chore: merge protolude into postgrest#4861
Conversation
Co-authored-by: Adam Wespiser <adamwespiser@gmail.com>
* Updates README to correctly list all supported stack LTS versions * Updates README to include LTS 19.14 and higher
* Release 0.3.3, with support for GHC 9.4 - allow base-4.18 (GHC 9.4) - allow ghc-prim 0.10 - allow bytestring 0.11.4 - test against GHC 9.4.4 Further CI tweaks: - cabal: update GHC versions to the latest releases of the respective major version - cabal: remove non-latest minor versions - stack: add lts-20 (GHC 9.2.5) - stack: bring the list of GHC versions up to date * Update Cabal CI --------- Co-authored-by: Robert Vollmert <rob@vllmrt.net>
* Test 9.6 and 9.8 in CI * Support GHC 9.10 * Support GHC 9.12 --------- Co-authored-by: Tom Ellis <tom-git@jaguarpaw.co.uk>
|
Unfortunately this approach doesn't quite work, yet. When I browse your branch on GitHub, the history/blame for the protolude files only goes back 1 commit - the one where you added the files. Here's what I'd do:
|
Later we merge the src/protolude directory with src/PostgREST. Signed-off-by: Taimoor Zaeem <taimoorzaeem@gmail.com>
7a4adb2 to
827121b
Compare
Ignore blame for commit that moved protolude files under src/protolude. Signed-off-by: Taimoor Zaeem <taimoorzaeem@gmail.com>
Right, your approach looks good. Thank you for sharing this. 👍 Roadmap
|
|
Did a quick check: Browsing your branch on GitHub now correctly tracks commit information, I can easily blame my way through. Nice. |
Signed-off-by: Taimoor Zaeem <taimoorzaeem@gmail.com>
Signed-off-by: Taimoor Zaeem <taimoorzaeem@gmail.com>
Currently, our vendored protolude has many failing checks for style, lint and hsie. Temporarily excluding it. We should reinstate these checks later. Signed-off-by: Taimoor Zaeem <taimoorzaeem@gmail.com>
718c1e0 to
a46ac79
Compare
|
Looks like there is a GHC 9.12.3 bug that only shows up when we build PostgREST with |
|
I'm not sure whether it'd be helpful to go on with the coverage-based removal of protolude code first to simplify that code-base. Maybe we then have a better chance of diagnosing this problem? Other ideas to go on from here:
|
Not sure about that either, but I will try to look into it.
I confirmed the error with 9.12.4 in my fork. Bug reported: https://gitlab.haskell.org/ghc/ghc/-/issues/27227
On my system it builds just fine. I think the profiled build is probably not the problem because it also fails in the regular build in NixOS - MacOS build in CI. |
|
GHC team traced this issue to haskell/cabal#7684. Looks like PR: haskell/cabal#11791 has been opened to fix the issue 🎉 . |
Towards #4775.