After an evaluation, GNOME has moved from Bugzilla to GitLab. Learn more about GitLab.
No new issues can be reported in GNOME Bugzilla anymore.
To report an issue in a GNOME project, go to GNOME GitLab.
Do not go to GNOME Gitlab for: Bluefish, Doxygen, GnuCash, GStreamer, java-gnome, LDTP, NetworkManager, Tomboy.
Bug 789619 - Please use @basename@ to fix build reproducibility and multiarch conflict
Please use @basename@ to fix build reproducibility and multiarch conflict
Status: RESOLVED FIXED
Product: atk
Classification: Platform
Component: build
git master
Other Linux
: Normal normal
: ---
Assigned To: ATK maintainer(s)
ATK maintainer(s)
: 790337 (view as bug list)
Depends on:
Blocks:
 
 
Reported: 2017-10-29 14:24 UTC by Samuel Thibault
Modified: 2018-08-30 09:40 UTC
See Also:
GNOME target: ---
GNOME version: ---


Attachments
proposed fix (2.02 KB, patch)
2017-10-29 14:24 UTC, Samuel Thibault
accepted-commit_now Details | Review

Description Samuel Thibault 2017-10-29 14:24:33 UTC
Created attachment 362487 [details] [review]
proposed fix

Hello,

atk-enum-types.* use @filename@, but this includes the build path, and thus bring unreproducibility, and also conflicts when installing multiple builds for multiple archs (e.g. 64bit+32bit) in the same system.

The attached patch makes it use @basename@ instead, which only includes the base of the file name (this is available in glib since 2.21).

Samuel
Comment 1 Michael Catanzaro 2017-11-15 15:24:16 UTC
*** Bug 790337 has been marked as a duplicate of this bug. ***
Comment 2 Alejandro Piñeiro Iglesias (IRC: infapi00) 2017-11-15 16:03:38 UTC
(In reply to Michael Catanzaro from comment #1)
> *** Bug 790337 has been marked as a duplicate of this bug. ***

At least that duplicated served as a ping for the original one.

 (In reply to Samuel Thibault from comment #0)
> Created attachment 362487 [details] [review]
> proposed fix

Sorry, I missed this bug+patch.

> Hello,
> 
> atk-enum-types.* use @filename@, but this includes the build path, and thus
> bring unreproducibility, and also conflicts when installing multiple builds
> for multiple archs (e.g. 64bit+32bit) in the same system.

Ok. Makes sense.

> The attached patch makes it use @basename@ instead, which only includes the
> base of the file name (this is available in glib since 2.21).

My only concern is how this change was not already done on gtk+. As far as I understand this would affect both gtk+ and atk, as gtk+ also have some templates. As gtk+ is more popular, I was expecting the same issue already raised. But I don't see the equivalent bug for gtk+. Or Im being too paranoid and it is just that was detected first on atk?
Comment 3 Samuel Thibault 2017-11-15 16:45:26 UTC
gtk+3.0 is currently not reproducible for other reasons too, so I guess nothing was reported there just by luck and it just happens that I noticed it on the atk package first.
Comment 4 Alejandro Piñeiro Iglesias (IRC: infapi00) 2017-11-16 07:14:52 UTC
(In reply to Samuel Thibault from comment #3)
> gtk+3.0 is currently not reproducible for other reasons too, so I guess
> nothing was reported there just by luck and it just happens that I noticed
> it on the atk package first.

Fair enough.
Comment 5 Alejandro Piñeiro Iglesias (IRC: infapi00) 2017-11-16 07:16:46 UTC
Review of attachment 362487 [details] [review]:

I don't have too much experience with the templates, but lgtm
Comment 6 Samuel Thibault 2018-08-28 14:06:29 UTC
So, can it be commited? :)
Comment 7 Alejandro Piñeiro Iglesias (IRC: infapi00) 2018-08-30 09:14:08 UTC
(In reply to Samuel Thibault from comment #6)
> So, can it be commited? :)

Ups sorry. Usually when the reviewer sets it to accepted-commit_now status, the patch author commits it.

Do you need me to push the commit?
Comment 8 Samuel Thibault 2018-08-30 09:21:37 UTC
I still don't have a gnome account, so please do :)
Comment 9 Alejandro Piñeiro Iglesias (IRC: infapi00) 2018-08-30 09:40:59 UTC
(In reply to Samuel Thibault from comment #8)
> I still don't have a gnome account, so please do :)

Patch pushed. Thanks for the patch. But fwiw, next time I would appreciate a patch created by git format-patch, so applying it would be as easy as git am <patch>

Closing bug.