- 04 Jun, 2017 11 commits
-
-
Niels Thykier authored
-
Niels Thykier authored
Signed-off-by:Niels Thykier <niels@thykier.net>
-
Niels Thykier authored
-
Signed-off-by: -
Niels Thykier authored
Signed-off-by: -
Niels Thykier authored
This cuts dh_clean runtime in half as measured in the debhelper source package with 1000 dummy packages added to d/control.
-
Niels Thykier authored
It made dh skip "-arch" and "-indep" targets. Signed-off-by: -
Signed-off-by: -
This was accidentally removed when optimising a common case. Signed-off-by: -
Niels Thykier authored
Signed-off-by: -
Niels Thykier authored
Signed-off-by:
-
- 26 May, 2017 2 commits
-
-
Michael Biebl authoredSigned-off-by:
Michael Biebl <biebl@debian.org> -
Niels Thykier authored
Signed-off-by:
-
- 07 May, 2017 2 commits
-
-
Niels Thykier authored
Signed-off-by: -
Signed-off-by:
-
- 06 May, 2017 6 commits
-
-
Niels Thykier authored
Signed-off-by: -
Niels Thykier authored
Signed-off-by: -
Niels Thykier authored
Signed-off-by: -
Niels Thykier authored
Apparently, libtool+automake has a race condition in some cases, where automake will try to relink a binary about be installed before the library is installed. The bug is many years old and apparently not a priority, so work around it in debhelper. Signed-off-by: -
Niels Thykier authored
Signed-off-by: -
Niels Thykier authored
There is no point in loading a sequence that conditionally adds a tool that no longer exists. Just give a deprecation warning for people still explicitly loading it and drop the rest. Signed-off-by:
-
- 30 Apr, 2017 1 commit
-
-
Niels Thykier authored
Signed-off-by:
-
- 29 Apr, 2017 1 commit
-
-
Niels Thykier authored
It is redundant (the original package also has it) plus it can cause non-overridable lintian warnings. Signed-off-by:
-
- 28 Apr, 2017 1 commit
-
-
Niels Thykier authored
Signed-off-by:
-
- 25 Apr, 2017 1 commit
-
-
Niels Thykier authored
Signed-off-by:
-
- 17 Apr, 2017 7 commits
-
-
Niels Thykier authored
Mostly useful if a derivative needs to fiddle with the calling of pod2man or po4a. Concrete case being DilOS, which overrides which PERL is used for the call to these tools. Signed-off-by: -
Niels Thykier authored
Signed-off-by: -
Niels Thykier authored
Debian has nothing general patterns that must be chmod'ed to 0755, but DilOS needs "*.so" and "*.so.*" to be 0755 (instead of 0644). As it only takes 2-3 lines of code to mode the patterns out of the logic, then lets have an empty list, so DilOS's changes can be smaller and limited to some data lists. Signed-off-by: -
Niels Thykier authored
Should also assist DilOS and other derivatives by moving the list out of the code logic. Signed-off-by: -
Niels Thykier authored
As a side-effect, dh_fixperms should do fewer fork+exec's for two reasons. One, we skip the find call if the target directory does not exist and two, the patch reorders some directory handling with similar permission requirements together. Signed-off-by: -
Niels Thykier authored
Signed-off-by: -
Niels Thykier authored
Signed-off-by:
-
- 12 Apr, 2017 4 commits
-
-
Niels Thykier authored
install_dir does a eqv. test now so there is no reason to repeat it in dh_auto_install. Also, bulking the directories into one call will ensure we only do one fork+exec. Signed-off-by: -
Niels Thykier authored
Signed-off-by: -
Signed-off-by: -
Niels Thykier authored
Suggested-by:
Michael Biebl <biebl@debian.org>
Signed-off-by:
-
- 10 Apr, 2017 4 commits
-
-
Niels Thykier authored
Signed-off-by: -
Niels Thykier authored
Clarify that dh_systemd_enable --no-enable does *not* control whether the service is started. Signed-off-by: -
Niels Thykier authored
Signed-off-by: -
Niels Thykier authored
With "enough" packages, pkgfile is still a very expensive part of dh. Mostly due to the can_skip check which often need to call pkgfile to determine if a helper can be skipped. By doing a "bulk check" we can reduce that overhead by about roughly 50% (20-25% of dh's total runtime). Signed-off-by:
-
