> ***************
> *** 53,55 ****
> --- 53,66 ----
> enforced... and svn_fs__things_different should work when comparing
> any two dag nodes at all. Going to try removing the restriction...
>
> +
> +
> + ------- Additional Comments From sussman@tigris.org 2002-07-01 14:08 PDT -------
> + Okay, I tweaked the fs function in r2386, and it fixed the problem as
> + predicted.
> +
> + All fs-tests pass. Heck, all of 'make check' passes. I'm gonna
> + backport this to the 2092 branch as well.
> +
> + Closing issue. Scott, update your working copy, rebuild, and try
> + again, should be all good now. Thanks for finding this!
Um...why is the dumper not using an apply_textdelta editor function?
That is, why is svn_repos_dir_delta doing all the work of figuring out
if file contents need to be transmitted, and the dumper ignoring it
and re-deriving the same information?!
I assume this would complicate the dumper just a little bit since it
wants to do everything node-related in add_file() and open_file(), but
I think that if those two functions just created the file_baton and
stored the action (add, replace, or change), and a new apply_textdelta
stuffed a bit saying that there actually were changes to the file's
contents, then all the real work could be done in a new close_file().
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Tue Jul 2 16:58:47 2002