Switch from exact snapshot to latest nightly#29
Merged
tbidne merged 11 commits intohaskell:masterfrom Mar 31, 2026
Merged
Conversation
The functionality was originally split into separate libraries for organization purposes. However this arguably does not buy us much, since the modules are already namespaces appropriately. Moreover, separate libraries were annoying as it slows down compilation/type checking, and makes HLS less effective.
Do not log terminal size error due to interference with optparse completions.
This allows us to annotate why a particular package is excluded.
We do this so minor version differences (i.e. different minor ghcs) do not cause failures, since ghc is in the stackage build plan.
Collaborator
Author
tbidne
force-pushed
the
dev
branch
3 times, most recently
from
4154165 to
fdebe3c
Compare
Bodigrim
approved these changes
Mar 27, 2026
Collaborator
Bodigrim
left a comment
Bodigrim
left a comment
There was a problem hiding this comment.
Very nice! Thank you!
I invited you to Haskell GitHub org with the intention to give you write access for this repo.
Collaborator
Author
|
Thanks! I accepted, though I think you'll also need to shuffle some permissions as I don't see an option to re-enable CI on the actions tab. |
Collaborator
|
Check again please. |
Collaborator
Author
|
It worked, thanks! I'll fix CI once I am back at my laptop (a few hours, most likely). |
tbidne
force-pushed
the
dev
branch
6 times, most recently
from
baaebeb to
e995918
Compare
Instead of pinning the snapshot to a specific date, we set it to 'nightly', which should allow more flexibility i.e. users can build whichever ghc corresponds to the latest snapshot. This should hopefully decrease the chance of users running into issues with ghc minor version upgrades e.g. bounds errors. Snapshot updates that upgrade the ghc major version will almost certainly require manual intervention here, but failures will likely be caught by the CI job that builds everything with --dry-run, so there should be some warning.
Collaborator
Author
|
CI is fixed, and I just built it again with the newest snapshot (9.12.4), so I'm going to go ahead and merge this. |
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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.
There are several updates here. First, there are some minor improvements (e.g. consolidating internal libs) and an update to ghc
9.12.3.More importantly, this implements the idea mentioned here, intending to resolve this. In more detail, this changes the snapshot from a specific date to just
nightly, meaning we always take the latest. The idea is that this should better allow users to runclc-stackagewith the latest ghc that corresponds to the current snapshot. There are a few changes that facilitate this:Boot packages added to
excluded_pkgs.jsonc(So e.g. we will not write an exact version constraint fortext, which could otherwise be a problem sinceghcis in the build plan). This is the main part that addresses the linked issue, as those are the most frequent pain points IME. As a proof-of-concept, I successfully built the package set here with9.12.2and9.12.3.Cabal
index-states dropped. Moreover,generated/cabal.projectincludes:which should mitigate other bounds issues.
There is a new CI job that builds everything with
--dry-run, using nix to supply the system dependencies.It's not perfect, but it is a pretty decent indication of how in-sync the current stackage nightly and package exclusion set here are.
If the job fails, there is very likely a real problem (nightly pulled in some package that needs to be excluded or has a new system dep). If the job succeeds, there is a good chance things are fine, though it is possible some new package fails to properly declare a new system dep at config time, so we wouldn't see a problem until someone tries to actually build it.
There is increased logging e.g. we try to warn the user when their ghc and the snapshot's do not match.
This is an alternative to this, in the sense that the new docs about ghc minor version should hopefully no longer be relevant. But @MangoIV please let me know if you think the docs (or anything else) can be improved.