[sane-devel] Re: [sane-commit] Changed format of commit mails

[Olaf Meeuwissen]

Henning Meier-Geinitz <henning at meier-geinitz.de> writes:

> Hi,
>
> On Thu, Jul 21, 2005 at 01:29:27PM +0200, Gerhard Jaeger wrote:
>> well in the this doesn't look too bad if it works ;)
>> Olivers' and my last commits didn't produce a correct output and CVS mumbles
>> something about deprecated format strings...
>> 
>> How could this be fixed?
>
> I don't really know. When I tested it it did work. I tested again 10
> times and got 2 failures and 8 good mails.
>
> Some background: Recently, in cvs the format of the variables given to
> scripts have been changed (e.g. in CVSROOT/loginfo). The old way gave
> everthing in ${sVv} as one argument splitted by spaces to the script
> while the new way uses separate arguments. If you still use the old
> way, you get a depracation warning. The new way can be selected by a
> switch in CVSROOT/config ("UseNewInfoFmtStrings=yes").

Running a local GForge, I noticed the same and checked it out.  In my
case it turned out to be a cron job that zaps whatever CVSROOT/config
is there and replaces it with a hardcoded one.  The script in question
is deb-specific/cvs_dump_update.pl and the job runs every 30 minutes.

I'm not sure Alioth uses an out-of-the-box GForge, but you may want to
contact the Alioth maintainer(s) or file a bug report.  I've attached
a diff of my local kludge.

> This switch is enabled. However, sometimes cvs seems to ignore it (or
> does not read the config at all). Therefore it assumes the old way of
> handling formats is enabled and emits the deprecation warning. As the
> commit log script expected the new format, several error messages were
> printed and the commit messages looked strange.
>
> I have no idea why cvs sometimes doesn't accept the keyword for the
> new format handling and sometimes does.
>
> Anyway, I've changed the behaviour now by using the old formats in the
> script manually (using "%1" instead of "%"). This should work with
> both the switch enabled and disabled. However, now a different warning
> is printed. Maybe there is a cvs guru out there who understands what's
> going on?

Hope this helps,
-- 
Olaf Meeuwissen                          EPSON AVASYS Corporation, LAN
FSF Associate Member #1962           sign up at http://member.fsf.org/
GnuPG key: 6BE37D90/AB6B 0D1F 99E7 1BF5 EB97  976A 16C7 F27D 6BE3 7D90
Penguin's lib!       -- I hack, therefore I am --               LPIC-2
-------------- next part --------------
Index: deb-specific/cvs_dump_update.pl
===================================================================
RCS file: /cvsroot/tux4ge/source/deb-specific/cvs_dump_update.pl,v
retrieving revision 1.1
retrieving revision 1.2
diff -u -r1.1 -r1.2
--- deb-specific/cvs_dump_update.pl	10 Feb 2005 01:38:04 -0000	1.1
+++ deb-specific/cvs_dump_update.pl	20 Jun 2005 08:31:05 -0000	1.2
@@ -126,6 +126,7 @@
 			if($verbose) { print("Enable pserver for $group_name:\t$userlist in $cvs_dir/CVSROOT/writers \n"); }
 			open (CONFIG,">$cvs_dir/CVSROOT/config");
 			print CONFIG "SystemAuth=yes\n";
+			print CONFIG "UseNewInfoFmtStrings=yes\n";
 			close CONFIG;
 		} else {
 			# turn off pserver writers
@@ -136,6 +137,7 @@
 			if($verbose) { print("Disable pserver for $group_name\n"); }
 			open (CONFIG,">$cvs_dir/CVSROOT/config");
 			print CONFIG "SystemAuth=no\n";
+			print CONFIG "UseNewInfoFmtStrings=yes\n";
 			close CONFIG;
 		}