[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: question about private header

From: Ben Collins-Sussman <sussman_at_red-bean.com>
Date: 2005-11-10 05:58:30 CET

On 11/9/05, Garrett Rooney <rooneg@electricjellyfish.net> wrote:
> On 11/9/05, Mary Jane Lemur <mjlemur@gmail.com> wrote:
> > hi list,
> >
> > i'm just starting to poke around svn source, and i noticed that in
> > libsvn_fs_base one of the files (reps-strings.c) directly #includes
> > ../libsvn_delta/delta.h. i guess it's because that C file calls
> > svn_txdelta__compose_windows... anyway, i read the hacking page and it seems
> > like that's a violation of svn library policy. why isn't the
> > compose_windows func something pubilc in svn_delta.h? it seems weird to me.
>
> We can break our own policies internally if we want to.

Yeah, we *can*, but *should* we?

>
> The Subversion libraries are only designed to work in a matched set
> (i.e. you need version 1.3.0 of libsvn_delta AND libsvn_fs_base, you
> can't use different versions), so poking around at the "not for public
> use" functions in one subversion library from another one is ok.
>
> But just because we do it, doesn't mean anyone else can ;-)
>

It seems sloppy. Why wouldn't we be following our own
conventions/rules about svn_foo_ and svn_foo__ ? Is this some special
exception? In other words, if I were to commit a fix for this right
now (making svn_txdelta_compose_windows() into a public API), would
someone object? Or is there some secret mainteance-nightmare reason
*not* to do so? It's not clear to me if this is an intentional thing,
or just a bit of sloppiness nobody's bothered to clean up.

If nobody objects, I'll fix tomorrow.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Thu Nov 10 05:59:15 2005

This is an archived mail posted to the Subversion Dev mailing list.