-
-
Notifications
You must be signed in to change notification settings - Fork 14.8k
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
z3: 4.8.17 -> 4.13.4; z3_4_12: 4.12.5 -> 4.12.6 #327052
Conversation
Rebased. |
ofborg complains about a test failure on Mac:
|
Weird. Technically the only way I affected that build was adding the -j $NIX_BUILD_CORES to checkPhase... |
I'll try another approach, probably just going to disable the parallel test build on MacOS. |
Wait, that doesn't seem like the problem:
The actual test runner isn't even run in parallel! |
Presumably this problem is only for legacy versions. One would hope. 🙃 |
Well it still fails after 3 retries. It is odd that it still passes on Linux. |
Let's find out whether it's the ordering of test source files that's causing an only-on-Darwin segfault... |
Possible places something was fixed: |
Now I wonder if the previous success was spurious... |
I also maintain KLEE and STP, and use z3.
Original commit message by Reno Dakota (https://github.com/paparodeo), rebased against latest z3 changes: z3 contains a bunch of typos in never used functions. The update to clang-19 now reports the typos -- wrong member function / variable names. There is an upstream patch which fixes some of these: Z3Prover/z3@2ce89e5 And the others we patch or remove as the upstream code has since been modified.
Resolved issues with Z3 Haskell libraries. |
|
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.
Haskell stuff LGTM. Skimmed the rest.
Sorry for the endless back and forth!
No problem! |
Failures are unrelated to z3: isabelle:
tlaps is the same thing. ugarit:
|
I think this should be ready to merge and backport. |
This would need to be backported without the default version upgrade (4.8.17 -> 4.13.4) since it's ostensibly a breaking change. |
I think the only real changes have been soundness fixes and compiler compatibility. We could do a manual backport that doesn't switch the default, though. Someone is bound to be depending on it. |
@numinit But packages in nixpkgs were broken by the update, right? Thus we should assume that it'd also affect downstream packages. |
Yes, it was all due to the Z3_TRUE/Z3_FALSE header change, though (and that was just the Haskell ones). I do agree about keeping the default on 4.8 in 24.11 in case there are other things even though it ended up mostly being fine otherwise. |
Backport failed for Please cherry-pick the changes locally and resolve any conflicts. git fetch origin release-24.11
git worktree add -d .worktree/backport-327052-to-release-24.11 origin/release-24.11
cd .worktree/backport-327052-to-release-24.11
git switch --create backport-327052-to-release-24.11
git cherry-pick -x 2cb20cbfc26056e46a86bc72742c4d63332a7834 5a15eeca82b9c33772380699b4d4e50203783bde ef1388f5283f7ed759d8b2859ca24f2d56c2c759 fc7bc1d7ccde12e4e89b82eac63137f33100c59c 2ad2eedce2138d3cbff95c1db32d5881221304ad 9d557ee2d49a44e134110b333ddf6088f3b7b933 3af07aa4d35df5029509c353220b6801867c882d df9943568857ce91906c76a17c5d7d440376d1ad c556717c444259c8305c182ea6f93ea7b9251dd0 |
Thanks for the great work! |
Description of changes
Resolves #326120.
Things done
nix.conf
? (See Nix manual)sandbox = relaxed
sandbox = true
nix-shell -p nixpkgs-review --run "nixpkgs-review rev HEAD"
. Note: all changes have to be committed, also see nixpkgs-review usage./result/bin/
)Add a 👍 reaction to pull requests you find important.