Daniel Rall writes:
> On Thu, 14 Jun 2007, Blair Zajac wrote:
>
> > But svn_fs_type() takes a char *, not a svn_fs_t * as an argument.
> >
> > svn_error_t *svn_fs_type(const char **fs_type, const char *path,
> > apr_pool_t *pool);
> >
> > So even if you had an svn_fs_t, you have to get the original path to the
> > repository and pass it here.
> >
> > Agreed that caching the type is a good idea.
>
> Bleh. (Sorry I didn't look at the API again this mornin'.)
>
> We could introduce something like:
>
> svn_error_t *svn_fs_type2(const char **fs_type, const svn_fs_t *fs,
> apr_pool_t *pool);
>
The problem here is that libsvn_repos needs to detect the fs type before
opening it, so introducing a new function and deprecating the old one
will not work here either. I can understand that getting the fs type
from the handle is convenient, but what is the *real* problem we're trying to
solve here? Is it just eliminating a bunch of I/O operations or is it
more than that? If it is just the I/O, then maybe that's so minor that
it's not worth cluttering the API for?
Thanks,
//Peter
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Fri Jun 15 08:39:54 2007