Skip to content

Navigation Menu

Explore
By company size
By use case
By industry
View all solutions
Topics
- AI
- DevOps
- Security
- Software Development
- View all
Explore
- GitHub Sponsors
  Fund open source developers
- The ReadME Project
  GitHub community articles
Repositories
- Enterprise platform
  AI-powered developer platform
Available add-ons
Pricing

Search code, repositories, users, issues, pull requests...

Search

Clear

Search syntax tips

Provide feedback

We read every piece of feedback, and take your input very seriously.

Include my email address so I can be contacted

Saved searches

Use saved searches to filter your results more quickly

Name

Query

To see all available qualifiers, see our documentation.

You signed in with another tab or window. Reload to refresh your session. You signed out in another tab or window. Reload to refresh your session. You switched accounts on another tab or window. Reload to refresh your session.

Dismiss alert

bullet-train-co / bullet_train Public

Notifications You must be signed in to change notification settings
Fork 285
Star 1.7k

Code
Issues 59
Pull requests 19
Actions
Projects
Security
Insights

Additional navigation options

Code
Issues
Pull requests
Actions
Projects
Security
Insights

Improvements for CI in GitHub Actions#1787

Merged

jagthedrummer merged 189 commits intomainbullet-train-co/bullet_train:mainfrom jeremy/parallel_testsbullet-train-co/bullet_train:jeremy/parallel_testsCopy head branch name to clipboard

Dec 10, 2024

+527-53

Conversation Commits189 (189)Checks Files changed

Commits

Commits on Dec 6, 2024

install parallel_tests
jagthedrummer
committed
configure db connections for parallel_tests
jagthedrummer
committed
mostly working...
jagthedrummer
committed
Try using paralell_tests to split test files in CI
jagthedrummer
committed
run parallel_tests from the right directory
jagthedrummer
committed
don't try to use a log that's not there
jagthedrummer
committed
what's in tmp/starter
jagthedrummer
committed
typo
jagthedrummer
committed
try this
jagthedrummer
committed
moare debugging
jagthedrummer
committed
don't require a runtime file, just use it if/when we have one
jagthedrummer
committed
try recording runtime data
jagthedrummer
committed
maybe this
jagthedrummer
committed
apparently comments in the middle of the command confuses GHA
jagthedrummer
committed
linter
jagthedrummer
committed
try setting group correctly
jagthedrummer
committed
maybe this way
jagthedrummer
committed
linter
jagthedrummer
committed
linter
jagthedrummer
committed
what's in that log
jagthedrummer
committed
Group by runtime if we can, and ensure the runtime log is there
jagthedrummer
committed
Upload the runtime log artifact
jagthedrummer
committed
Start on combining runtime logs
jagthedrummer
committed
Add to the right pipeline, and actually combine the runtime logs
jagthedrummer
committed
Make it dependent on actually running tests
jagthedrummer
committed
rename the job
jagthedrummer
committed
give the job a name
jagthedrummer
committed
try using a cache to keep the runtime log around
jagthedrummer
committed
only restore the runtime log on test nodes
jagthedrummer
committed
combine files properly
jagthedrummer
committed
trigger ci
jagthedrummer
committed
show line count of combined log
jagthedrummer
committed
combine the logs to the right place
jagthedrummer
committed
trigger ci
jagthedrummer
committed
what if we run multiple parallel test runners on each CI node
jagthedrummer
committed
oops
jagthedrummer
committed
try this for parallel tests
jagthedrummer
committed
that didn't work...
jagthedrummer
committed
what if we don't specify db stuff so rigidly
jagthedrummer
committed
try to get db connections working with multiple runners per node
jagthedrummer
committed
specify localhost in CI
jagthedrummer
committed
4 groups, 2 per node
jagthedrummer
committed
what if we had twelve runners
jagthedrummer
committed
what does parallel:setup do without a number
jagthedrummer
committed
maybe 16 then
jagthedrummer
committed
try to get failures reported correctly
jagthedrummer
committed
maybe this will actually catch failures
jagthedrummer
committed
maybe this
jagthedrummer
committed
can't pass both...
jagthedrummer
committed
does this help
jagthedrummer
committed
what if this is causing problems
jagthedrummer
committed
maybe bundle exec always exits cleanly
jagthedrummer
committed
try simplifying this
jagthedrummer
committed
I bet that was it
jagthedrummer
committed
don't need that now
jagthedrummer
committed
maybe if we generate a fake secret...
jagthedrummer
committed
Fix a test for parallel use
jagthedrummer
committed
Fix for mailers in tests with parallel_test
jagthedrummer
committed
trigger ci
jagthedrummer
committed
trigger ci
jagthedrummer
committed
try reporting a single summary for all tests
jagthedrummer
committed
report summaries individually to make sure things match
jagthedrummer
committed
making a test fail to make sure the combined summary is reported correctly
jagthedrummer
committed
fix the failing test
jagthedrummer
committed
Combined summary for the internal pipeline
jagthedrummer
committed
Unique names for capturing summary artifacts
jagthedrummer
committed
try combining coverage data
jagthedrummer
committed
quote a string with a colon in it
jagthedrummer
committed
maybe this will work
jagthedrummer
committed
linter
jagthedrummer
committed
what's in the coverage directory
jagthedrummer
committed
have to explicitly enable hidden files
jagthedrummer
committed
try to get it working
jagthedrummer
committed
don't error out if the coverage dir isn't there yet
jagthedrummer
committed
need bundle exec
jagthedrummer
committed
try showing a summary of coverage on the PR
jagthedrummer
committed
try to point to the right thing
jagthedrummer
committed
make sure these have data in them
jagthedrummer
committed
what does the default formatter do
jagthedrummer
committed
knock down the minimum coverage
jagthedrummer
committed
try to get better info from the simplecov action
jagthedrummer
committed
knock down the per file coverage limit
jagthedrummer
committed
try using the official standardrb workflow
jagthedrummer
committed
try to get coverage data reporting more reliably
jagthedrummer
committed
try to keep minitest reporters from clobbering each other
jagthedrummer
committed
make the linter happy
jagthedrummer
committed
try to see if all the test reports are there
jagthedrummer
committed
is it more consistent with just one runner per node
jagthedrummer
committed
what if we just run one test node
jagthedrummer
committed
try two runner on one node
jagthedrummer
committed
maybe the brackwets cause problems
jagthedrummer
committed
try this for more parallelism
jagthedrummer
committed
not sure we need to serialize stdout
jagthedrummer
committed
try 12
jagthedrummer
committed
Where does the simplecov report end up if the internal workflow isn't running any jobs
jagthedrummer
committed
try this
jagthedrummer
committed
re-enable these
jagthedrummer
committed
what if we make our own summary
jagthedrummer
committed
derp
jagthedrummer
committed
ignore all report dirs created by parallel_tests
jagthedrummer
committed
Try conveying some actual info
jagthedrummer
committed
Try to add some file info
jagthedrummer
committed
fix
jagthedrummer
committed
clean it up a bit
jagthedrummer
committed
moar data
jagthedrummer
committed
what if we put the file table in a details block
jagthedrummer
committed
hmm
jagthedrummer
committed
maybe r needs to come first
jagthedrummer
committed
maybe this will do it
jagthedrummer
committed
clean up
jagthedrummer
committed
some clean up and polish
jagthedrummer
committed
more cleanup and polish
jagthedrummer
committed
upload coverage data for superscaffolding tests
jagthedrummer
committed
maybe we can generate two reports to get better data
jagthedrummer
committed
still need to require it
jagthedrummer
committed
Add group info
jagthedrummer
committed
oops
jagthedrummer
committed
only show 10 files
jagthedrummer
committed
can we dynamically build the matrix
jagthedrummer
committed
oops
jagthedrummer
committed
does this work
jagthedrummer
committed
try to calculate number of test runners per node dynamically based on repo visibility
jagthedrummer
committed
simplify the script a bit
jagthedrummer
committed
does this arrangement change anything
jagthedrummer
committed
Revert "does this arrangement change anything"
jagthedrummer
committed
Can we extract the matrix calculation into a workflow...
jagthedrummer
committed
are we getting output
jagthedrummer
committed
oops
jagthedrummer
committed
add runs-on
jagthedrummer
committed
that doesn't work
jagthedrummer
committed
env vars are always strings
jagthedrummer
committed
those too
jagthedrummer
committed
maybe this
jagthedrummer
committed
can we use a workflow from another repo
jagthedrummer
committed
maybe if we specify a branch
jagthedrummer
committed
try using it as an action, not a workflow
jagthedrummer
committed
fix
jagthedrummer
committed
need to add a ref
jagthedrummer
committed
try to connect the outputs
jagthedrummer
committed
slim things down a bit
jagthedrummer
committed
make sure things change as expected
jagthedrummer
committed
use the right variable name
jagthedrummer
committed
trigger CI
jagthedrummer
committed
some clean up
jagthedrummer
committed
use v1
jagthedrummer
committed
can't pass an empty with block apparently
jagthedrummer
committed
explicitly pass the number of runners per node to parallel setup
jagthedrummer
committed
centralize knowledge about where artifacts live in the action
jagthedrummer
committed
bug fix
jagthedrummer
committed
what if we add an environment
jagthedrummer
committed
nope
jagthedrummer
committed
break a couple of tests
jagthedrummer
committed
unbreak the tests
jagthedrummer
committed
rename this
jagthedrummer
committed
make the linter fail
jagthedrummer
committed
clean up
jagthedrummer
committed
make the linter happy again
jagthedrummer
committed
delete artifacts after we have them
jagthedrummer
committed
configure active storage for parallel tests
jagthedrummer
committed
delete artifacts last, so that retries are maybe possible
jagthedrummer
committed
try uploading the full coverage report as an artifact
jagthedrummer
committed
Add a convenience script for viewing the coverage report
jagthedrummer
committed
link to the coverage report in the summary
jagthedrummer
committed
Revert "Add a convenience script for viewing the coverage report"
jagthedrummer
committed
Add a readme to the report dir
jagthedrummer
committed
try to get things laying out better
jagthedrummer
committed
what's going on...
jagthedrummer
committed
try to get gem groups in core
jagthedrummer
committed

Commits on Dec 9, 2024

use the same path as bin/hack
jagthedrummer
committed
linter
jagthedrummer
committed
don't fail the workflow if deleting an artifact goes sideways
jagthedrummer
committed
can we get some coverage data about super scaffolding
jagthedrummer
committed
try to get better coverage info
jagthedrummer
committed
does this help
jagthedrummer
committed
that seems to cause problems
jagthedrummer
committed
hmm, maybe it's eager load and not the other thing
jagthedrummer
committed
linter
jagthedrummer
committed
maybe we should always upload artifacts
jagthedrummer
committed
that seems to prevent it from reporting sometimes
jagthedrummer
committed
always combine and clean up
jagthedrummer
committed
maybe a recycle symbol instead of the peek-a-boo guy
jagthedrummer
committed
oops
jagthedrummer
committed
we need this to fail if super scaffolding fails
jagthedrummer
committed

Commits on Dec 10, 2024

let's not do that for now
jagthedrummer
committed
some clean up
jagthedrummer
committed
remove the old parallel-ci script that we don't use anymore
jagthedrummer
committed
cleanup
jagthedrummer
committed
fix for grouping
jagthedrummer
committed
don't include avo in coverage reports
jagthedrummer
committed

Footer

© 2025 GitHub, Inc.

Footer navigation

Terms
Privacy
Security
Status
Docs
Contact

You can’t perform that action at this time.