diff options
author | Massimiliano Culpo <massimiliano.culpo@gmail.com> | 2020-05-12 22:56:01 +0200 |
---|---|---|
committer | GitHub <noreply@github.com> | 2020-05-12 13:56:01 -0700 |
commit | 11fa61665f2c921d6c9af82d9556e2985a2f3bd4 (patch) | |
tree | bc94815517c74ef1cd41d8b0751ca808dd3498b5 /.flake8_packages | |
parent | 6b41fb88a988cf35099ea5f5ecf7a3b42a5d7e1b (diff) | |
download | spack-11fa61665f2c921d6c9af82d9556e2985a2f3bd4.tar.gz spack-11fa61665f2c921d6c9af82d9556e2985a2f3bd4.tar.bz2 spack-11fa61665f2c921d6c9af82d9556e2985a2f3bd4.tar.xz spack-11fa61665f2c921d6c9af82d9556e2985a2f3bd4.zip |
travis: use bionic as default for Linux (#16521)
Modifications:
- [x] Travis now uses `bionic` as a default (`xenial` used for Python 3.5, `trusty` for Python 2.6)
- [x] Shell unit tests have been factored into their own run
- [x] `kcov` is built only for tests that upload coverage results
Overall with this we shave 3-4 mins. on each run and add an additional run of about 3 min. For some reason `kcov` 38 fails forwarding output when used with Python unit tests, so I used v34 for that and v38 (latest) for shell testing. Previously we were using v25.
Diffstat (limited to '.flake8_packages')
0 files changed, 0 insertions, 0 deletions