rspamd: fix builds on non-x86_64-linux platforms #103973
Conversation
LuaJIT is built in rspamd only on x86_64-linux, and LuaJIT support became enabled by default in 2.6, breaking builds without it. This commit explicitly disables LuaJIT support on non-x86_64 architectures.
|
/marvin opt-in |
|
Hi! I'm an experimental bot. My goal is to guide this PR through its stages, hopefully ending with a merge. You can read up on the usage here. |
|
/status needs_reviewer |
|
@ajs124 this PR uses an experimental bot to accelerate code review. If you would merge this PR if you had permissions, write |
|
The PR author cannot set the status to If you are not the PR author and you are reading this, please review the usage of this bot. You may be able to help. Please make an honest attempt to resolve all outstanding issues before setting to |
There was a problem hiding this comment.
Change looks good, but the ofborg is still building, and might not finish because it seems to be building the world on aarch64:
https://logs.nix.ci/?attempt_id=f70c6e0c-25ad-4986-b237-685aa8e6ac5f&key=nixos%2Fnixpkgs.103973
@kisik21 What do I need to tell the Marvin if I do have merge permissions and would merge, but first want to give the ofborg the chance to finish the build (e.g. "merge if CI goes green")?
|
@timokau consider giving Marvin merge perms and an auto-merge command on ofborg approval? |
|
Also you could try to build this PR yourself - it's a fairly small change, and if you're on unstable it should build pretty quick :3 |
|
@kisik21 no, I consider that out of scope. Its not necessary for marvin's core functionality ("Making sure your PR gets a review and your reviews don't get lost.") and I don't want to have that kind of responsibility. A merge bot has been / is discussed a lot though, for example this might interest you. @nh2 it doesn't have any support for that right now. I could imagine a simply reminder functionality at some point in the future. I have thought about some ofBorg integration anyway (it could ask the PR author to fix ofBorg errors before requesting a reviewer). Reminders are pretty far down the backlog for now though. |
|
Since it was broken before and I assume @kisik21 has tested this locally, I'll ignore the running ofBorg check now. It has been queued for 6h already. Thanks for fixing this! |
@timokau @kisik21 My request isn't to allow the bot to auto-merge (that'd be nice for another day), but to address the fact that the usage doesn't describe what to do. It says:
so the first thing any commiter is thinking is "'does not have the permission to do so' doesn't apply to me, so what do I do now, well the document doesn't say, so I guess ... nothing?" If "nothing" is the right answer, it'd still be nice to write that into the doc. |
|
Marvin just augments the usual PR review process. So in cases where it can't help (yet), the usual process (or lack thereoff) applies. In my case that would be "keep a tab open for a while, then check if something might be wrong with ofBorg if the check still hasn't passed". Still, any contribution to or clarification of the contributing document is appreciated, feel free to open a PR :) |
Motivation for this change
Fixes #102751 - architectures other than
x86_64need some love too! (also, it's a package I use on my production machine and I finally figured out how to fix it)LuaJIT is built in rspamd only on x86_64-linux, and LuaJIT support became enabled by default in 2.6, breaking builds without it. This commit explicitly disables LuaJIT support on non-x86_64 architectures.
Things done
sandboxinnix.confon non-NixOS linux)nixpkgs#legacyPackages.x86_64-linux.rspamd.tests.rspamd.simplenixpkgs#legacyPackages.i686-linux.rspamd.tests.rspamd.simplenixpkgs#legacyPackages.aarch64-linux.rspamd.tests.rspamd.simplenix-shell -p nixpkgs-review --run "nixpkgs-review wip"./result/bin/)nix path-info -Sbefore and after)cc @avnik @fpletz @globin