On Friday 18 July 2008 Maurice van der Pot wrote:
> On Thu, Jul 17, 2008 at 01:13:38PM +0200, Philipp Marek wrote:
> > Does it hurt that much to allow property changes on the directory to be
> > committed? It takes only a few bytes in the repository (AFAIK).
> > Why do you want to avoid that?
>
> It's not so much the committing that I want to avoid as it is the
> showing up in the output of fsvs status. It would be perfect if the need
> to commit was directly linked to whether or not fsvs status generated
> any output.
There's the option "empty_commit" - that can avoid completely no-op commits.
And if you additionally use "-ftext" on commit you shouldn't get meta-only 
changes - but that means that changed rights per se won't be committed, too.

You could try
        fsvs ci -o empty_commit=no -C -f text,owner,group,mode
That should commit exactly what you want - all changes except mtime-only, 
and won't create empty commits.

But please note: the "mode" specification is r1820, only a few minutes 
old - and untested.
(You might want to do a backport of this change, to use a released 
version, and not the current trunk - which is a bit in flux. See the 
end of the mail.)

> It's like compiler warnings: I like to get rid of any messages even if
> they're not really important to make sure that I will not overlook the
> important ones later.
Right.

> > Otherwise you'd need a way to selectively ignore changes on entries ...
> > Hmmm.
>
> Something to look forward to? =)
Does the above suffice? Or do you have the distinct requirement to ignore 
mtime changes on specific entries only?


Regards,

Phil


-- 
Versioning your /etc, /home or even your whole installation?
             Try fsvs (fsvs.tigris.org)!



Index: src/fsvs.c
===================================================================
--- src/fsvs.c  (Revision 1819)
+++ src/fsvs.c  (Revision 1820)
@@ -207,8 +207,8 @@
  * operations, modification of the work done on given entries.
  *
  * It requires a specification at the end, which can be any combination of
- * \c any, \c text, \c new, \c deleted, \c meta, \c mtime, \c group or \c
- * owner.
+ * \c any, \c text, \c new, \c deleted, \c meta, \c mtime, \c group, \c mode
+ * or \c owner.
  *
  * By giving eg. the value \c text, with a \ref status action only entries
  * that are new or changed are shown; with \c mtime,group only entries
Index: src/options.c
===================================================================
--- src/options.c       (Revision 1819)
+++ src/options.c       (Revision 1820)
@@ -73,6 +73,7 @@ const struct opt___val_str_t opt___filte
        { .val=FS_META_CHANGED,                                                 
                        .string="meta" },
        { .val=FS_META_MTIME,                                                   
                        .string="mtime" },
        { .val=FS_META_OWNER,                                                   
                        .string="owner" },
+       { .val=FS_META_UMODE,                                                   
                                .string="mode" },
        { .val=FS_META_GROUP,                                                   
                        .string="group" },
        { .val=FS_NEW,                                                          
                                 .string="new" },
        { .val=FS_REMOVED,                                                      
                                 .string="deleted" },
Index: CHANGES
===================================================================
--- CHANGES     (Revision 1819)
+++ CHANGES     (Revision 1820)
@@ -17,6 +17,7 @@ Changes since 1.1.16
 - New command "rel-ignore" for converting input to relative ignore
        patterns.
 - Try to get the directory mtimes correct again.
+- The filter option now allows "mode", too.

 Changes since 1.1.15
 - FSVS_WARNINGS removed. Use FSVS_WARNING.

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to