Bug#724306: Bug #724306: pu: package dpkg/1.16.11
On Thu, 2013-09-26 at 04:46 +0200, Guillem Jover wrote: On Tue, 2013-09-24 at 19:47:16 +0100, Adam D. Barratt wrote: Is there a plan for fixing #717983 in unstable in the near future? As things currently stand, 1.16.11 would have to be pushed in to testing as part of the point release. Yes, sorry about that, I didn't want to push an upload yet due to some problems with the commit mailing list, and because when manpages-it got fixed the problem became pretty much non-urgent. But on a second thought it might be blocking packages depending on 1.17.0 features, so I could have for example lowered the severity (even if temporarily). In any case I was planning on releaseing 1.17.2 in a couple of days anyway, so that should turn this into a non-issue. unstable still appears to have 1.17.1, so it looks like we may well be pushing 1.6.12 in to testing on Saturday. Regards, Adam -- To UNSUBSCRIBE, email to debian-release-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/1381345913.29291.25.ca...@jacala.jungle.funky-badger.org
Bug#724306: Bug #724306: pu: package dpkg/1.16.11
On Tue, 2013-10-01 at 06:11 +0200, Guillem Jover wrote: On Mon, 2013-09-30 at 18:57:15 +0100, Adam D. Barratt wrote: On Mon, 2013-09-30 at 16:59 +0200, Guillem Jover wrote: Thanks, unfortunately 724949 just came in a day after the upload, it involves improper caching of the «dpkg --print-architecture» and «gcc -dumpmachine» output, affecting the performance of wanna-build. This was already fixed in 1.17.0, so it has been tested for a while. I was wondering if you'd be fine with a quick 1.16.12 upload, with the attached diff? Yes, that looks okay. Thanks, uploaded. and flagged for acceptance. Regards, Adam -- To UNSUBSCRIBE, email to debian-release-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/1380656232.5700.12.ca...@jacala.jungle.funky-badger.org
Bug#724306: Bug #724306: pu: package dpkg/1.16.11
Hi! On Sat, 2013-09-28 at 08:13:29 +0100, Adam D. Barratt wrote: Control: tags -1 + pending On Sat, 2013-09-28 at 05:47 +0200, Guillem Jover wrote: On Thu, 2013-09-26 at 05:37:30 +0100, Adam D. Barratt wrote: On Thu, 2013-09-26 at 04:46 +0200, Guillem Jover wrote: On Tue, 2013-09-24 at 19:47:16 +0100, Adam D. Barratt wrote: This looks okay overall; thanks. I'm assuming that the changes have been tested on a stable system, particularly the Replaces. Yes. Let me know if and when you want this uploaded to the stable queue. Please feel free to go ahead. Done, thanks! Flagged for acceptance. Thanks, unfortunately 724949 just came in a day after the upload, it involves improper caching of the «dpkg --print-architecture» and «gcc -dumpmachine» output, affecting the performance of wanna-build. This was already fixed in 1.17.0, so it has been tested for a while. I was wondering if you'd be fine with a quick 1.16.12 upload, with the attached diff? (Just for future reference, would you have preferred a separate bug report?) Thanks, Guillem diff --git a/debian/changelog b/debian/changelog index a8c0c87..1f4d107 100644 --- a/debian/changelog +++ b/debian/changelog @@ -1,3 +1,10 @@ +dpkg (1.16.12) stable; urgency=low + + * Fix value caching in Dpkg::Arch by not shadowing the variables. +Closes: #724949 + + -- Guillem Jover guil...@debian.org Mon, 30 Sep 2013 16:52:37 +0200 + dpkg (1.16.11) stable; urgency=low [ Raphaël Hertzog ] diff --git a/scripts/Dpkg/Arch.pm b/scripts/Dpkg/Arch.pm index bfc19f4..bfee423 100644 --- a/scripts/Dpkg/Arch.pm +++ b/scripts/Dpkg/Arch.pm @@ -59,7 +59,7 @@ my %debarch_to_debtriplet; # dpkg-architecture itself, by avoiding computing the DEB_BUILD_ # variables when they are not requested. - my $build_arch = `dpkg --print-architecture`; + $build_arch = `dpkg --print-architecture`; syserr(dpkg --print-architecture failed) if $? 8; chomp $build_arch; @@ -75,7 +75,7 @@ my %debarch_to_debtriplet; { return $gcc_host_gnu_type if defined $gcc_host_gnu_type; - my $gcc_host_gnu_type = `\${CC:-gcc} -dumpmachine`; + $gcc_host_gnu_type = `\${CC:-gcc} -dumpmachine`; if ($? 8) { $gcc_host_gnu_type = ''; } else { commit dbe1c7762def447088c3d3a29eea0d7012af525f Author: Guillem Jover guil...@debian.org Date: Mon Sep 30 16:52:53 2013 +0200 Release 1.16.12 commit 8dafb822bb93de1ababd850360844986c9e0e900 Author: Guillem Jover guil...@debian.org Date: Tue Jan 1 19:30:36 2013 +0100 Dpkg::Arch: Fix value caching by not shadowing the variables Cherry picked from commit a64bfa733075a7140193f5a4b9d4292234dd230e. The effect of not caching the values has a severe impact on performance on code repeatedly calling (directly or indirectly) the get_raw_build_arch() and get_raw_host_arch() functions. Addresses Variables::ProhibitReusedNames. Closes: #724949
Bug#724306: Bug #724306: pu: package dpkg/1.16.11
On Mon, 2013-09-30 at 16:59 +0200, Guillem Jover wrote: On Sat, 2013-09-28 at 08:13:29 +0100, Adam D. Barratt wrote: Flagged for acceptance. Thanks, unfortunately 724949 just came in a day after the upload, it involves improper caching of the «dpkg --print-architecture» and «gcc -dumpmachine» output, affecting the performance of wanna-build. This was already fixed in 1.17.0, so it has been tested for a while. I was wondering if you'd be fine with a quick 1.16.12 upload, with the attached diff? Yes, that looks okay. (Just for future reference, would you have preferred a separate bug report?) Where the changes are distinct from those in the original report, separate bug are definitely preferable; so in this case, yes. Regards, Adam -- To UNSUBSCRIBE, email to debian-release-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/1380563835.29203.11.ca...@jacala.jungle.funky-badger.org
Bug#724306: Bug #724306: pu: package dpkg/1.16.11
On Mon, 2013-09-30 at 18:57:15 +0100, Adam D. Barratt wrote: On Mon, 2013-09-30 at 16:59 +0200, Guillem Jover wrote: Thanks, unfortunately 724949 just came in a day after the upload, it involves improper caching of the «dpkg --print-architecture» and «gcc -dumpmachine» output, affecting the performance of wanna-build. This was already fixed in 1.17.0, so it has been tested for a while. I was wondering if you'd be fine with a quick 1.16.12 upload, with the attached diff? Yes, that looks okay. Thanks, uploaded. (Just for future reference, would you have preferred a separate bug report?) Where the changes are distinct from those in the original report, separate bug are definitely preferable; so in this case, yes. Ok, sorry then, will do that next time. Regards, Guillem -- To UNSUBSCRIBE, email to debian-release-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20131001041149.ga4...@gaara.hadrons.org
Bug#724306: Bug #724306: pu: package dpkg/1.16.11
Control: tags -1 + pending On Sat, 2013-09-28 at 05:47 +0200, Guillem Jover wrote: On Thu, 2013-09-26 at 05:37:30 +0100, Adam D. Barratt wrote: On Thu, 2013-09-26 at 04:46 +0200, Guillem Jover wrote: On Tue, 2013-09-24 at 19:47:16 +0100, Adam D. Barratt wrote: This looks okay overall; thanks. I'm assuming that the changes have been tested on a stable system, particularly the Replaces. Yes. Let me know if and when you want this uploaded to the stable queue. Please feel free to go ahead. Done, thanks! Flagged for acceptance. Regards, Adam -- To UNSUBSCRIBE, email to debian-release-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/1380352409.22427.11.ca...@jacala.jungle.funky-badger.org
Bug#724306: Bug #724306: pu: package dpkg/1.16.11
On Thu, 2013-09-26 at 05:37:30 +0100, Adam D. Barratt wrote: On Thu, 2013-09-26 at 04:46 +0200, Guillem Jover wrote: On Tue, 2013-09-24 at 19:47:16 +0100, Adam D. Barratt wrote: Control: tags -1 + confirmed [...] This looks okay overall; thanks. I'm assuming that the changes have been tested on a stable system, particularly the Replaces. Yes. Let me know if and when you want this uploaded to the stable queue. Please feel free to go ahead. Done, thanks! Regards, Guillem -- To UNSUBSCRIBE, email to debian-release-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20130928034752.ga13...@gaara.hadrons.org
Bug#724306: Bug #724306: pu: package dpkg/1.16.11
Hi! On Tue, 2013-09-24 at 19:47:16 +0100, Adam D. Barratt wrote: Control: tags -1 + confirmed For some reason, this didn't make it to debian-release; that's usually related to attachment size, but they don't seem /that/ big... This has been the case with all my last dpkg requests, I guess there's something in them that bothers the mailing list setup… :/ This looks okay overall; thanks. I'm assuming that the changes have been tested on a stable system, particularly the Replaces. Yes. Let me know if and when you want this uploaded to the stable queue. Is there a plan for fixing #717983 in unstable in the near future? As things currently stand, 1.16.11 would have to be pushed in to testing as part of the point release. Yes, sorry about that, I didn't want to push an upload yet due to some problems with the commit mailing list, and because when manpages-it got fixed the problem became pretty much non-urgent. But on a second thought it might be blocking packages depending on 1.17.0 features, so I could have for example lowered the severity (even if temporarily). In any case I was planning on releaseing 1.17.2 in a couple of days anyway, so that should turn this into a non-issue. Thanks, Guillem -- To UNSUBSCRIBE, email to debian-release-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20130926024640.ga4...@gaara.hadrons.org
Bug#724306: Bug #724306: pu: package dpkg/1.16.11
On Thu, 2013-09-26 at 04:46 +0200, Guillem Jover wrote: On Tue, 2013-09-24 at 19:47:16 +0100, Adam D. Barratt wrote: Control: tags -1 + confirmed [...] This looks okay overall; thanks. I'm assuming that the changes have been tested on a stable system, particularly the Replaces. Yes. Let me know if and when you want this uploaded to the stable queue. Please feel free to go ahead. Regards, Adam -- To UNSUBSCRIBE, email to debian-release-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/1380170250.30209.58.ca...@jacala.jungle.funky-badger.org
Bug#724306: Bug #724306: pu: package dpkg/1.16.11
Control: tags -1 + confirmed For some reason, this didn't make it to debian-release; that's usually related to attachment size, but they don't seem /that/ big... On Mon, 2013-09-23 at 17:16 +0200, Guillem Jover wrote: dpkg (1.16.11) stable; urgency=low [ Raphaël Hertzog ] * Fix usage of non-existent _() function in multiple places of the Perl code. Thanks to Lincoln Myers linc...@netapp.com for the patch. Closes: #708607 [ Guillem Jover ] * Fix chmod() arguments order in Dpkg::Source::Quilt. Closes: #710265 Thanks to Pablo Oliveira pa...@sifflez.org. * Only ignore older packages if the existing version is informative. This allows any program using libdpkg to parse the available file to see again packages with versions lesser than 0-0 (like 0~0-0). Closes: #676664 * Fix use after free in dpkg_arch_load_list() on libdpkg. Reported by Pedro Ribeiro ped...@gmail.com. [ Updated programs translations ] * Vietnamese (Trần Ngọc Quân). Closes: #715334 [ Added man page translations ] * Italian (Beatrice Torracca). Closes: #711647 [ Updated man page translations ] * Japanese (TAKAHASHI Motonobu). Closes: #704240 This looks okay overall; thanks. I'm assuming that the changes have been tested on a stable system, particularly the Replaces. Is there a plan for fixing #717983 in unstable in the near future? As things currently stand, 1.16.11 would have to be pushed in to testing as part of the point release. Regards, Adam -- To UNSUBSCRIBE, email to debian-release-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/1380048436.17826.13.ca...@jacala.jungle.funky-badger.org