Bug #15973: APT pinning broken for stretch-backports

Bug #15973

APT pinning broken for stretch-backports

Added by segfault 2018-09-23 19:11:18 . Updated 2018-10-04 10:43:47 .

Status:

Resolved

Priority:

Normal

Assignee:

Category:

Target version:

Tails_3.9.1

Start date:

2018-09-23

Due date:

% Done:

100%

Feature Branch:

segfault:bugfix/15973-fix-backports-apt-pinning

Type of work:

Code

Blueprint:

Starter:

Affected tool:

Deliverable for:

Description

The APT pinning for stretch-backports is ignored, because we set the origin to “Debian” instead of “Debian Backports”. I see that this was done on purpose in 8efc9e33, because our snapshots repo uses “Debian” as its origin. But in config/chroot_local-includes/lib/live/config/1500-reconfigure-APT, we change the repo to the official Debian repo, which uses “Debian Backports” as the origin for backports.

I see two solutions: Use a different archive for backports in our repo, which uses “o=Debian Backports”. Or replace the origin for the backport pins in config/chroot_local-includes/lib/live/config/1500-reconfigure-APT.

Subtasks

Related issues

Blocked by Tails - ~~Bug #15837~~: APT is not configured to install packages from our repo	Resolved	2018-08-24
Blocks Tails - ~~Feature #15334~~: Core work 2018Q3: Foundations Team	Resolved	2018-02-20

History

#1 Updated by segfault 2018-09-23 19:11:42

Target version set to Tails_3.10.1

#2 Updated by segfault 2018-09-23 19:12:03

Assignee set to segfault

I would like to take this one.

#3 Updated by segfault 2018-09-23 19:46:20

Assignee changed from segfault to intrigeri
QA Check set to Ready for QA
Feature Branch set to bugfix/15973-fix-backports-apt-pinning

#4 Updated by segfault 2018-09-23 20:04:15

blocked by ~~Bug #15837~~: APT is not configured to install packages from our repo added

#5 Updated by segfault 2018-09-23 20:04:18

Feature Branch changed from bugfix/15973-fix-backports-apt-pinning to segfault:bugfix/15973-fix-backports-apt-pinning

I based the branch on bugfix/15837-fix-apt-preferences, so ~~Bug #15837~~ should be merged before this one.

#6 Updated by intrigeri 2018-09-28 08:15:38

Assignee changed from intrigeri to segfault
QA Check changed from Ready for QA to Dev Needed

I believe [a,n] does not do what you think it does: this “character class” will match any single char among “a”, “n” (so far so good)… and “,” (incorrect).

#7 Updated by intrigeri 2018-09-28 08:16:07

Status changed from Confirmed to In Progress
% Done changed from 0 to 50

#8 Updated by intrigeri 2018-09-28 08:24:20

This being said: good catch and thanks for cleaning up after me :)

#9 Updated by segfault 2018-09-29 09:18:24

Assignee changed from segfault to intrigeri
QA Check changed from Dev Needed to Ready for QA

intrigeri wrote:
> I believe [a,n] does not do what you think it does: this “character class” will match any single char among “a”, “n” (so far so good)… and “,” (incorrect).

Fixed in ec50ebb27690ee08f3b78ac7727edb23a35e2eb8.

#10 Updated by intrigeri 2018-09-30 14:20:12

% Done changed from 50 to 70

Code review passes, will now test!

#11 Updated by intrigeri 2018-09-30 14:23:47

blocks ~~Feature #15334~~: Core work 2018Q3: Foundations Team added

#12 Updated by intrigeri 2018-09-30 16:13:35

Assignee changed from intrigeri to segfault
QA Check changed from Ready for QA to Dev Needed

intrigeri wrote:
> Code review passes

Actually not: commit:d4bd6d00f5d912a0e701a1e849b85b7a2a2b4a1f introduces a regression in the “Custom APT repository” regexp. deb[.]tails[.]boum[.]org was correct but deb.tails.boum.org matches e.g. debXtailsYboumZorg, which it should not.