[Date Prev][Date Next] [Chronological] [Thread] [Top]

Re: liblber docs

At 01:51 AM 2002-01-13, Howard Chu wrote:
>I started going over the lber man pages, to bring them up to date and realized
>that some of my recent changes are a bit of a mess. In particular, liblber has
>code to read and write BVarrays but the functions to construct and free them
>in slapd - bvarray_add, bvarray_free.  They belong in liblber/memory.c, but now
>I have to face the fact that I hate the name "BVarray" and don't want to have
>to use functions named ber_bvarray_add() etc... Anyone got a suggestion for a
>shorter, more convenient name?

To be consistent with the draft spec, BVarray should be named BerValueArray
or BerVAarray (see name space requirements).  And routines managing them
should likely be called ber_bvarray_*() etc..  Sorry.

>Also I noticed that the encode and decode man pages talk about the library only
>dealing with single-octet tags, but the code itself has handled multi-octet
>tags for quite a long time.

That's because for the IETF draft we needed to support tags of up to 4-octet
tags (to better handle LDAPv3 extensions).

>Of course, nothing in the source tree actually
>needs/uses multi-octet support. (Perhaps because the docs have always said
>single-octet only...?)

The docs are out dated.  We need multi-octet tag support.