- 29 Jan, 2018 1 commit
-
-
Mikaël Salson authored
Default to algo (which is similar to the former behaviour). The default rule was test_browser which was likely to fail as it required xvfb-run
-
- 12 Oct, 2017 3 commits
-
-
Mathieu Giraud authored
Completes c661b652. See #2630.
-
Mathieu Giraud authored
See #2630.
-
Mathieu Giraud authored
See #2630. We should check whether all rules are still used.
-
- 15 Sep, 2017 1 commit
-
-
Mathieu Giraud authored
See #2611.
-
- 28 Aug, 2017 3 commits
-
-
Mathieu Giraud authored
-
Mathieu Giraud authored
-
Mathieu Giraud authored
-
- 17 Mar, 2017 1 commit
-
-
Mikaël Salson authored
Follows 53be475d. Our test job uses the coverage so we need to remove -O0 from the compilation option even when we compute the coverage.
-
- 16 Mar, 2017 1 commit
-
-
Mikaël Salson authored
.should_get → .should-get
-
- 15 Mar, 2017 2 commits
-
-
Mathieu Giraud authored
-
Mathieu Giraud authored
-
- 13 Mar, 2017 2 commits
-
-
Mathieu Giraud authored
`README.org` is somewhat long, we may help people in finding what they are looking for. Completes ac1027f0, see #2237.
-
Mathieu Giraud authored
-
- 09 Mar, 2017 1 commit
-
-
Mikaël Salson authored
Closes #2237
-
- 31 Jan, 2017 1 commit
-
-
Mathieu Giraud authored
-
- 06 Jan, 2017 1 commit
-
-
Mikaël Salson authored
make browser (for instance) could not have been launched because the directory browser existed.
-
- 18 Oct, 2016 1 commit
-
-
Mikaël Salson authored
-
- 29 Sep, 2016 1 commit
-
-
Mathieu Giraud authored
When the global test suite is passing, individual failed tests (counted in SHOULD_*_EXPECTED_FAILS) can be marked as 'TODO' to make continuous integration happy
-
- 28 Sep, 2016 1 commit
-
-
Mathieu Giraud authored
-
- 24 Sep, 2016 5 commits
-
-
Mathieu Giraud authored
-
Mathieu Giraud authored
-
Mathieu Giraud authored
-
Mathieu Giraud authored
We prefer to have clean snapshots.
-
Mathieu Giraud authored
We run should-vdj-to-tap.py unbuffered (python -u) to have the best tee experience.
-
- 23 Sep, 2016 1 commit
-
-
Mikaël Salson authored
The *.data in germline subdirectories were not included.
-
- 05 Aug, 2016 2 commits
-
-
Mathieu Giraud authored
-
Mathieu Giraud authored
Commits ad0fe37e and 5815bf0d were very lazy, they are reverted. It is safer to still check all these tests, allowing some failed tests corresponding to recent or challenging .should-vdj sequences. Te results of the tests were carefully compared to the ones before ad0fe37e/5815bf0d. The difference (6 tests) come from some missing D segments due to 0c3cd0e.
-
- 13 Jul, 2016 2 commits
-
-
Mathieu Giraud authored
The other files are generated through a 'make' in tools/tests
-
Mathieu Giraud authored
Follows 7b955792
-
- 28 Jun, 2016 1 commit
-
-
Mikaël Salson authored
-
- 10 Jun, 2016 1 commit
-
-
Mathieu Giraud authored
(Same comment than ad0fe37e) These tests were extended with many sequences that should be further checked. Moreover they will be soon moved to their own git repository. We have to find a new way to measure the quality of the build even when some tests fail.
-
- 10 May, 2016 1 commit
-
-
- 04 Mar, 2016 1 commit
-
-
Mathieu Giraud authored
-
- 01 Mar, 2016 1 commit
-
-
Mathieu Giraud authored
-
- 04 Oct, 2015 2 commits
-
-
Mathieu Giraud authored
-
Mathieu Giraud authored
-
- 25 Sep, 2015 2 commits
-
-
Mikaël Salson authored
We don't want to re-generate the recombinations each time we launch the process. It would certainly be convenient to have the recombinations generated when the files do not exist. But I didn't succeed to specify a joker in the dependance list.
-
Mikaël Salson authored
Random recombinations with random deletions and insertions at the junction and mismatches in the sequence (1%)
-
- 09 Sep, 2015 1 commit
-
-
Mathieu Giraud authored
Since 43d2a3fe, errors in make_tools were ignored. The syntax is now ugly, but it works.
-