[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

gothicx-guest commit problems



        Hi!

 I stumbled upon some quite strange commits to the wesnoth part (and a
misleading reply to a bugreport) and investigated a bit further to have
found out that there were several strange commits going on, like changes
to debian/changelog changing an already released changelog entry and
name it in the commit message as "watchfile added".

 I tried to talk to Kmos and more or less got a bit of pouting response,
partly evading my questions on what the reasons behind it was, getting
for a lowering of a versioned cdbs dependency (commit message: debhelper
version bumped) a reply like: "the cdbs i changed because I think it
should work too without the -1.1" - which when having known something
about cdbs and reading up on it quickly shows that it is a very bad
idea.

 I've now done the following:
$> for i in $(svn log svn+ssh://svn.debian.org/svn/pkg-games/packages/ \
 | grep gothicx|awk '{print $1}' | cut -b2- ); do svn log -r$i \
 svn+ssh://svn.debian.org/svn/pkg-games/packages/; svn diff \
 -r$(($i - 1)):$i svn+ssh://svn.debian.org/svn/pkg-games/packages/; \
 echo; echo;echo; done > gothicx.log

 And I found some more strange commits in other packages clearly can't
have been tested:

#v-
r4412 | gothicx-guest | 2007-10-14 02:01:03 +0200 (Son, 14 Okt 2007) | 1 line

update watch file
------------------------------------------------------------------------
Index: trunk/vodovod/debian/watch
===================================================================
--- trunk/vodovod/debian/watch  (Revision 4411)
+++ trunk/vodovod/debian/watch  (Revision 4412)
@@ -4,6 +4,4 @@
 # Compulsory line, this is a version 3 file
 version=3
 
-# Examine a Webpage 
-# <Webpage URL> <string match>
-http://home.gna.org/vodovod/ .*vodovod-(.*)-src\.tar\.gz
+http://download.gna.org/vodovod-(.*)-src\.tar\.gz
#v-

 This completely breaks the watch file, there is no such match anymore.

#v+
r4411 | gothicx-guest | 2007-10-14 01:54:58 +0200 (Son, 14 Okt 2007) | 1 line

update watch file
------------------------------------------------------------------------
Index: trunk/rrootage/debian/watch
===================================================================
--- trunk/rrootage/debian/watch (Revision 4410)
+++ trunk/rrootage/debian/watch (Revision 4411)
@@ -4,4 +4,4 @@
 # Compulsory line, this is a version 3 file
 version=3
 
-http://sf.net/rrootage/rRootage-(.*)\.tar\.gz
+http://sf.net/rrootage/rRootage-(.+)\.tar\.gz
#v-

 A pretty useless change, especially since this change wasn't done
consistent at all, in every other commited watch file .* was used, so
what's the deal with this change?

#v+
r4408 | gothicx-guest | 2007-10-14 01:49:18 +0200 (Son, 14 Okt 2007) | 1
line

add watch files
------------------------------------------------------------------------
Index: trunk/wesnoth/debian/changelog
===================================================================
--- trunk/wesnoth/debian/changelog      (Revision 4407)
+++ trunk/wesnoth/debian/changelog      (Revision 4408)
@@ -1,7 +1,7 @@
 wesnoth (1:1.2.7-2) unstable; urgency=low
 
   * Epoch this version, there was a fumble with an development release
   * upload
-    to unstable, closes: #446505
+    to unstable (Closes: #446505).
 
  -- Gerfried Fuchs <rhonda@debian.at>  Sat, 13 Oct 2007 17:12:44 +0200
#v-

 Said "watch file" change in the changelog. No, the watch file wasn't
touched at all, and it was and still is happily living there ever after.

#v+
r4407 | gothicx-guest | 2007-10-14 01:38:53 +0200 (Son, 14 Okt 2007) | 1 line

modified watch
------------------------------------------------------------------------
Index: trunk/wormux/debian/watch
===================================================================
--- trunk/wormux/debian/watch   (Revision 4406)
+++ trunk/wormux/debian/watch   (Revision 4407)
@@ -6,18 +6,5 @@
 # Compulsory line, this is a version 3 file
 version=3
 
-# Uncomment to examine a Webpage 
-# <Webpage URL> <string match>
-#http://www.example.com/downloads.php #PACKAGE#-(.*)\.tar\.gz
-
-# Uncomment to examine a Webserver directory
 #http://www.example.com/pub/#PACKAGE#-(.*)\.tar\.gz
-http://download.gna.org/wormux/wormux-([\d\.]+)\.tar\.gz
-
-# Uncommment to examine a FTP server
-#ftp://ftp.example.com/pub/#PACKAGE#-(.*)\.tar\.gz debian uupdate
-
-# Uncomment to find new files on sourceforge, for debscripts >= 2.9
-# http://sf.net/#PACKAGE#/#PACKAGE#-(.*)\.tar\.gz
-
-
+http://download.gna.org/wormux/wormux-static-([\d\.]+)\.tar\.bz2
#v-

 ... I wonder why one would want to watch wormux-static, which is a
_binary_ package? At least it got me aware of having to update it for
matching the beta packages, too.

 Lots of compat/debhelper bumps, I highly doubt that those changes got
tested or readed up for required/possible changes to debian/rules ...

 r4403 is pretty nasty, it just says "Bump compat to version 5" but
includes control file changes like removing Homepage from the long
description of rrootage which we discussed about _not_ to do before
lenny release, and completely dropped a second URL that was in the
description. It also removes changes to the rrootage.desktop file,
removing Encoding=UTF-8 line from it?! Changing Priority in pixfrogger,
sauerbraten (which isn't bad in itself, but doesn't match the log
message at all, or least get documented in changelog...), ...

 The worst of all, pretty seldom any changes got documented in
debian/changelog files.

 r4393 is a funny one, too:

#v+
-Upstream Author: 
+Upstream Author(s):
#v-

 Sorry, I am only at not even 40% of that output and already need a
break. And please don't get it wrong, I don't say that he commited only
breakage, but the amount of it is simply overwhelming.

 I wonder how other people in the team see such things, if this is the
kind of quality we want the packages to be touched with. Kmos told me
he won't touch wesnoth anymore - but that unfortunately doesn't change
the core problem of breaking stuff...

 Sorry for the length of the mail, but it's not like this is something
that I want or think I should keep to myself, if we want to work as a
team and deliver quality to our users.

 So long,
Rhonda



Reply to: