[arch-dev-public] pacman-4.1 - Arch-dev-public - lists.archlinux.org

newer
[arch-dev-public] Integrity Check...

[arch-dev-public] pacman-4.1

older
[arch-dev-public] Signoff report...

Allan McRae

1 Apr 2013 1 Apr '13

6:39 a.m.

Hi, Pacman-4.1.0 is released and in the [testing] repos. See my blog [1] for details of what is new in this release. Cheers, Allan [1] http://allanmcrae.com/2013/04/pacman-4-1-released/

Show replies by date

Rémy Oudompheng

1 Apr 1 Apr

12:26 p.m.

2013/4/1 Allan McRae :

Hi,

Pacman-4.1.0 is released and in the [testing] repos. See my blog [1] for details of what is new in this release.

Cheers, Allan

[1] http://allanmcrae.com/2013/04/pacman-4-1-released/

I have updated pyalpm accordingly. Thanks for the release blog post! It's been a while since I haven't followed pacman development and I am thrilled to see such transformations happening. Rémy.

Sébastien Luttringer

4 Apr 4 Apr

9:42 p.m.

On Mon, Apr 1, 2013 at 8:39 AM, Allan McRae wrote:

Hi,

Pacman-4.1.0 is released and in the [testing] repos. See my blog [1] for details of what is new in this release.

With this commit[1], can we assume that install script should now be written in bash? [1] https://projects.archlinux.org/svntogit/packages.git/commit/trunk?h=packages/pacman&id=b41b136a374fc85c559e06120b5dc82cd4d5aa28 -- Sébastien "Seblu" Luttringer https://www.seblu.net GPG: 0x2072D77A

Allan McRae

10:25 p.m.

On 05/04/13 07:42, Sébastien Luttringer wrote:

On Mon, Apr 1, 2013 at 8:39 AM, Allan McRae wrote:

...
Hi,

Pacman-4.1.0 is released and in the [testing] repos. See my blog [1] for details of what is new in this release.

With this commit[1], can we assume that install script should now be written in bash?

[1] https://projects.archlinux.org/svntogit/packages.git/commit/trunk?h=packages/pacman&id=b41b136a374fc85c559e06120b5dc82cd4d5aa28

What did you think it was written in previously? We have always used bashisms. Allan

Sébastien Luttringer

5 Apr 5 Apr

9:36 a.m.

On 05/04/13 07:42, Sébastien Luttringer wrote:

...
On Mon, Apr 1, 2013 at 8:39 AM, Allan McRae wrote:

...
Hi,

Pacman-4.1.0 is released and in the [testing] repos. See my blog [1] for details of what is new in this release.

With this commit[1], can we assume that install script should now be written in bash?

[1]

https://projects.archlinux.org/svntogit/packages.git/commit/trunk?h=packages/pacman&id=b41b136a374fc85c559e06120b5dc82cd4d5aa28

...
What did you think it was written in previously? We have always used bashisms.

Posix shell. Pacman starts /bin/sh, a symlink to bash, which switch it in

On Fri, Apr 5, 2013 at 12:25 AM, Allan McRae wrote: posix mode. Actually, this mode doesn't disallow bashism, it modify some minor behaviour of bash. I also remember when I was a TU padawan, a jedi master, ask me to avoid bashism in install scripts because /bin/sh can be modified by user to something else and we should be posix as possible. I never checked that before but we have currently at minimum 50 install scripts with bashism. I will be more relax writing my install scripts :) Thanks for clarification. -- Sébastien "Seblu" Luttringer https://www.seblu.net GPG: 0x2072D77A

Allan McRae

9:47 a.m.

On 05/04/13 19:36, Sébastien Luttringer wrote:

I never checked that before but we have currently at minimum 50 install scripts with bashism.

Every single install script for info files is non "sh" compatible, as users of dash as /bin/sh will testify to...

Thomas Bächler

8:34 a.m.

Am 01.04.2013 08:39, schrieb Allan McRae:

Hi,

Pacman-4.1.0 is released and in the [testing] repos. See my blog [1] for details of what is new in this release.

Cheers, Allan

[1] http://allanmcrae.com/2013/04/pacman-4-1-released/

The new default CPPFLAGS seem to break configure. cpp complains that -D_FORTIFY_SOURCE=2 cannot be used without -O.

Allan McRae

8:57 a.m.

On 05/04/13 18:34, Thomas Bächler wrote:

Am 01.04.2013 08:39, schrieb Allan McRae:

...
Hi,

Pacman-4.1.0 is released and in the [testing] repos. See my blog [1] for details of what is new in this release.

Cheers, Allan

[1] http://allanmcrae.com/2013/04/pacman-4-1-released/

The new default CPPFLAGS seem to break configure. cpp complains that -D_FORTIFY_SOURCE=2 cannot be used without -O.

I thought it fixed issues (e.g. curl). Or are we both just taking one package and wildly generalising...?

Thomas Bächler

9:09 a.m.

Am 05.04.2013 10:57, schrieb Allan McRae:

On 05/04/13 18:34, Thomas Bächler wrote:

...
Am 01.04.2013 08:39, schrieb Allan McRae:

...
Hi,

Pacman-4.1.0 is released and in the [testing] repos. See my blog [1] for details of what is new in this release.

Cheers, Allan

[1] http://allanmcrae.com/2013/04/pacman-4-1-released/

The new default CPPFLAGS seem to break configure. cpp complains that -D_FORTIFY_SOURCE=2 cannot be used without -O.

I thought it fixed issues (e.g. curl).

Or are we both just taking one package and wildly generalising...?

We are. This is what I got: configure:6470: checking how to run the C preprocessor configure:6501: gcc -E -D_FORTIFY_SOURCE=2 conftest.c In file included from /usr/include/limits.h:26:0, from /usr/lib/gcc/x86_64-unknown-linux-gnu/4.8.0/include-fixed/limits.h:168, from /usr/lib/gcc/x86_64-unknown-linux-gnu/4.8.0/include-fixed/syslimits.h:7, from /usr/lib/gcc/x86_64-unknown-linux-gnu/4.8.0/include-fixed/limits.h:34, from conftest.c:13: /usr/include/features.h:330:4: warning: #warning _FORTIFY_SOURCE requires compiling with optimization (-O) [-Wcpp] # warning _FORTIFY_SOURCE requires compiling with optimization (-O) ^ configure:6501: $? = 0 configure: failed program was: | /* confdefs.h */ | #define PACKAGE_NAME "shellinabox" | #define PACKAGE_TARNAME "shellinabox" | #define PACKAGE_VERSION "2.10" | #define PACKAGE_STRING "shellinabox 2.10" | #define PACKAGE_BUGREPORT "markus@shellinabox.com" | #define PACKAGE_URL "" | #define VCS_REVISION "239" | #define PACKAGE "shellinabox" | #define VERSION "2.10" | /* end confdefs.h. */ | #ifdef __STDC__ | # include | #else | # include | #endif | Syntax error Confusing here is that there is no error, just a warning, but it still starts retrying different preprocessor command and eventually fails. Appending -O2 to the CPPFLAGS fixed the build.

Allan McRae

10:38 a.m.

On 05/04/13 19:09, Thomas Bächler wrote:

Am 05.04.2013 10:57, schrieb Allan McRae:

...
On 05/04/13 18:34, Thomas Bächler wrote:

...
Am 01.04.2013 08:39, schrieb Allan McRae:

...
Hi,

Pacman-4.1.0 is released and in the [testing] repos. See my blog [1] for details of what is new in this release.

Cheers, Allan

[1] http://allanmcrae.com/2013/04/pacman-4-1-released/

The new default CPPFLAGS seem to break configure. cpp complains that -D_FORTIFY_SOURCE=2 cannot be used without -O.

I thought it fixed issues (e.g. curl).

Or are we both just taking one package and wildly generalising...?

We are. This is what I got:

configure:6470: checking how to run the C preprocessor configure:6501: gcc -E -D_FORTIFY_SOURCE=2 conftest.c In file included from /usr/include/limits.h:26:0, from /usr/lib/gcc/x86_64-unknown-linux-gnu/4.8.0/include-fixed/limits.h:168, from /usr/lib/gcc/x86_64-unknown-linux-gnu/4.8.0/include-fixed/syslimits.h:7, from /usr/lib/gcc/x86_64-unknown-linux-gnu/4.8.0/include-fixed/limits.h:34, from conftest.c:13: /usr/include/features.h:330:4: warning: #warning _FORTIFY_SOURCE requires compiling with optimization (-O) [-Wcpp] # warning _FORTIFY_SOURCE requires compiling with optimization (-O) ^ configure:6501: $? = 0 configure: failed program was: | /* confdefs.h */ | #define PACKAGE_NAME "shellinabox" | #define PACKAGE_TARNAME "shellinabox" | #define PACKAGE_VERSION "2.10" | #define PACKAGE_STRING "shellinabox 2.10" | #define PACKAGE_BUGREPORT "markus@shellinabox.com" | #define PACKAGE_URL "" | #define VCS_REVISION "239" | #define PACKAGE "shellinabox" | #define VERSION "2.10" | /* end confdefs.h. */ | #ifdef __STDC__ | # include | #else | # include | #endif | Syntax error

Confusing here is that there is no error, just a warning, but it still starts retrying different preprocessor command and eventually fails. Appending -O2 to the CPPFLAGS fixed the build.

Yes... this is my pet hate - people who use -Werror and expect their code to be magically immune to any future warnings. And this is even more annoying because it is quite hidden... Fix: sed "s#AC_LANG_WERROR##" configure.ac autoreconf -i Allan

4053

Age (days ago)

4057

Last active (days ago)

Download

9 comments

4 participants

tags

participants (4)

Allan McRae
Rémy Oudompheng
Sébastien Luttringer
Thomas Bächler