Update triggers for current state

Jake Bailey
2023-08-22 09:18:03 -07:00
parent 39647bc117
commit 8771215452

@@ -5,7 +5,6 @@ If you're a TS maintainer, you can respond to a PR with a comment similar to
to trigger a specialized on-demand build on the PR.
The currently recognized commands are:
* [`test this`](https://typescript.visualstudio.com/TypeScript/_build?definitionId=11) - This runs the internal RWC suite against the PR (this takes around 20 minutes). If the tests fail, a diff will be produced and a PR against the internal RWC suite repo will be made with the diff.
* [`run dt`](https://typescript.visualstudio.com/TypeScript/_build?definitionId=23) - The runs the definitely typed linter using the PR TS build sharded across 4 worker containers (this takes around 25 minutes).
* [`run dt slower`](https://typescript.visualstudio.com/TypeScript/_build?definitionId=18) - This is the same as the above, but only on a single worker (this takes around 90 minutes). This is useful if the results aren't needed promptly and the build queue is busy.
* [`user test this`](https://typescript.visualstudio.com/TypeScript/_build?definitionId=47) - This runs the nightly-tested `user` suite against the PR and against main (this takes around 30 minutes). The bot will post a summary comment comparing results from the two.
@@ -17,8 +16,8 @@ The currently recognized commands are:
* [`test tsserver top100`](https://typescript.visualstudio.com/TypeScript/_build?definitionId=47) - This runs the top 100 TypeScript repos on GitHub, by stars, against the PR and against main (this takes around 30 minutes). The bot will post a summary comment comparing results from the two. This variant tests tsserver, not tsc.
Note that `100` can be replaced with any other number up to 3 digits. For example, `test top200`, `test top50`, or `test top999` will all work.
* [`perf test`](https://typescript.visualstudio.com/TypeScript/_build?definitionId=22) - This queues a build on our perf server using the code from the PR - once started (which will only happen once any currently running build is done), this takes around 40 minutes. The bot should post the results of the perf test run back into the triggering PR once done.
* [`perf test faster`](https://typescript.visualstudio.com/TypeScript/_build?definitionId=45) - This is the same as the above, but only runs the node 12 host tests - so the results should be less complete, but come back in around 1/3rd the time.
* [`perf test`](https://typescript.visualstudio.com/TypeScript/_build?definitionId=69) - This queues a build on our perf server using the code from the PR - once started (which will only happen once any currently running build is done), this takes around 25 minutes. The bot should post the results of the perf test run back into the triggering PR once done.
* [`perf test faster`](https://typescript.visualstudio.com/TypeScript/_build?definitionId=69) - This is the same as the above, but only runs tsc tests.
* [`pack this`](https://typescript.visualstudio.com/TypeScript/_build?definitionId=19) - This creates a build which does a build, runs an LKG, runs normal tests, and then packs the result into an installable tarball (which can be downloaded from the build artifacts on the azure pipelines build status page), perfect for installing with `npm i <URL to tarball>` to test with.
* [`cherry-pick this to branchname`](https://typescript.visualstudio.com/TypeScript/_build?definitionId=30) - This launches a task to squash the commits from the PR and then open a new PR that cherry-picks the change into branch `branchname`. This takes about 5 minutes as the build agent needs to clone the input PR. The bot should reply if something goes wrong, or otherwise once the new PR is open.
* [`cherry-pick this to branchname and LKG`](https://typescript.visualstudio.com/TypeScript/_build?definitionId=30) - Same as above, but an LKG commit will be added onto the PR after the squashed cherry-pick commit.