Bug#774794: Debian Bug#774794: Re: dpkg: Add breaks for new trigger cycles
Hi, Guillem Jover guil...@debian.org wrote: And to the gconf maintainers, does its trigger really need to be synchronous (i.e. do trigggering packages need to await the trigger processing)? Or in other words do those other triggering packages require the trigger to be processed to be able to consider themselves configured and able to satisfy dependencies? The package (typically an end-user application) is not usable until the trigger has been run. That said, the dependencies of a package using GConf would not usually require the trigger to have run for their own postinst to run successfully. I cannot exclude a corner case where a dependency would use a binary needing registered GConf schemas in its postinst, but it is definitely not the typical use case. So it could probably be migrated to a noawait trigger. The same holds for the /usr/share/glib-2.0/schemas trigger in libglib2.0-0 (but probably not the other triggers of the same package). Cheers, -- Joss -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#774794: Debian Bug#774794: Re: dpkg: Add breaks for new trigger cycles
[ CCing gconf maintainers. ] Hi! On Tue, 2015-02-03 at 10:19:29 +0100, Andreas Beckmann wrote: On 2015-02-03 07:12, Niels Thykier wrote: That said, I have no status on the cycles mentioned by Andreas Beckmann (CC'ed) in this bug report. If they are still relevant, we should have bugs for them. I can still reproduce these problems, not sure which package to be blamed for and whether its trigger cycles or something different. The debian-security-support and doc-base instances are just missing Breaks in dpkg, which I've added. The only remaining problem is the gconf2 one, which from a cursory check looks suspiciously like the dbus problem. (Sorry for not having checked that one before, it looked like something else, because it's not really a trigger cycle.) Andreas, if you have such system readily available, it would be nice to know if apt-get can get out of that state or if it gets stuck there. And to the gconf maintainers, does its trigger really need to be synchronous (i.e. do trigggering packages need to await the trigger processing)? Or in other words do those other triggering packages require the trigger to be processed to be able to consider themselves configured and able to satisfy dependencies? Thanks, Guillem -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#774794: Debian Bug#774794: Re: dpkg: Add breaks for new trigger cycles
On 2015-02-03 07:12, Niels Thykier wrote: That said, I have no status on the cycles mentioned by Andreas Beckmann (CC'ed) in this bug report. If they are still relevant, we should have bugs for them. I can still reproduce these problems, not sure which package to be blamed for and whether its trigger cycles or something different. Andreas -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#774794: Debian Bug#774794: Re: dpkg: Add breaks for new trigger cycles
On 2015-01-07 20:12, Niels Thykier wrote: On 2015-01-07 18:41, Niels Thykier wrote: Package: dpkg Version: 1.17.23 Severity: serious Control: block -1 by 774791 Hi, This is a bug to track missing Breaks for new trigger cycles (i.e. cycles found after the 1.17.23 upload). * grace (#774558, fixed in sid and testing) * php5 (#774559, fixed in sid - has a FTBFS on ppc64el) * xine-ui (#774791, not fixed yet) Thanks in advance, ~Niels Please add gitweb to the above list (#774803). ~Niels Hi The above packages have now been fixed in unstable and there currently no bugs left blocking this bug. That said, I have no status on the cycles mentioned by Andreas Beckmann (CC'ed) in this bug report. If they are still relevant, we should have bugs for them. Thanks, ~Niels -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#774794: Debian Bug#774794: Re: dpkg: Add breaks for new trigger cycles
Hi! On Wed, 2015-01-07 at 20:12:28 +0100, Niels Thykier wrote: On 2015-01-07 18:41, Niels Thykier wrote: Package: dpkg Version: 1.17.23 Severity: serious Control: block -1 by 774791 This is a bug to track missing Breaks for new trigger cycles (i.e. cycles found after the 1.17.23 upload). * grace (#774558, fixed in sid and testing) * php5 (#774559, fixed in sid - has a FTBFS on ppc64el) * xine-ui (#774791, not fixed yet) Please add gitweb to the above list (#774803). Thanks, had that already in mind when you pointed them out. I'll queue a patch for the 1.17.x branch adding Breaks for these, and another one changing the = Breaks to ones. I'll probably wait for an upload until the dbus issue has been checked by the apt team (mail upcoming). Thanks also for filing those other bug reports! Regards, Guillem -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#774794: Debian Bug#774794: Re: dpkg: Add breaks for new trigger cycles
On 2015-01-07 18:41, Niels Thykier wrote: Package: dpkg Version: 1.17.23 Severity: serious Control: block -1 by 774791 Hi, This is a bug to track missing Breaks for new trigger cycles (i.e. cycles found after the 1.17.23 upload). * grace (#774558, fixed in sid and testing) * php5 (#774559, fixed in sid - has a FTBFS on ppc64el) * xine-ui (#774791, not fixed yet) Thanks in advance, ~Niels Please add gitweb to the above list (#774803). ~Niels -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org