Closed mirekdlugosz closed 1 year ago
Merging #434 (69a34f9) into main (2bffdac) will not change coverage. Report is 1 commits behind head on main. The diff coverage is
n/a
.
@@ Coverage Diff @@
## main #434 +/- ##
=======================================
Coverage 79.11% 79.11%
=======================================
Files 5 5
Lines 249 249
=======================================
Hits 197 197
Misses 52 52
:mega: We’re building smart automated test selection to slash your CI/CD build times. Learn more
since we will only want to update playwright alongside jenkins, shouldn't we lock playwright on pyproject.toml as well to avoid future updates for other dependencies inadvertently also upgrading playwright?
I'm on the fence. I see your point, but:
poetry install
with poetry.lock
committed, which means that poetry will use exact versions we have in lockfile, even if newer versions are available in PyPI. If we version Playwright in pyproject, that means we can't only run poetry lock
, we have to also bump version in pyproject. Version in pyproject will also benefit any other package manager that understands pyproject format, but we don't care about these.All in all, I'm willing to take a risk and not version this explicitly. If this becomes a problem down the line, then we can fix it properly.
Good points! Leaving my 👍🏽 , then
This is done mainly for Playwright (UI tests).
Since today, Jenkins agent downloads Playwright browsers during image build. This speeds up pipeline execution a little (as we don't have to download large files).
Playwright browser versions are tied to Playwright version. As I have used newer Playwright when building Jenkins agent image, now pipeline still wants to download browsers - because our pinned Playwright is older, and requires older browsers. After this PR, it won't be necessary.
If we update Playwright version in the future, it's best to also update Jenkins agent image.