bliss@tigris.org wrote:
>Author: bliss
>Date: Fri May 14 17:14:14 2004
>New Revision: 9716
>
>Modified:
> trunk/subversion/libsvn_wc/adm_crawler.c
>Log:
>Fix issue #1854 (user data loss on case-insensitive file systems)
>
>* subversion/libsvn_wc/adm_crawler.c
> (report_revisions): Call svn_io_check_path for entries that appear
> to be missing to be sure that they are really missing.
>
>
>Modified: trunk/subversion/libsvn_wc/adm_crawler.c
>==============================================================================
>--- trunk/subversion/libsvn_wc/adm_crawler.c (original)
>+++ trunk/subversion/libsvn_wc/adm_crawler.c Fri May 14 17:14:14 2004
>@@ -214,7 +214,7 @@
> apr_ssize_t klen;
> void *val;
> const svn_wc_entry_t *current_entry;
>- svn_node_kind_t *dirent_kind;
>+ svn_node_kind_t *dirent_kind, dirent_kind_storage;
> svn_boolean_t missing = FALSE;
>
> /* Clear the iteration subpool here because the loop has a bunch
>@@ -250,7 +250,17 @@
> /* Is the entry on disk? Set a flag if not. */
> dirent_kind = (svn_node_kind_t *) apr_hash_get (dirents, key, klen);
> if (! dirent_kind)
>- missing = TRUE;
>+ {
>+ /* It is possible on a case insensitive system that the
>+ entry is not really missing, so we call our trusty but
>+ expensive friend svn_io_check_path to be sure. */
>+ SVN_ERR (svn_io_check_path (this_full_path, &dirent_kind_storage,
>+ iterpool));
>+ if (dirent_kind_storage == svn_node_none)
>+ missing = TRUE;
>+ else
>+ dirent_kind = &dirent_kind_storage;
>+ }
>
>
Would it make sense for Subversion to fix the case of such files on
case-insensitive filesystems?
-- Brane
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Sat May 15 09:06:18 2004