( const char *dn );

Expires: May 2001                                              [Page 55]

C LDAP API        C LDAP Application Program Interface  17 November 2000

Parameters are:

ld      The session handle.

entry   The entry whose name is to be retrieved, as returned by
        ldap_first_entry() or ldap_next_entry().

dn      The dn to explode, such as returned by ldap_get_dn().  If NULL,
        a zero length DN is used.

rdn     The rdn to explode, such as returned in the components of the
        array returned by ldap_explode_dn().  If NULL, a zero length DN
        is used.

notypes A boolean parameter, if non-zero indicating that the dn or rdn
        components are to have their type information stripped off
        (i.e., "cn=Babs" would become "Babs").

ldap_get_dn() will return NULL if there is some error parsing the dn,
setting error parameters in the session handle ld to indicate the error.
It returns a pointer to newly allocated space that the caller SHOULD
free by calling ldap_memfree() when it is no longer in use.  Note the
format of the DNs returned is given by [5].  The root DN is returned as
a zero length string ("").

ldap_explode_dn() returns a NULL-terminated char * array containing the
RDN components of the DN supplied, with or without types as indicated by
the notypes parameter. The components are returned in the order they
appear in the dn.  The array returned SHOULD be freed when it is no
longer in use by calling ldap_value_free().

ldap_explode_rdn() returns a NULL-terminated char * array containing the
components of the RDN supplied, with or without types as indicated by
the notypes parameter. The components are returned in the order they
appear in the rdn.  The array returned SHOULD be freed when it is no
longer in use by calling ldap_value_free().

ldap_dn2ufn() converts the DN into the user friendly format described in
[14]. The UFN returned is newly allocated space that SHOULD be freed by
a call to ldap_memfree() when no longer in use.

16.5.  Retrieving controls from an entry

ldap_get_entry_controls() is used to extract LDAP controls from an
entry.

Expires: May 2001                                              [Page 56]

C LDAP API        C LDAP Application Program Interface  17 November 2000

int ldap_get_entry_controls(
                   LDAP            *ld,
                   LDAPMessage     *entry,
                   LDAPControl     ***serverctrlsp
           );

Parameters are:

ld           The session handle.

entry        The entry to extract controls from, as returned by
             ldap_first_entry() or ldap_next_entry().

serverctrlsp This result parameter will be filled in with an allocated
             array of controls copied out of entry. The control array
             SHOULD be freed by calling ldap_controls_free().  If ser-
             verctrlsp is NULL, no controls are returned.  If no con-
             trols were returned, *serverctrlsp is set to NULL.

ldap_get_entry_controls() returns an LDAP result code that indicates
whether the reference could be successfully parsed (LDAP_SUCCESS if all
goes well). If ldap_get_entry_controls() returns a value other than
LDAP_SUCCESS, the value of the serverctrlsp output parameter is unde-
fined.

16.6.  Parsing References

ldap_parse_reference() is used to extract referrals and controls from a
SearchResultReference message.

int ldap_parse_reference(
                   LDAP            *ld,
                   LDAPMessage     *ref,
                   char            ***referralsp,
                   LDAPControl     ***serverctrlsp,
                   int             freeit
           );

Parameters are:

ld           The session handle.

ref          The reference to parse, as returned by ldap_result(),
             ldap_first_reference(), or ldap_next_reference().

Expires: May 2001                                              [Page 57]

C LDAP API        C LDAP Application Program Interface  17 November 2000

referralsp   This result parameter will be filled in with an allocated
             array of character strings.  The elements of the array are
             the referrals (typically LDAP URLs) contained in ref.  The
             array SHOULD be freed when no longer in used by calling
             ldap_value_free().  If referralsp is NULL, the referral
             URLs are not returned.  If no referrals were returned,
             *referralsp is set to NULL.

serverctrlsp This result parameter will be filled in with an allocated
             array of controls copied out of ref. The control array
             SHOULD be freed by calling ldap_controls_free().  If ser-
             verctrlsp is NULL, no controls are returned.  If no con-
             trols were returned, *serverctrlsp is set to NULL.

freeit       A boolean that determines whether the ref parameter is
             disposed of or not.  Pass any non-zero value to have this
             routine free ref after extracting the requested informa-
             tion.  This is provided as a convenience; you can also use
             ldap_msgfree() to free the result later.

ldap_parse_reference() returns an LDAP result code that indicates
whether the reference could be successfully parsed (LDAP_SUCCESS if all
goes well).  If a value other than LDAP_SUCCESS is returned, the value
of the referralsp and serverctrlsp result parameters are undefined.

17.  Encoded ASN.1 Value Manipulation

This section describes routines which MAY be used to encode and decode
BER-encoded ASN.1 values, which are often used inside of control and
extension values.

With the exceptions of two new functions ber_flatten() and ber_init(),
these functions are compatible with the University of Michigan LDAP 3.3
implementation of BER.

Note that the functions defined in this section all provide a method for
determining success or failure but generally do not provide access to
specific error codes.  Therefore, applications that require precise
error information when encoding or decoding ASN.1 values SHOULD NOT use
these functions.

17.1.  BER Data Structures and Types

The following additional integral types are defined for use in manipula-
tion of BER encoded ASN.1 values:

Expires: May 2001                                              [Page 58]

C LDAP API        C LDAP Application Program Interface  17 November 2000

typedef <impl_tag_t> ber_tag_t;     /* for BER tags */

typedef <impl_int_t> ber_int_t;     /* for BER ints, enums, and Booleans */

typedef <impl_unit_t> ber_uint_t;   /* unsigned equivalent of ber_uint_t */

typedef <impl_slen_t> ber_slen_t;   /* signed equivalent of ber_len_t */

Note that the actual definition for these four integral types is imple-
mentation specific; that is, `<impl_tag_t>', `<impl_int_t>',
`<impl_uint_t>', and `<impl_slen_t>' MUST each be replaced with an
appropriate implementation-specific type.

The `ber_tag_t' type is an unsigned integral data type that is large
enough to hold the largest BER tag supported by the API implementation.
The width (number of significant bits) of `ber_tag_t' MUST be at least
32, greater than or equal to that of `unsigned int' (so that integer
promotions won't promote it to `int'), and no wider than that of
`unsigned long'.

The `ber_int_t' and `ber_uint_t' types are the signed and unsigned vari-
ants of an integral type that is large enough to hold integers for pur-
poses of BER encoding and decoding.  The width of `ber_int_t' MUST be at
least 32 and no larger than that of `long'.

The `ber_slen_t' type is the signed variant of the `ber_len_t' integral
type, i.e. if `ber_len_t' is unsigned long, then `ber_slen_t' is signed
long.  The `<impl_slen_t>' in the `ber_len_t' typedef MUST be replaced
with an appropriate type.  Note that `ber_slen_t' is not used directly
in the C LDAP API but is provided for the convenience of application
developers and for use by extensions to the API.

typedef struct berval {
                   ber_len_t       bv_len;
                   char            *bv_val;
           } BerValue;

As defined earlier in the section "Common Data Structures", a berval
structure contains an arbitrary sequence of bytes and an indication of
its length.  The bv_len element is an unsigned integer.  The bv_val is
not necessarily zero-terminated.  Applications MAY allocate their own
berval structures.

As defined earlier in the section "Common Data Structures", the BerEle-
ment structure is an opaque structure:

typedef struct berelement BerElement;

Expires: May 2001                                              [Page 59]

C LDAP API        C LDAP Application Program Interface  17 November 2000

It contains not only a copy of the encoded value, but also state infor-
mation used in encoding or decoding.  Applications cannot allocate their
own BerElement structures.  The internal state is neither thread-
specific nor locked, so two threads SHOULD NOT manipulate the same
BerElement value simultaneously.

A single BerElement value cannot be used for both encoding and decoding.

17.2.  Memory Disposal and Utility Functions

void ber_bvfree( struct berval *bv );

ber_bvfree() frees a berval structure returned from this API.  Both the
bv->bv_val string and the berval structure itself are freed.  If bv is
NULL, this call does nothing.

void ber_bvecfree( struct berval **bv );

ber_bvecfree() frees an array of berval structures returned from this
API.  Each of the berval structures in the array are freed using
ber_bvfree(), then the array itself is freed.  If bv is NULL, this call
does nothing.

struct berval *ber_bvdup( const struct berval *bv );

ber_bvdup() returns a copy of a berval structure.  The bv_val field in
the returned berval structure points to a different area of memory than
the bv_val field in the bv argument.  The NULL pointer is returned on
error (e.g. out of memory).

void ber_free( BerElement *ber, int fbuf );

ber_free() frees a BerElement which is returned from the API calls
ber_alloc_t() or ber_init().  Each BerElement SHOULD be freed by the
caller.  The second argument fbuf SHOULD always be set to 1 to ensure
that the internal buffer used by the BER functions is freed as well as
the BerElement container itself.  If ber is NULL, this call does noth-
ing.

17.3.  Encoding

BerElement *ber_alloc_t( int options );

ber_alloc_t() constructs and returns BerElement.  The NULL pointer is
returned on error.  The options field contains a bitwise-or of options
which are to be used when generating the encoding of this BerElement.
One option is defined and SHOULD always be supplied:

Expires: May 2001                                              [Page 60]

C LDAP API        C LDAP Application Program Interface  17 November 2000

#define LBER_USE_DER 0x01

When this option is present, lengths will always be encoded in the
minimum number of octets.  Note that this option does not cause values
of sets to be rearranged in tag and byte order or default values to be
removed, so these functions are not sufficient for generating DER output
as defined in X.509 and X.680.  If the caller takes responsibility for
ordering values of sets correctly and removing default values, DER out-
put as defined in X.509 and X.680 can be produced.

Unrecognized option bits are ignored.

The BerElement returned by ber_alloc_t() is initially empty.  Calls to
ber_printf() will append bytes to the end of the ber_alloc_t().

int ber_printf( BerElement *ber, const char *fmt, ... );

The ber_printf() routine is used to encode a BER element in much the
same way that sprintf() works.  One important difference, though, is
that state information is kept in the ber argument so that multiple
calls can be made to ber_printf() to append to the end of the BER ele-
ment. ber MUST be a pointer to a BerElement returned by ber_alloc_t().
ber_printf() interprets and formats its arguments according to the for-
mat string fmt.  ber_printf() returns -1 if there is an error during
encoding and a non-negative number if successful.  As with sprintf(),
each character in fmt refers to an argument to ber_printf().

The format string can contain the following format characters:

't'     Tag.  The next argument is a ber_tag_t specifying the tag to
        override the next element to be written to the ber.  This works
        across calls.  The integer tag value SHOULD contain the tag
        class, constructed bit, and tag value.  For example, a tag of
        "[3]" for a constructed type is 0xA3U.  All implementations MUST
        support tags that fit in a single octet (i.e., where the tag
        value is less than 32) and they MAY support larger tags.

'b'     Boolean.  The next argument is an ber_int_t, containing either 0
        for FALSE or 0xff for TRUE.  A boolean element is output.  If
        this format character is not preceded by the 't' format modif-
        ier, the tag 0x01U is used for the element.

'e'     Enumerated.  The next argument is a ber_int_t, containing the
        enumerated value in the host's byte order.  An enumerated ele-
        ment is output.  If this format character is not preceded by the
        't' format modifier, the tag 0x0AU is used for the element.

'i'     Integer.  The next argument is a ber_int_t, containing the

Expires: May 2001                                              [Page 61]

C LDAP API        C LDAP Application Program Interface  17 November 2000

integer in the host's byte order.  An integer element is output.
        If this format character is not preceded by the 't' format
        modifier, the tag 0x02U is used for the element.

'B'     Bitstring.  The next two arguments are a char * pointer to the
        start of the bitstring, followed by a ber_len_t containing the
        number of bits in the bitstring.  A bitstring element is output,
        in primitive form.  If this format character is not preceded by
        the 't' format modifier, the tag 0x03U is used for the element.

'X'     Reserved and not to be used.  In older revisions of this specif-
        ication,

'n'     Null.  No argument is needed.  An ASN.1 NULL element is output.
        If this format character is not preceded by the 't' format
        modifier, the tag 0x05U is used for the element.

'o'     Octet string.  The next two arguments are a char *, followed by
        a ber_len_t with the length of the string.  The string MAY con-
        tain null bytes and are do not have to be zero-terminated.   An
        octet string element is output, in primitive form.  If this for-
        mat character is not preceded by the 't' format modifier, the
        tag 0x04U is used for the element.

's'     Octet string.  The next argument is a char * pointing to a
        zero-terminated string.  An octet string element in primitive
        form is output, which does not include the trailing '\0' (null)
        byte. If this format character is not preceded by the 't' format
        modifier, the tag 0x04U is used for the element.

'v'     Several octet strings.  The next argument is a char **, an array
        of char * pointers to zero-terminated strings.  The last element
        in the array MUST be a NULL pointer. The octet strings do not
        include the trailing '\0' (null) byte.  Note that a construct
        like '{v}' is used to get an actual SEQUENCE OF octet strings.
        The 't' format modifier cannot be used with this format charac-
        ter.

'V'     Several octet strings.  A NULL-terminated array of struct berval
        *'s is supplied.  Note that a construct like '{V}' is used to
        get an actual SEQUENCE OF octet strings. The 't' format modifier
        cannot be used with this format character.

'{'     Begin sequence.  No argument is needed.  If this format charac-
        ter is not preceded by the 't' format modifier, the tag 0x30U is
        used.

'}'     End sequence.  No argument is needed.  The 't' format modifier

Expires: May 2001                                              [Page 62]

C LDAP API        C LDAP Application Program Interface  17 November 2000

cannot be used with this format character.

'['     Begin set.  No argument is needed.  If this format character is
        not preceded by the 't' format modifier, the tag 0x31U is used.

']'     End set.  No argument is needed.  The 't' format modifier cannot
        be used with this format character.

Each use of a '{' format character SHOULD be matched by a '}' character,
either later in the format string, or in the format string of a subse-
quent call to ber_printf() for that BerElement.  The same applies to the
'[' and ']' format characters.

Sequences and sets nest, and implementations of this API MUST maintain
internal state to be able to properly calculate the lengths.

int ber_flatten( BerElement *ber, struct berval **bvPtr );

The ber_flatten routine allocates a struct berval whose contents are a
BER encoding taken from the ber argument. The bvPtr pointer points to
the returned berval structure, which SHOULD be freed using ber_bvfree().
This routine returns 0 on success and -1 on error.

The ber_flatten API call is not present in U-M LDAP 3.3.

The use of ber_flatten on a BerElement in which all '{' and '}' format
modifiers have not been properly matched is an error (i.e., -1 will be
returned by ber_flatten() if this situation is exists).

17.4.  Encoding Example

The following is an example of encoding the following ASN.1 data type:

Example1Request ::= SEQUENCE {
           s     OCTET STRING, -- must be printable
           val1  INTEGER,
           val2  [0] INTEGER DEFAULT 0
      }

int encode_example1(const char *s, ber_int_t val1, ber_int_t val2,
               struct berval **bvPtr)
      {
           BerElement *ber;
           int rc = -1;

*bvPtr = NULL;  /* in case of error */

Expires: May 2001                                              [Page 63]

C LDAP API        C LDAP Application Program Interface  17 November 2000

ber = ber_alloc_t(LBER_USE_DER);

if (ber == NULL) return -1;

if (ber_printf(ber,"{si",s,val1) == -1) {
                   goto done;
           }

if (val2 != 0) {
                   if (ber_printf(ber,"ti",(ber_tag_t)0x80,val2) == -1) {
                           goto done;
                   }
           }

if (ber_printf(ber,"}") == -1) {
                   goto done;
           }

rc = ber_flatten(ber,bvPtr);

done:
           ber_free(ber,1);
           return rc;
      }

17.5.  Decoding

The following two macros are available to applications: LBER_ERROR and
LBER_DEFAULT.  Both of these macros MUST be #define'd as ber_tag_t
integral values that are treated as invalid tags by the API implementa-
tion.  It is RECOMMENDED that the values of LBER_ERROR and LBER_DEFAULT
be the same and that they be defined as values where all octets have the
value 0xFF.  ISO C guarantees that these definitions will work:

#define LBER_ERROR   ((ber_tag_t)-1)
           #define LBER_DEFAULT ((ber_tag_t)-1)

The intent is that LBER_ERROR and LBER_DEFAULT are both defined as the
integer value that has all octets set to 0xFF, as such a value is not a
valid BER tag.

BerElement *ber_init( const struct berval *bv );

The ber_init function constructs a BerElement and returns a new BerEle-
ment containing a copy of the data in the bv argument.  ber_init returns
the NULL pointer on error.

Expires: May 2001                                              [Page 64]

C LDAP API        C LDAP Application Program Interface  17 November 2000

ber_tag_t ber_scanf( BerElement *ber, const char *fmt, ... );

The ber_scanf() routine is used to decode a BER element in much the same
way that sscanf() works.  One important difference, though, is that some
state information is kept with the ber argument so that multiple calls
can be made to ber_scanf() to sequentially read from the BER element.
The ber argument SHOULD be a pointer to a BerElement returned by
ber_init().  ber_scanf interprets the bytes according to the format
string fmt, and stores the results in its additional arguments.
ber_scanf() returns LBER_ERROR on error, and a different value on suc-
cess.  If an error occurred, the values of all the result parameters are
undefined.

The format string contains conversion specifications which are used to
direct the interpretation of the BER element.  The format string can
contain the following characters:

'a'     Octet string.  A char ** argument MUST be supplied.  Memory is
        allocated, filled with the contents of the octet string, zero-
        terminated, and the pointer to the string is stored in the argu-
        ment.  The returned value SHOULD be freed using ldap_memfree.
        The tag of the element MUST indicate the primitive form (con-
        structed strings are not supported) but is otherwise ignored and
        discarded during the decoding.  This format cannot be used with
        octet strings which could contain null bytes.

'O'     Octet string.  A struct berval ** argument MUST be supplied,
        which upon return points to an allocated struct berval contain-
        ing the octet string and its length.  ber_bvfree() SHOULD be
        called to free the allocated memory.  The tag of the element
        MUST indicate the primitive form (constructed strings are not
        supported) but is otherwise ignored during the decoding.

'b'     Boolean.  A pointer to a ber_int_t MUST be supplied. The
        ber_int_t value stored will be 0 for FALSE or nonzero for TRUE.
        The tag of the element MUST indicate the primitive form but is
        otherwise ignored during the decoding.

'e'     Enumerated.  A pointer to a ber_int_t MUST be supplied. The
        enumerated value stored will be in host byte order.  The tag of
        the element MUST indicate the primitive form but is otherwise
        ignored during the decoding.  ber_scanf() will return an error
        if the value of the enumerated value cannot be stored in a
        ber_int_t.

'i'     Integer.  A pointer to a ber_int_t MUST be supplied. The
        ber_int_t value stored will be in host byte order.  The tag of
        the element MUST indicate the primitive form but is otherwise

Expires: May 2001                                              [Page 65]

C LDAP API        C LDAP Application Program Interface  17 November 2000

ignored during the decoding.  ber_scanf() will return an error
        if the integer cannot be stored in a ber_int_t.

'B'     Bitstring.  A char ** argument MUST be supplied which will point
        to the allocated bits, followed by a ber_len_t * argument, which
        will point to the length (in bits) of the bitstring returned.
        ldap_memfree SHOULD be called to free the bitstring.  The tag of
        the element MUST indicate the primitive form (constructed bit-
        strings are not supported) but is otherwise ignored during the
        decoding.

'n'     Null.  No argument is needed.  The element is verified to have a
        zero-length value and is skipped.  The tag is ignored.

't'     Tag.  A pointer to a ber_tag_t MUST be supplied.  The ber_tag_t
        value stored will be the tag of the next element in the BerEle-
        ment ber, represented so it can be written using the 't' format
        of ber_printf().  The decoding position within the ber argument
        is unchanged by this; that is, the fact that the tag has been
        retrieved does not affect future use of ber.

'v'     Several octet strings.  A char *** argument MUST be supplied,
        which upon return points to an allocated NULL-terminated array
        of char *'s containing the octet strings.  NULL is stored if the
        sequence is empty.  ldap_memfree SHOULD be called to free each
        element of the array and the array itself.  The tag of the
        sequence and of the octet strings are ignored.

'V'     Several octet strings (which could contain null bytes).  A
        struct berval *** MUST be supplied, which upon return points to
        a allocated NULL-terminated array of struct berval *'s contain-
        ing the octet strings and their lengths.  NULL is stored if the
        sequence is empty. ber_bvecfree() can be called to free the
        allocated memory.  The tag of the sequence and of the octet
        strings are ignored.

'x'     Skip element.  The next element is skipped.  No argument is
        needed.

'{'     Begin sequence.  No argument is needed.  The initial sequence
        tag and length are skipped.

'}'     End sequence.  No argument is needed.

'['     Begin set.  No argument is needed.  The initial set tag and
        length are skipped.

']'     End set.  No argument is needed.

Expires: May 2001                                              [Page 66]

C LDAP API        C LDAP Application Program Interface  17 November 2000

ber_tag_t ber_peek_tag( BerElement *ber,
                   ber_len_t *lenPtr );

ber_peek_tag() returns the tag of the next element to be parsed in the
BerElement argument.  The length of this element is stored in the
*lenPtr argument.  LBER_DEFAULT is returned if there is no further data
to be read.  The decoding position within the ber argument is unchanged
by this call; that is, the fact that ber_peek_tag() has been called does
not affect future use of ber.

ber_tag_t ber_skip_tag( BerElement *ber, ber_len_t *lenPtr );

ber_skip_tag() is similar to ber_peek_tag(), except that the state
pointer in the BerElement argument is advanced past the first tag and
length, and is pointed to the value part of the next element.  This rou-
tine SHOULD only be used with constructed types and situations when a
BER encoding is used as the value of an OCTET STRING.  The length of the
value is stored in *lenPtr.

ber_tag_t ber_first_element( BerElement *ber,
                   ber_len_t *lenPtr, char **opaquePtr );

ber_tag_t ber_next_element( BerElement *ber,
                   ber_len_t *lenPtr, char *opaque );

ber_first_element() and ber_next_element() are used to traverse a SET,
SET OF, SEQUENCE or SEQUENCE OF data value. ber_first_element() calls
ber_skip_tag(), stores internal information in *lenPtr and *opaquePtr,
and calls ber_peek_tag() for the first element inside the constructed
value. LBER_DEFAULT is returned if the constructed value is empty.
ber_next_element() positions the state at the start of the next element
in the constructed type.  LBER_DEFAULT is returned if there are no
further values.

The len and opaque values SHOULD NOT be used by applications other than
as arguments to ber_next_element(), as shown in the example below.

17.6.  Decoding Example

The following is an example of decoding an ASN.1 data type:

Example2Request ::= SEQUENCE {
           dn    OCTET STRING, -- must be printable
           scope ENUMERATED { b (0), s (1), w (2) },
           ali   ENUMERATED { n (0), s (1), f (2), a (3) },
           size  INTEGER,
           time  INTEGER,

Expires: May 2001                                              [Page 67]

C LDAP API        C LDAP Application Program Interface  17 November 2000

tonly BOOLEAN,
           attrs SEQUENCE OF OCTET STRING, -- must be printable
           [0] SEQUENCE OF SEQUENCE {
              type  OCTET STRING -- must be printable,
              crit  BOOLEAN DEFAULT FALSE,
              value OCTET STRING
           } OPTIONAL }

#define TAG_CONTROL_LIST 0xA0U /* context specific cons 0 */

int decode_example2(struct berval *bv)
      {
           BerElement *ber;
           ber_len_t len;
           ber_tag_t res;
           ber_int_t scope, ali, size, time, tonly;
           char *dn = NULL, **attrs = NULL;
           int i,rc = 0;

ber = ber_init(bv);
           if (ber == NULL) {
                   fputs("ERROR ber_init failed\n", stderr);
                   return -1;
           }

res = ber_scanf(ber,"{aiiiib{v}",&dn,&scope,&ali,
                           &size,&time,&tonly,&attrs);

if (res == LBER_ERROR) {
                   fputs("ERROR ber_scanf failed\n", stderr);
                   ber_free(ber,1);
                   return -1;
           }

/* *** use dn */
           ldap_memfree(dn);

for (i = 0; attrs != NULL && attrs[i] != NULL; i++) {
                   /* *** use attrs[i] */
                   ldap_memfree(attrs[i]);
           }
           ldap_memfree((char *)attrs);

if (ber_peek_tag(ber,&len) == TAG_CONTROL_LIST) {
                   char *opaque;
                   ber_tag_t tag;

for (tag = ber_first_element(ber,&len,&opaque);

Expires: May 2001                                              [Page 68]

C LDAP API        C LDAP Application Program Interface  17 November 2000

tag != LBER_DEFAULT;
                        tag = ber_next_element (ber,&len,opaque)) {

ber_len_t tlen;
                           ber_tag_t ttag;
                           char *type;
                           ber_int_t crit;
                           struct berval *value;

if (ber_scanf(ber,"{a",&type) == LBER_ERROR) {
                                   fputs("ERROR cannot parse type\n", stderr);
                                   break;
                           }
                           /* *** use type */
                           ldap_memfree(type);

ttag = ber_peek_tag(ber,&tlen);
                           if (ttag == 0x01U) {  /* boolean */
                                   if (ber_scanf(ber,"b",
                                                 &crit) == LBER_ERROR) {
                                           fputs("ERROR cannot parse crit\n",
                                               stderr);
                                           rc = -1;
                                           break;
                                   }
                           } else if (ttag == 0x04U) { /* octet string */
                                   crit = 0;
                           } else {
                                   fputs("ERROR extra field in controls\n",
                                       stderr );
                                   break;
                           }

if (ber_scanf(ber,"O}",&value) == LBER_ERROR) {
                                   fputs("ERROR cannot parse value\n", stderr);
                                   rc = -1;
                                   break;
                           }
                           /* *** use value */
                           ber_bvfree(value);
                   }
           }

if ( rc == 0 ) {        /* no errors so far */
                   if (ber_scanf(ber,"}") == LBER_ERROR) {
                           rc = -1;
                   }
           }

Expires: May 2001                                              [Page 69]

C LDAP API        C LDAP Application Program Interface  17 November 2000

ber_free(ber,1);

return rc;
       }

18.  Security Considerations

LDAPv2 supports security through protocol-level authentication using
clear-text passwords.  LDAPv3 adds support for SASL [12] (Simple Authen-
tication Security Layer) methods.  LDAPv3 also supports operation over a
secure transport layer using Transport Layer Security TLS [9].  Readers
are referred to the protocol documents for discussion of related secu-
rity considerations.

Implementations of this API SHOULD be cautious when handling authentica-
tion credentials.  In particular, keeping long-lived copies of creden-
tials without the application's knowledge is discouraged.

19.  Acknowledgements

Many members of the IETF ASID and LDAPEXT working groups as well as
members of the Internet at large have provided useful comments and
suggestions that have been incorporated into this document.  Chris
Weider deserves special mention for his contributions as co-author of
earlier revisions of this document.

The original material upon which this specification is based was sup-
ported by the National Science Foundation under Grant No.  NCR-9416667.

20.  Copyright

This document and translations of it may be copied and furnished to oth-
ers, and derivative works that comment on or otherwise explain it or
assist in its implementation may be prepared, copied, published and dis-
tributed, in whole or in part, without restriction of any kind, provided
that the above copyright notice and this paragraph are included on all
such copies and derivative works.  However, this document itself may not
be modified in any way, such as by removing the copyright notice or
references to the Internet Society or other Internet organizations,
except as needed for the  purpose of developing Internet standards in
which case the procedures for copyrights defined in the Internet Stan-
dards process must be followed, or as required to translate it into

Expires: May 2001                                              [Page 70]

C LDAP API        C LDAP Application Program Interface  17 November 2000

languages other than English.

The limited permissions granted above are perpetual and will not be
revoked by the Internet Society or its successors or assigns.

This document and the information contained herein is provided on an "AS
IS" basis and THE INTERNET SOCIETY AND THE INTERNET ENGINEERING TASK
FORCE DISCLAIMS ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDING BUT NOT
LIMITED TO ANY WARRANTY THAT THE USE OF THE INFORMATION HEREIN WILL NOT
INFRINGE ANY RIGHTS OR ANY IMPLIED WARRANTIES OF MERCHANTABILITY OR FIT-
NESS FOR A PARTICULAR PURPOSE.

21.  Bibliography

[1]  Bradner, S., "Key words for use in RFCs to Indicate Requirement
     Levels", RFC 2119, March 1997.

[2]  M. Wahl, T. Howes, S. Kille, "Lightweight Directory Access Protocol
     (v3)", RFC 2251, December 1997.

[3]  M. Wahl, A. Coulbeck, T. Howes, S. Kille, W. Yeong, C. Robbins,
     "Lightweight Directory Access Protocol (v3): Attribute Syntax
     Definitions", RFC 2252, December 1997.

[4]  The Directory: Selected Attribute Syntaxes.  CCITT, Recommendation
     X.520.

[5]  M. Wahl, S. Kille, T. Howes, "Lightweight Directory Access Protocol
     (v3): A UTF-8 String Representation of Distinguished Names", RFC
     2253, December 1997.

[6]  F. Yergeau, "UTF-8, a transformation format of Unicode and ISO
     10646", RFC 2044, October 1996.

[7]  K. Simonsen, "Character Mnemonics and Character Sets," RFC 1345,
     June 1992.

[8]  "Programming Languages - C", ANSI/ISO Standard 9899, revised 1997.

[9]  J. Hodges, R. Morgan, M. Wahl, "Lightweight Directory Access Proto-
     col (v3):  Extension for Transport Layer Security", INTERNET-DRAFT
     (work in progress) <draft-ietf-ldapext-ldapv3-tls-05.txt>, June
     1999.

[10] R. Hinden, S. Deering, "IP Version 6 Addressing Architecture," RFC
     1884, December 1995.

Expires: May 2001                                              [Page 71]

C LDAP API        C LDAP Application Program Interface  17 November 2000

[11] A. Herron, T. Howes, M. Wahl, A. Anantha, "LDAP Control Extension
     for Server Side Sorting of Search Results", INTERNET-DRAFT (work in
     progress) <draft-ietf-ldapext-sorting-02.txt>, 5 April 1999.

[12] J. Meyers, "Simple Authentication and Security Layer (SASL)", RFC
     2222, October 1997.

[13] T. Howes, "The String Representation of LDAP Search Filters," RFC
     2254, December 1997.

[14] S. Kille, "Using the OSI Directory to Achieve User Friendly Nam-
     ing," RFC 1781, March 1995.

22.  Authors' Addresses

Mark Smith (document editor)
   Netscape Communications Corp.
   901 San Antonio Rd.
   Palo Alto, CA  94303-4900
   Mail Stop SCA17 - 201
   USA
   +1 650 937-3477
   mcs@netscape.com

Tim Howes
   Loudcloud, Inc.
   599 N. Mathilda Avenue
   Sunnyvale, CA 94085
   USA
   +1 408 744-7300
   howes@loudcloud.com

Andy Herron
   Microsoft Corp.
   1 Microsoft Way
   Redmond, WA 98052
   USA
   +1 425 882-8080
   andyhe@microsoft.com

Mark Wahl
   Sun Microsystems, Inc.
   8911 Capital of Texas Hwy, Suite 4140
   Austin, TX 78759
   USA
   +1 626 919 3600
   Mark.Wahl@sun.com

Expires: May 2001                                              [Page 72]

C LDAP API        C LDAP Application Program Interface  17 November 2000

Anoop Anantha
   Microsoft Corp.
   1 Microsoft Way
   Redmond, WA 98052
   USA
   +1 425 882-8080
   anoopa@microsoft.com

23.  Appendix A - Sample C LDAP API Code

#include <stdio.h>
   #include <ldap.h>

main()
   {
           LDAP            *ld;
           LDAPMessage     *res, *e;
           int             i, rc;
           char            *a, *dn;
           BerElement      *ptr;
           char            **vals;

/* open an LDAP session */
           if ( (ld = ldap_init( "dotted.host.name", LDAP_PORT )) == NULL )
                   return 1;

/* authenticate as nobody */
           if (( rc = ldap_simple_bind_s( ld, NULL, NULL )) != LDAP_SUCCESS ) {
                   fprintf( stderr, "ldap_simple_bind_s: %s\n",
                       ldap_err2string( rc ));
                   ldap_unbind( ld );
                   return 1;
           }

/* search for entries with cn of "Babs Jensen", return all attrs  */
           if (( rc = ldap_search_s( ld, "o=University of Michigan, c=US",
               LDAP_SCOPE_SUBTREE, "(cn=Babs Jensen)", NULL, 0, &res ))
               != LDAP_SUCCESS ) {
                   fprintf( stderr, "ldap_search_s: %s\n",
                       ldap_err2string( rc ));
                   if ( res == NULL ) {
                           ldap_unbind( ld );
                           return 1;
                   }
           }

/* step through each entry returned */

Expires: May 2001                                              [Page 73]

C LDAP API        C LDAP Application Program Interface  17 November 2000

for ( e = ldap_first_entry( ld, res ); e != NULL;
               e = ldap_next_entry( ld, e ) ) {
                   /* print its name */
                   dn = ldap_get_dn( ld, e );
                   printf( "dn: %s\n", dn );
                   ldap_memfree( dn );

/* print each attribute */
                   for ( a = ldap_first_attribute( ld, e, &ptr ); a != NULL;
                       a = ldap_next_attribute( ld, e, ptr ) ) {
                           printf( "\tattribute: %s\n", a );

/* print each value */
                           vals = ldap_get_values( ld, e, a );
                           for ( i = 0; vals[i] != NULL; i++ ) {
                                   printf( "\t\tvalue: %s\n", vals[i] );
                           }
                           ldap_value_free( vals );
                           ldap_memfree( a );
                   }
                   if ( ptr != NULL ) {
                           ber_free( ptr, 0 );
                   }
           }
           /* free the search results */
           ldap_msgfree( res );

/* close and free connection resources */
           ldap_unbind( ld );

return 0;
   }

24.  Appendix B - Namespace Consumed By This Specification

The following 2 prefixes are used in this specification to name func-
tions:
   ldap_
   ber_

The following 6 prefixes are used in this specification to name struc-
tures, unions, and typedefs:
   ldap
   LDAP
   mod_vals_u
   ber
   Ber

Expires: May 2001                                              [Page 74]

C LDAP API        C LDAP Application Program Interface  17 November 2000

timeval

The following 3 prefixes are used in this specification to name #defined
macros:
   LDAP
   LBER_
   mod_

25.  Appendix C - Summary of Requirements for API Extensions

As the LDAP protocol is extended, this C LDAP API will need to be
extended as well.  For example, an LDAPv3 control extension has already
been defined for server-side sorting of search results [7].  This appen-
dix summarizes the requirements for extending this API.

25.1.  Compatibility

Extensions to this document SHOULD NOT, by default, alter the behavior
of any of the APIs specified in this document.  If an extension option-
ally changes the behavior of any existing C LDAP API function calls, the
behavior change MUST be well documented.  If an extension that operates
on an LDAP session affects a chain of messages that was previously
obtained by a call to ldap_result() or by calling a synchronous search
routine, this MUST be well documented.

25.2.  Style

Extensions to this API SHOULD follow the general style and naming con-
ventions used in this document.  For example, function names SHOULD
start with "ldap_" or "ber_" and consist entirely of lowercase letters,
digits, and underscore ('_') characters.  It is RECOMMENDED that private
and experimental extensions use only the following prefixes for macros,
types, and function names:
       LDAP_X_
       LBER_X_
       ldap_x_
       ber_x_
and that these prefixes not be used by standard extensions.

25.3.  Dependence on Externally Defined Types

Extensions to this API SHOULD minimize dependencies on types and macros
that are defined in system headers and generally use only intrinsic
types that are part of the C language, types defined in this specifica-
tion, or types defined in the extension document itself.

Expires: May 2001                                              [Page 75]

C LDAP API        C LDAP Application Program Interface  17 November 2000

25.4.  Compile Time Information

Extensions to this API SHOULD conform to the requirements contained in
the "Retrieving Information at Compile Time" section of this document.
That is, extensions SHOULD define a macro of the form:

#define LDAP_API_FEATURE_x level

so that applications can detect the presence or absence of the extension
at compile time and also test the version or level of the extension pro-
vided by an API implementation.

25.5.  Runtime Information

Extensions to this API SHOULD conform to the requirements contained in
the "Retrieving Information During Execution" section of this document.
That is, each extension SHOULD be given a character string name and that
name SHOULD appear in the ldapai_extensions array field of the LDAPAPI-
Info structure following a successful call to ldap_get_option() with an
option parameter value of LDAP_OPT_API_INFO.  In addition, information
about the extension SHOULD be available via a call to ldap_get_option()
with an option parameter value of LDAP_OPT_API_FEATURE_INFO.

25.6.  Values Used for Session Handle Options

Extensions to this API that add new session options (for use with the
ldap_get_option() and ldap_set_option() functions) SHOULD meet the
requirements contained in the last paragraph of the "LDAP Session Handle
Options" section of this document.  Specifically, standards track docu-
ments MUST use values for option macros that are between 0x1000 and
0x3FFF inclusive and private and experimental extensions MUST use values
for the option macros that are between 0x4000 and 0x7FFF inclusive.

26.  Appendix D - Known Incompatibilities with RFC 1823

This appendix lists known incompatibilities between this API specifica-
tion and the one contained in RFC 1823, beyond the additional API func-
tions added in support of LDAPv3.

26.1.  Opaque LDAP Structure

In RFC 1823, some fields in the LDAP structure were exposed to applica-
tion programmers.  To provide a cleaner interface and to make it easier
for implementations to evolve over time without sacrificing binary com-
patibility with older applications, the LDAP structure is now entirely
opaque.  The new ldap_set_option() and ldap_get_option() calls can be

Expires: May 2001                                              [Page 76]

C LDAP API        C LDAP Application Program Interface  17 November 2000

used to manipulate per-session and global options.

26.2.  Additional Result Codes

The following new result code macros were introduced to support LDAPv3:
   LDAP_REFERRAL
   LDAP_ADMINLIMIT_EXCEEDED
   LDAP_UNAVAILABLE_CRITICAL_EXTENSION
   LDAP_CONFIDENTIALITY_REQUIRED
   LDAP_SASL_BIND_IN_PROGRESS
   LDAP_AFFECTS_MULTIPLE_DSAS
   LDAP_CONNECT_ERROR
   LDAP_NOT_SUPPORTED
   LDAP_CONTROL_NOT_FOUND
   LDAP_NO_RESULTS_RETURNED
   LDAP_MORE_RESULTS_TO_RETURN
   LDAP_CLIENT_LOOP
   LDAP_REFERRAL_LIMIT_EXCEEDED

26.3.  Freeing of String Data with ldap_memfree()

All strings received from the API (e.g., those returned by the
ldap_get_dn() or ldap_dn2ufn() functions) SHOULD be freed by calling
ldap_memfree() not free().  RFC 1823 did not define an ldap_memfree()
function.

26.4.  Changes to ldap_result()

The meaning of the all parameter to ldap_result has changed slightly.
Nonzero values from RFC 1823 correspond to LDAP_MSG_ALL (0x01).  There
is also a new possible value, LDAP_MSG_RECEIVED (0x02).

The result type LDAP_RES_MODDN is now returned where RFC 1823 returned
LDAP_RES_MODRDN.  The actual value for these two macros is the same
(0x6D).

26.5.  Changes to ldap_first_attribute() and ldap_next_attribute

Each non-NULL return value SHOULD be freed by calling ldap_memfree()
after use.  In RFC 1823, these two functions returned a pointer to a
per-session buffer, which was not very thread-friendly.

After the last call to ldap_first_attribute() or ldap_next_attribute(),
the value set in the ptr parameter SHOULD be freed by calling ber_free(

Expires: May 2001                                              [Page 77]

C LDAP API        C LDAP Application Program Interface  17 November 2000

ptr, 0 ).  RFC 1823 did not mention that the ptr value SHOULD be freed.

The type of the ptr parameter was changed from void * to BerElement *.

26.6.  Changes to ldap_modrdn() and ldap_modrdn_s() Functions

In RFC 1823, the ldap_modrdn() and ldap_modrdn_s() functions include a
parameter called deleteoldrdn.  This does not match the great majority
of implementations, so in this specification the deleteoldrdn parameter
was removed from ldap_modrdn() and ldap_modrdn_s().  Two additional
functions that support deleteoldrdn and are widely implemented as well
were added to this specification: ldap_modrdn2() and ldap_modrdn2_s().

26.7.  Changes to the berval structure

In RFC 1823, the bv_len element of the berval structure was defined as
an `unsigned long'.  In this specification, the type is implementation-
specific, although it MUST be an unsigned integral type that is at least
32 bits in size.  See the appendix "Data Types and Legacy Implementa-
tions" for additional considerations.

26.8.  API Specification Clarified

RFC 1823 left many things unspecified, including behavior of various
memory disposal functions when a NULL pointer is presented, requirements
for headers, values of many macros, and so on.  This specification is
more complete and generally tighter than the one in RFC 1823.

26.9.  Deprecated Functions

A number of functions that are in RFC 1823 are labeled as "deprecated"
in this specification.  In most cases, a replacement that provides
equivalent functionality has been defined.  The deprecated functions
are:

ldap_bind()
           Use ldap_simple_bind() or ldap_sasl_bind() instead.

ldap_bind_s()
           Use ldap_simple_bind_s() or ldap_sasl_bind_s() instead.

ldap_kerberos_bind() and ldap_kerberos_bind_s()
           No equivalent functions are provided.

Expires: May 2001                                              [Page 78]

C LDAP API        C LDAP Application Program Interface  17 November 2000

ldap_modrdn() and ldap_modrdn2()
           Use ldap_rename() instead.

ldap_modrdn_s() and ldap_modrdn2_s()
           Use ldap_rename_s() instead.

ldap_open()
           Use ldap_init() instead.

ldap_perror()
           Use ldap_get_option( ld, LDAP_OPT_RESULT_CODE, &rc ) followed
           by fprintf( stderr, "%s: %s", msg, ldap_err2string( rc ))
           instead.

ldap_result2error()
           Use ldap_parse_result() instead.

27.  Appendix E - Data Types and Legacy Implementations

The data types associated with the length of a ber value (ber_len_t),
and the tag (ber_tag_t) have been defined in this specification as
unsigned integral types of implementation-specific size.  The data type
used for encoding and decoding ber integer, enumerated, and boolean
values has been defined in this specification as a signed integral type
of implementation-specific size.  This was done so that source and
binary compatibility of the C LDAP API can be maintained across ILP32
environments (where int, long, and pointers are all 32 bits in size) and
LP64 environments (where ints remain 32 bits but longs and pointers grow
to 64 bits).

In older implementations of the C LDAP API, such as those based on RFC
1823, implementors may have chosen to use an `unsigned long' for length
and tag values.  If a long data type was used for either of these items,
a port of an application to a 64-bit operating system using the LP64
data model would find the size of the types used by the C LDAP API to
increase.  Also, if the legacy implementation had chosen to implement
the tag and types as an unsigned int, adoption of a specification that
mandated use of unsigned longs would cause a source incompatibility in
an LP64 application.  By using implementation-specific data types, the C
LDAP API implementation is free to choose the correct data type and the
ability to maintain source compatibility.

For example, suppose a legacy implementation chose to define the return
value of ber_skip_tag() as an unsigned long but wishes to have the
library return a 32-bit quantity in both ILP32 and LP64 data models.
The following typedefs for ber_tag_t will provide a fixed sized data
structure while preserving existing ILP32 source -- all without

Expires: May 2001                                              [Page 79]

C LDAP API        C LDAP Application Program Interface  17 November 2000

generating compiler warnings:
           #include <limits.h>     /* provides UINT_MAX in ISO C */
           #if UINT_MAX >= 0xffffffffU
               typedef unsigned int ber_tag_t;
           #else
               typedef unsigned long ber_tag_t;
           #endif

Similar code can be used to define appropriate ber_len_t, ber_int_t,
ber_slen_t and ber_uint_t types.

28.  Appendix F - Changes Made Since Last Document Revision

The previous version of this document was draft-ietf-ldapext-ldap-c-
api-04.txt, dated 8 October 1999.  This appendix lists all of the
changes made to that document to produce this one.

28.1.  API Changes

"Header Requirements" section: added requirement that the simple pro-
   gram provided must execute as well as compile without errors.

"LDAP Session Handle Options" section: changed the name of the
   LDAP_OPT_ERROR_NUMBER option to LDAP_OPT_RESULT_CODE.  Allow
   LDAP_OPT_ON to be defined as an implementation specific value (to
   avoid problems on architectures where the value ((void *)1) is not
   usable).

"Initializing an LDAP Session" section: allow use of the value zero
   for the `portno' parameter to mean "use port 389."

"Searching" section: added LDAP_DEFAULT_SIZELIMIT (-1) to allow
   application programmers to use the sizelimit from the LDAP session
   handle with ldap_search_ext() and ldap_search_ext_s().

"Modifying an entry" section: moved mod_vals union out of LDAPMod and
   added mod_vals_u_t typedef so users of the API can declare variables
   using the union type.  "Handling Errors and Parsing Results" section:
   added text to require that ldap_err2string() MUST NOT return NULL.

"A Client Control That Governs Referral Processing" section: modified
   the text to specify that a ber_uint_t value should be used to hold
   the flags.

Expires: May 2001                                              [Page 80]

C LDAP API        C LDAP Application Program Interface  17 November 2000

28.2.  Editorial Changes and Clarifications

"Overview of LDAP API Use and General Requirements" section: added
   text to clarify our use of the term "asynchronous."

"Retrieving Information During Execution" section: added text
   describing the `ldapai_vendor_name' and `ldapai_vendor_version'
   fields (text was accidently deleted during a previous round of
   edits).

"LDAP Session Handle Options" section: improved the text that
   describes the LDAP_OPT_TIMELIMIT, LDAP_OPT_SIZELIMIT, and
   LDAP_OPT_RESULT_CODE options.  Provided details and an example of the
   correct LDAP_OPT_HOST_NAME string to return when the `portno' passed
   to ldap_init() is not zero or 389.

"Result Codes" section: renamed section (was "LDAP Error Codes").

"Authenticating to the directory" section: clarified that the `dn',
   `cred', and `passwd' parameters can be NULL.  Added text indicate
   that the `servercredp' is set to NULL if an API error occurs.

"Performing LDAP Operations" section: replaced "All functions take a
   session handle" with "Most functions...."

"Search" section: removed the detailed discussion of the session han-
   dle options (already covered in the "Retrieving Information During
   Execution" section).  Also removed the word "global" when discussing
   the session default value for the `timeout' parameter.  Also clari-
   fied that a NULL `base' parameter means use a zero-length string for
   the base DN.

"Comparing a Value Against an Entry" section: corrected the "success-
   ful" return codes for ldap_compare_ext_s() and ldap_compare_s() (was
   LDAP_SUCCESS; changed to LDAP_COMPARE_TRUE or LDAP_COMPARE_FALSE).

"Extended Operations" section: added text to indicate that the
   `retoidp' and `retdatap' result parameters are set to NULL if an API
   error occurs in ldap_extended_operation_s().

"Handling Errors and Parsing Results" section: added text to say that
   the `matcheddnp' result parameter will be set to NULL if the server
   does not return a matched DN string.  Added text to indicate that
   serverctrlsp can be NULL.  Added text to indicate that *retoidpp,
   *retdatap, *referralsp, and *serverctrlsp will be set to NULL if no
   items of that type are returned.  Removed specific reference to
   LDAP_NO_SUCH_OBJECT result code when discussing the `matcheddnp'
   result parameter and added clarifying note about "" vs. NULL.

Expires: May 2001                                              [Page 81]

C LDAP API        C LDAP Application Program Interface  17 November 2000

"Parsing References" section: added text to indicate that *refer-
   ralsp, and *serverctrlsp will be set to NULL if no items of that type
   are returned.

"Obtaining Results and Peeking Inside LDAP Messages" section: added
   text to say that LDAPMessage chains MAY be tied to a session handle.

"BER Data Structures and Types" section: removed note about
   ber_uint_t not being used in this document (it is now).  Changed text
   to simplify the description of ber_slen_t.  Removed misleading sen-
   tence about the width of ber_uint_t.

"Encoded ASN.1 Value Manipulation / Encoding" section: added note
   that 'X' is reserved.  Also fixed a few small bugs in the example
   code.

"Encoded ASN.1 Value Manipulation / Decoding" section: clarified the
   requirements for LBER_ERROR and LBER_DEFAULT (expressed using octets
   instead of bits).  Also fixed a few small bugs in the example code.

Added the following text to all descriptions of the `serverctrls' and
   `clientctrls' parameters: ", or NULL if no <server/client> controls
   are to be used."

Added the following text to the description of all `dn' and `rdn'
   parameters: "If NULL, a zero length DN is sent to the server."

Replaced many occurrences of the phrase "error code" with "result
   code" throughout the document.

Added text to indicate that the value of the `msgidp' result parame-
   ter is undefined if an error occurs in the following functions:
   ldap_sasl_bind(), ldap_search_ext(), ldap_compare_ext(),
   ldap_modify_ext(), ldap_add_ext(), ldap_delete_ext(),
   ldap_extended_operation().

Added text to indicate that the `res' result parameter is set to NULL
   if an API error occurs in the following functions: ldap_result(),
   ldap_search_s(), ldap_search_st().

Added text to indicate that all result parameters have undefined
   values if an API error is returned by the following functions:
   ldap_parse_result(), ldap_parse_sasl_bind_result(),
   ldap_parse_extended_result(), ldap_parse_reference(), ber_scanf().

Added angle brackets around ficticious impl_XXX_t types to make it
   more obvious that these are not real "C" types, e.g., typedef
   <impl_len_t> ber_len_t'.

Expires: May 2001                                              [Page 82]

C LDAP API        C LDAP Application Program Interface  17 November 2000

Appendix B: Added mod_vals_u and removed PLDAP from the struct,
   unions, and typedefs prefix list.

Appendix C: Added note in "Compatibility" section about extensions
   possible affecting chains of messages and the fact that that must be
   well documented.  Appendix D: Improved text for ldap_perror() (what
   to use instead).

"Authors" section: updated contact information for Mark Smith, Tim
   Howes, and Mark Wahl.

Fixed a few obvious typos, improved indentation, added missing blank
   lines, and so on.

Expires: May 2001                                              [Page 83]