On Thu, Dec 7, 2017 at 11:03 AM Julian Foad <julianfoad_at_apache.org> wrote:
> Queries/suggestions on svn_client_status6() and svn_client_patch_func_t...
> Index: subversion/include/svn_client.h
> --- subversion/include/svn_client.h (revision 1817399)
> +++ subversion/include/svn_client.h (working copy)
> @@ -2514,9 +2514,9 @@ typedef svn_error_t *(*svn_client_status
> * *result_rev is not meaningful unless @a check_out_of_date is
> * set).
> - * - If @a check_working_copy is not set, do not scan the working
> - * copy for local modifications. This parameter will be ignored
> - * unless @a check_out_of_date is set. When set, the status
> + * - If @a check_working_copy is false, do not scan the working
> + * copy for local modifications. This parameter will be assumed true
> + * unless @a check_out_of_date is set. When false, the status
> * report will not contain any information about local changes in
> * the working copy; this includes local deletions and
> * replacements.
> @@ -7456,18 +7456,24 @@ svn_client_min_max_revisions(svn_revnum_
> - * The callback invoked by svn_client_patch() before attempting to patch
> - * the target file at @a canon_path_from_patchfile (the path as parsed
> - * the patch file, but in canonicalized form). The callback can set
> - * @a *filtered to @c TRUE to prevent the file from being patched, or else
> + * The callback invoked by svn_client_patch() when patching each target
> + *
> + * Called after putting the patch result and any reject in temporary
> + * before moving those files to the real location to complete the
> + *
> + * The callback can set @a *filtered to @c TRUE to prevent moving the
> + * temporary files to the real location to complete the patching, or else
> * must set it to @c FALSE.
> + * @a canon_path_from_patchfile is the path as parsed from the patch file,
> + * but in canonicalized form.
> + *
> * The callback is also provided with @a patch_abspath, the path of a
> * temporary file containing the patched result, and with @a
> * the path to a temporary file containing the diff text of any hunks
> * which were rejected during patching.
Definitely much clearer language!
> - * Because the callback is invoked before the patching attempt is made,
> + * ### ? Because the callback is invoked before the patching attempt is
> * there is no guarantee that the target file will actually be patched
> * successfully. Client implementations must pay attention to notification
> * feedback provided by svn_client_patch() to find out which paths were
The callback is invoked before the patching attempt is made and therefore
no guarantee that the target file will actually be patched successfully.
> - Julian
Received on 2017-12-08 03:26:23 CET