-
-
Notifications
You must be signed in to change notification settings - Fork 367
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Wingman: Low gas warning #2038
Wingman: Low gas warning #2038
Conversation
@pepeiborra the dynamically linked builds are now failing in CI with an undefined linker reference to my new constructor. Usually when I see this it means my cabal file's modules are not up to date, but that doesn't appear to be the case here. Any idea what's up? |
"Linking" the tactics test suite fails to find the symbol for the new constructor. I'm not sure what "linking" means here since we are asking Cabal to use dynamic linking for executables, it could be some sort of validation step, or TH if any. We are doing dynamic linking with stack: this is probably not well trodden and involves telling cabal to link haskell packages dynamically (for executables) without stack being aware. What could be going wrong here? If you have the time, please put together a minimal repro and send it upstream to stack or wherever it belongs. In terms of fixing it, there's a couple of options:
I don't have time to investigate any further, hopefully you or any of the usual maintainers can take a look! |
For anyone who might be reading the thread:
This is what appears in CI, and in my local builds:
|
IMO this is a non-option as it makes CircleCI be a serious bottleneck of PR processes. When the cache had a problem, CircleCI took about 5 hours, and we often have more than 3 somewhat active PRs, so the CI process will take about a day just by itself if we disable the cache. |
Hmm i think the problem could be caused by @isovector could you test locally if |
Let see how much time cost build local packages from scratch: jneira#25 |
Ok, with the actual full cache (https://app.circleci.com/pipelines/github/jneira/haskell-language-server/642/workflows/0dfabe4c-3b34-49a3-8df6-e799103089db):
Caching only deps (https://app.circleci.com/pipelines/github/jneira/haskell-language-server/644/workflows/f66595fc-2e5e-46c5-b549-842e59340e84):
Times are not uniform in the 2nd case so maybe it could be closed to full cache. |
The numbers look good @jneira, please send a PR to merge it |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
thanks for your phenomal work on wingman 🙂
Several jobs were stalled since 17 hours with the message:
As reported by @Ailrun in some pr i dont find out right now |
At least the testing workflow of this one has started to run 🤞 |
Make Wingman inform you if it ran out of gas when attempting to synthesize an expression, rather than just failing with a "couldn't be solved" error message.
Feature request from @masaeedu.