From 1be510f9529cb082f802408b472a77d074b394c0 Mon Sep 17 00:00:00 2001 From: Nicolas Vigier Date: Sun, 14 Apr 2013 13:46:12 +0000 Subject: Add zarb MLs html archives --- zarb-ml/mageia-dev/2011-September/008179.html | 128 ++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 zarb-ml/mageia-dev/2011-September/008179.html (limited to 'zarb-ml/mageia-dev/2011-September/008179.html') diff --git a/zarb-ml/mageia-dev/2011-September/008179.html b/zarb-ml/mageia-dev/2011-September/008179.html new file mode 100644 index 000000000..bf0e29115 --- /dev/null +++ b/zarb-ml/mageia-dev/2011-September/008179.html @@ -0,0 +1,128 @@ + + + + [Mageia-dev] [RPM] cauldron core/release apache-2.2.20-1.mga2 + + + + + + + + + +

[Mageia-dev] [RPM] cauldron core/release apache-2.2.20-1.mga2

+ nicolas vigier + boklm at mars-attacks.org +
+ Sat Sep 17 14:55:11 CEST 2011 +

+
+ +
On Thu, 01 Sep 2011, Anssi Hannula wrote:
+
+> On 01.09.2011 15:23, Guillaume Rousse wrote:
+> > Le 01/09/2011 13:10, Mageia Team a écrit :
+> >> guillomovitch<guillomovitch>  2.2.20-1.mga2:
+> >> + Revision: 137516
+> >> - add missing LSB headers to htcacheclean init script
+> > 
+> > The script handling the changelog need some serious fixes. Usually it
+> > cumulates the changes for several releases, while here it only take
+> > accounts the changes between previous failed submission attempt, and
+> > this one.
+> 
+> What is happening is that markreleases for mga1 are done for the
+> cauldron tree, causing changelogs for cauldron releases to be grouped
+> according to submissions of mga1 updates_testing:
+> 
+> $ mgarepo rpmlog -s apache
+> * Thu Sep 01 2011 guillomovitch <guillomovitch> 2.2.20-1.mga2
+> + Revision: 137516
+> - add missing LSB headers to htcacheclean init script
+> 
+> * Wed Aug 31 2011 guillomovitch <guillomovitch> 2.2.17-5.1.mga1
+> + Revision: 137231
+> - new version
+> - drop suexec PATH patch
+> [...]
+> 
+> 
+> As I noted in #mageia-sysadm several months ago, the steps needed to fix
+> this are:
+> 
+> 1) disable markrelease on these submits
+> 2) propedit the wrong markreleases (look for post-final mga1
+> markreleases in cauldron tree) to have invalid log msgs so that mgarepo
+> will ignore them
+> 3) add markrelease to the final update move (from _testing to updates)
+> script, or make the youri action run markrelease in the originating tree
+> instead of cauldron
+
+I think it's difficult to do markrelease at final update move, as we
+don't know the commit number for the markrelease.
+
+Maybe this change in youri config would allow to have the markrelease
+done on the right directory for updates :
+
+Index: modules/buildsystem/templates/submit-upload.conf
+===================================================================
+--- modules/buildsystem/templates/submit-upload.conf	(revision 2006)
++++ modules/buildsystem/templates/submit-upload.conf	(working copy)
+@@ -14,11 +14,12 @@
+         upload_state: queue
+         queue: queue
+         noarch: i586
+-        svn: svn+ssh://svn.<%= domain %>/svn/packages/cauldron/
+         1:
+             arch: i586 x86_64
++            svn: svn+ssh://svn.<%= domain %>/svn/packages/updates/1/
+         cauldron:
+             arch: i586 x86_64
++            svn: svn+ssh://svn.<%= domain %>/svn/packages/cauldron/
+ 
+ # targets definition
+ targets:
+
+
+But this would not be good for backports. The URL used for submit is
+not saved, so it's difficult to do the markrelease on the right
+directory. However if youri were using a database to store all the infos
+about submits (URL, commit number, submitter, etc ...), this would be
+possible. It would also be possible to do the markrelease when the
+package is moved, with the move script using youri database.
+
+
+ + +
+

+ +
+More information about the Mageia-dev +mailing list
+ -- cgit v1.2.1