Re: CVS update: subversion/subversion/libsvn_fs dag.c editor.c

[Karl Fogel <kf...@galois.collab.net>]

Branko, we should talk very soon -- unfortunately, Ben and I have to
run out right now, but the quick summary is:

What we're doing in dag.c right now is implementing what's necessary
for the commit tree editor to work.  This is a much smaller subset of
what tree.c will do, because the editor is depth-first and a single
drive (no multi-accessor stuff going on).

More later,
-K

Branko =?ISO-8859-2?Q?=C8ibej?= <br...@xbc.nu> writes:
> kfogel@tigris.org wrote:
> 
> >   (dag_node_t): Start defining this struct.
> 
> Eek, that was a close shave. I could fairly hear the commit whistling 
> past my ears while I was typing C-x C-q!
> 
> >    /* Initializing a filesystem.  */
> >    
> >   +struct dag_node_t
> >   +{
> >   +  svn_fs_t *fs;
> >   +  svn_fs_id_t *id;
> >   +  apr_pool_t *pool;
> >   +  /* kff todo: working here */
> >   +};
> >   +
> 
> I think we'll have to know the node's path so that we can create "copy" 
> nodes from dag_node_t's sources. We have that information in 
> svn_fs_node_t; we can
> 
> a)  move it to dag_node_t instead, or
> b)  pass it to svn_fs__dag_make_copy explicitly (and hope to get it right).
> 
> I'd prefer a), but it would cause a few changes in tree.c.
> 
> -- 
> Brane �ibej
>     home:   <br...@xbc.nu>             http://www.xbc.nu/brane/
>     work:   <br...@hermes.si>   http://www.hermes-softlab.com/
>      ACM:   <br...@acm.org>            http://www.acm.org/