>> Will we be adding a function to svn_fs.h - presumably named something
>> like svn_fs_txn_mergeinfo()? - that allows external callers to call this
>> vtable entry? (I'm slightly surprised that we didn't include it in this
>> commit, actually, since that would be the ideal place to document what
>> this call actually does).
>>
>
> While this API is currently used only inside the FS backends (which
> access fs-loader.h's struct declarations through relative includes),
> it might make sense to include a svn_fs.h API for this new hook for
> general consumption. Thanks for the review, Malcom.
>
> Kamesh, how does this fit in with your second patch? Should we add
> the API separately?
>
> - Dan
>
I could not see the need to expose this API to svn_fs.h as I failed to
see the utility value with this API outside libsvn_fs_util.
I would like to restate the fact that this API gives sophisticated
HASH('path Vs svn:mergeinfo propval'.....) which can only be consumed by
libsvn_fs_util.
Ultimately I would like to remove it if possible to derive the same
stuff by some other sophisticated means.
Slightly different thought (May be offtopic)
If I understand correctly ultimately we might have 'merge-tracking'
implemented without 'svn:mergeinfo'.
In that case I would like "root_vtable_t's change_merge_info" moved to
"txn_vtable_t". Currently I could not see the use case for
"root_vtable_t's change_merge_info".
Then also I could not see the need to expose "txn_vtable_t's
get_mergeinfo" via svn_fs.h as the equivalent info might be available as
some 'wc xml files'.
My second patch does not expose this hook.
With regards
Kamesh Jayachandran
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Tue Sep 19 07:39:24 2006