Bug#1055510: Best way to coordinate this fix
On 2023-11-10 at 02:45:14, Helmut Grohne (hel...@subdivi.de) wrote: > Thank you. The package built and dumat has imported it. I locally forked > its analysis database pretending that systemd would not declare a > conflict for molly-guard and reran it on that database. It does not > report any issues for molly-guard 0.8. I also checked the underlying > database and see that it recognizes the duplicated diversions there. Thanks for all of the help Helmut! I have uploaded 0.8.1 to unstable. Francois -- https://fmarier.org/
Bug#1055510: Best way to coordinate this fix
Hi Francois, On Thu, Nov 09, 2023 at 10:27:51PM -0800, Francois Marier wrote: > I've just uploaded to experimental. If there are any tests you can easily > run there, please do so. Thank you. The package built and dumat has imported it. I locally forked its analysis database pretending that systemd would not declare a conflict for molly-guard and reran it on that database. It does not report any issues for molly-guard 0.8. I also checked the underlying database and see that it recognizes the duplicated diversions there. > I've upgraded in unstable from the current version to 0.8 without problems, > so that should in theory work when I eventually upload to unstable. >From a /usr-merge pov, I have no objections to uploading this to unstable immediately. Do note that backporting 0.8 to bookworm should not be done without reverting the file moves that I added. I hope that backporting molly-guard is not an important user story. Helmut
Bug#1055510: Best way to coordinate this fix
On 2023-11-08 at 21:15:58, Helmut Grohne (hel...@subdivi.de) wrote: > Thank you. I suggest going via experimental first. I've just uploaded to experimental. If there are any tests you can easily run there, please do so. I've upgraded in unstable from the current version to 0.8 without problems, so that should in theory work when I eventually upload to unstable. Francois -- https://fmarier.org/
Bug#1055510: Best way to coordinate this fix
Hi Francois, On Wed, Nov 08, 2023 at 01:00:33PM -0800, Francois Marier wrote: > If anybody would like to double-check (or test) the package I've prepared, > it's in salsa right now. Thank you. I suggest going via experimental first. https://wiki.debian.org/UsrMerge recommends uploading to experimental when moving files from / to /usr. This allows some qa already and in particular dumat will have a second look at it. That look is fairly limited at this time, because systemd conflicts with molly-guard now, but I can locally make it pretend it would not and feed back the result to you. Alternatively, Luca can perform another experimental upload of systemd adding a version constraint to the conflict for molly-guard. Does that extra detour make sense to you? Helmut
Bug#1055510: Best way to coordinate this fix
On Thu, Nov 9 2023 at 12:13:25 AM +01:00:00, Simó Albert i Beltran wrote: Please take a look at https://salsa.debian.org/debian/molly-guard/-/commit/c1120c0c3602955abe02d4d810985ad13d02cdba Sorry, I meant https://salsa.debian.org/debian/molly-guard/-/commit/c1120c0c3602955abe02d4d810985ad13d02cdba#note_440065
Bug#1055510: Best way to coordinate this fix
Please take a look at https://salsa.debian.org/debian/molly-guard/-/commit/c1120c0c3602955abe02d4d810985ad13d02cdba
Bug#1055510: Best way to coordinate this fix
If anybody would like to double-check (or test) the package I've prepared, it's in salsa right now. Francois
Bug#1055510: Best way to coordinate this fix
Control: tags -1 + patch On Tue, Nov 07, 2023 at 09:35:54PM +0100, Helmut Grohne wrote: > On Tue, Nov 07, 2023 at 08:11:19PM +, Luca Boccassi wrote: > > On Tue, 7 Nov 2023 at 20:04, Francois Marier wrote: > > > What's the best way to coordinate a fix for this? > > > > > > I assume that we shouldn't upload a new molly-guard packages until the > > > files > > > have actually moved in the systemd package? > > > > > > Should we wait until systemd is in unstable to push a new molly-guard out? > > If the workaround is correctly implemented, it can go to unstable > directly. Note that it's not simply changing the diversions from the old > location to the new location. The workaround is duplicating them. At > that point, you can no longer use --rename and have to do the renaming > by hand. I'm happy to review a patch. For opensysusers I sent one, but I > haven't gotten down to molly-guard yet. Proposed patch attached. Helmut diff --minimal -Nru molly-guard-0.7.2/Makefile molly-guard-0.7.2+nmu1/Makefile --- molly-guard-0.7.2/Makefile 2019-07-09 18:50:53.0 +0200 +++ molly-guard-0.7.2+nmu1/Makefile 2023-11-08 14:16:52.0 +0100 @@ -1,7 +1,7 @@ PREFIX?=/usr cfgdir?=/etc/molly-guard -libdir?=/lib -sbindir?=/sbin +libdir?=$(PREFIX)/lib +sbindir?=$(PREFIX)/sbin REALPATH?=$(libdir)/molly-guard all: molly-guard.8 shutdown diff --minimal -Nru molly-guard-0.7.2/debian/changelog molly-guard-0.7.2+nmu1/debian/changelog --- molly-guard-0.7.2/debian/changelog 2019-07-09 18:50:53.0 +0200 +++ molly-guard-0.7.2+nmu1/debian/changelog 2023-11-08 14:47:40.0 +0100 @@ -1,3 +1,11 @@ +molly-guard (0.7.2+nmu1) UNRELEASED; urgency=medium + + * Non-maintainer upload. + * Duplicate aliased diversions for DEP17 M18. (Closes: #-1) + * Move all files to /usr. + + -- Helmut Grohne Wed, 08 Nov 2023 14:47:40 +0100 + molly-guard (0.7.2) unstable; urgency=medium * upload to unstable diff --minimal -Nru molly-guard-0.7.2/debian/molly-guard.lintian-overrides molly-guard-0.7.2+nmu1/debian/molly-guard.lintian-overrides --- molly-guard-0.7.2/debian/molly-guard.lintian-overrides 2019-07-09 18:50:53.0 +0200 +++ molly-guard-0.7.2+nmu1/debian/molly-guard.lintian-overrides 2023-11-08 14:47:40.0 +0100 @@ -1,9 +1,9 @@ -molly-guard: binary-without-manpage sbin/coldreboot -molly-guard: binary-without-manpage sbin/halt -molly-guard: binary-without-manpage sbin/pm-hibernate -molly-guard: binary-without-manpage sbin/pm-suspend -molly-guard: binary-without-manpage sbin/pm-suspend-hybrid -molly-guard: binary-without-manpage sbin/poweroff -molly-guard: binary-without-manpage sbin/reboot -molly-guard: binary-without-manpage sbin/shutdown -molly-guard: diversion-for-unknown-file usr/sbin/* preinst:29 +molly-guard: no-manual-page [usr/sbin/coldreboot] +molly-guard: no-manual-page [usr/sbin/halt] +molly-guard: no-manual-page [usr/sbin/pm-hibernate] +molly-guard: no-manual-page [usr/sbin/pm-suspend] +molly-guard: no-manual-page [usr/sbin/pm-suspend-hybrid] +molly-guard: no-manual-page [usr/sbin/poweroff] +molly-guard: no-manual-page [usr/sbin/reboot] +molly-guard: no-manual-page [usr/sbin/shutdown] +molly-guard: diversion-for-unknown-file sbin/* preinst:* diff --minimal -Nru molly-guard-0.7.2/debian/molly-guard.postrm molly-guard-0.7.2+nmu1/debian/molly-guard.postrm --- molly-guard-0.7.2/debian/molly-guard.postrm 2019-07-09 18:50:53.0 +0200 +++ molly-guard-0.7.2+nmu1/debian/molly-guard.postrm2023-11-08 14:47:40.0 +0100 @@ -20,7 +20,11 @@ case "$1" in remove) for cmd in halt poweroff reboot shutdown coldreboot ; do -dpkg-divert --package molly-guard --rename --remove /sbin/$cmd +dpkg-divert --package molly-guard --no-rename --remove /sbin/$cmd +dpkg-divert --package molly-guard --no-rename --remove "/usr/sbin/$cmd" + if test -e "/usr/lib/molly-guard/$cmd"; then +mv "/usr/lib/molly-guard/$cmd" "/usr/sbin/$cmd" +fi done for cmd in pm-hibernate pm-suspend pm-suspend-hybrid ; do diff --minimal -Nru molly-guard-0.7.2/debian/molly-guard.preinst molly-guard-0.7.2+nmu1/debian/molly-guard.preinst --- molly-guard-0.7.2/debian/molly-guard.preinst2019-07-09 18:50:53.0 +0200 +++ molly-guard-0.7.2+nmu1/debian/molly-guard.preinst 2023-11-08 14:47:40.0 +0100 @@ -14,7 +14,7 @@ case "$1" in install|upgrade) -mkdir -p /lib/molly-guard +mkdir -p /usr/lib/molly-guard # Cleanup erroneous diversions added in 0.6.0 for cmd in pm-hibernate pm-suspend pm-suspend-hybrid ; do @@ -22,11 +22,27 @@ done for cmd in halt poweroff reboot shutdown coldreboot ; do -dpkg-divert --package molly-guard --divert /lib/molly-guard/$cmd --rename /sbin/$cmd +dpkg-divert --package molly-guard --divert "/usr/lib/molly-guard/$cmd" --no-rename --add "/usr/sbin/$cmd" +#
Bug#1055510: Best way to coordinate this fix
Hi Francois, On Tue, Nov 07, 2023 at 08:11:19PM +, Luca Boccassi wrote: > On Tue, 7 Nov 2023 at 20:04, Francois Marier wrote: > > What's the best way to coordinate a fix for this? > > > > I assume that we shouldn't upload a new molly-guard packages until the files > > have actually moved in the systemd package? > > > > Should we wait until systemd is in unstable to push a new molly-guard out? If the workaround is correctly implemented, it can go to unstable directly. Note that it's not simply changing the diversions from the old location to the new location. The workaround is duplicating them. At that point, you can no longer use --rename and have to do the renaming by hand. I'm happy to review a patch. For opensysusers I sent one, but I haven't gotten down to molly-guard yet. Helmut
Bug#1055510: Best way to coordinate this fix
On Tue, 7 Nov 2023 at 20:04, Francois Marier wrote: > > Hi Luca, > > What's the best way to coordinate a fix for this? > > I assume that we shouldn't upload a new molly-guard packages until the files > have actually moved in the systemd package? > > Should we wait until systemd is in unstable to push a new molly-guard out? Hi, I believe you can upload to unstable straight away, as it would be adding new diversions which should be fine if nothing is there. For confirmation CC'ing Helmut. Kind regards, Luca Boccassi
Bug#1055510: Best way to coordinate this fix
Hi Luca, What's the best way to coordinate a fix for this? I assume that we shouldn't upload a new molly-guard packages until the files have actually moved in the systemd package? Should we wait until systemd is in unstable to push a new molly-guard out? Francois -- https://fmarier.org/