This PR implements a number of improvements for our Travis CI: #361
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
added 8 commits
December 21, 2016 15:16
This prevents Travis log collector to add separate expansion marks to the echo output and the actuall log output.
this script is intended only for use in Travis CI and contains
configuration of the test run requested:
* it can run linter step separately by specifying TASK_TO_RUN="lint"
environment variable in .travis.yml. In this case it also runs
pep8 checker on the commits in PR.
* other steps are run in developer mode in order to skip pylint run
and speed up the task
* in all cases the CI result log is populated and can be displayed
if the job fails
In order to speed our Travis CI gating even further, the lint step has been split to a separate job that can be run in parallel with the test runs. The test runs are in turn launched in developer mode to speed them up.
This is a preparatory work for the future requirement of running Python2/3 jobs simultaneously.
|
Bump for review. |
frasertweedale
approved these changes
Jan 3, 2017
|
I assume the licence headers did not break the automember tests so this could be pushed. Just a brief question: would it be reasonable to get the line number of "========= FAILURES =========" and tail the "$CI_TRAVIS_LOG" from the end to it? |
|
@stlaz thanks for an idea, that will actually make the log output much more useful. I will try to think about it. |
|
@martbab My naive solution is to do something like LINE=`grep -n -m 1 $CI_TRAVIS_LOG -e "=== FAILURES ===" | cut -d: -f1`
LINES=`wc -l $CI_TRAVIS_LOG`
tail -n `expr $LINES - $LINE` $CI_TRAVIS_LOG |
|
@stlaz I have implemented a simple log trimming which keeps only pytest failures if present. The original behavior is kept as a fallback for the case if the setup fails. |
martbab
force-pushed
the
travis-ci-improvements
branch
from
4b4f52c
to
b2b46d0
Compare
|
The change LGTM, ACK, we'll see how it works :) |
stlaz
added
ack
stlaz
reviewed
Jan 5, 2017
| function truncate_log_to_test_failures() { | ||
| # chop off everything in the CI_RESULTS_LOG preceding pytest error output | ||
| # if there are pytest errors in the log | ||
| $error_fail_regexp='\(=== ERRORS ===\)\|\(=== FAILURES ===\)' |
There was a problem hiding this comment.
Should $error_fail_regexp miss "$"?
There was a problem hiding this comment.
Oops sorry I forgot how to bash.
stlaz
reviewed
Jan 5, 2017
| # if there are pytest errors in the log | ||
| $error_fail_regexp='\(=== ERRORS ===\)\|\(=== FAILURES ===\)' | ||
|
|
||
| if grep -e "$error_fail_regexp" $CI_RESULTS_LOG |
If we get to the `run-tests` phase, we no longer care about messages from previous steps so we can truncate the log output to only show pytest failures/errors, reducing log size. As a fallback, the previous behaviour to print last 5000 log lines was kept in the case the CI tests fail during setup.
martbab
force-pushed
the
travis-ci-improvements
branch
from
b2b46d0
to
39a784e
Compare
|
I have no more remarks on this, hopefully final ACK. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
split out the test runner part into a standalone script, .travis.yml should
now only define test matrix, set environment variables and process output
after failure
mark the project as Python one, implement support for running builds using
different python version (future-proofing against incoming Python2/3 CI) and
cache job dependencies
use separate job for pep8/linters. This shaves off ca 6-8 minutes from
overall build time. You should get CI results in 26 min compared to previous
33 min