-
Notifications
You must be signed in to change notification settings - Fork 105
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Use commit timestamp #63
Conversation
Please ping me when this is passing tests and ready to review! |
2c3f318
to
dcdf8e5
Compare
OK, other than the 'bazar' typo this looks excellent! Is it ready to merge? |
35f1455
to
8a54147
Compare
@aspiers Do you understand why this fails? Travis reports that python-dateutil is installed successfully ... |
@jblunck I have no idea :-( |
Any news on this? Not sure about the failure. looks like python-2.6 is used but later python 2.7 is referenced? |
@olafhering Good catch - that's exactly the problem. The bug was introduced by #73 but sadly Travis CI didn't catch it at the time; the Travis environment happened to have Python 2.7, but when it was accidentally used it worked due to lack of any external module dependencies. |
Proposed solution in #82. Any volunteers? :) |
#82 is now fixed; please can you rebase this? Thanks! |
@jblunck Do you think you'll be able to rebase this? |
This is the required preparation of the framework for the tests that check tar_scm for always creating a binary identical tarball from the same commit. Signed-off-by: Jan Blunck <jblunck@infradead.org>
Otherwise the tarball will not be binary identical even if it is created from the same source tag. In some cases this will lead to unnecessary rebuilds. Signed-off-by: Jan Blunck <jblunck@infradead.org>
8a54147
to
b6682e3
Compare
Mercurial has a bug in the localdate filter that makes it apply the current offset from UTC to historic timestamps for timezones that have daylight savings enabled. Signed-off-by: Jan Blunck <jblunck@infradead.org>
b6682e3
to
8833bd4
Compare
Again, I have no clue why I didn't notice this was rebased. Maybe there is some issue with how I receive github notifications, but it seems to work fine everywhere else. Anyway sorry for the delay, and thanks a lot for your great contribution! |
Unfortunately Travis CI missed this because it deliberately doesn't retest the updated merge commit GitHub generates when a PR is updated: https://docs.travis-ci.com/user/pull-requests#My-Pull-Request-isn%E2%80%99t-being-built
This branch depends on fix_hgtest branch. Additionally to that I'm going to refactor the additional unittest after merging the fix_include_exclude branch.
For an enterprise setup it is important that the generated tarballs are binary identical if they are generated from the same source revision. This commit makes tar_scm use the commit timestamp as a reference for the mtime of the files included in the tarball. The unittest is using a fixed, well known value instead to check that the mtime is taken from the source repository.