Skip to content
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

[ci] Add 32bit job for 4.14.2 #1698

Merged
merged 3 commits into from
Oct 1, 2024
Merged

[ci] Add 32bit job for 4.14.2 #1698

merged 3 commits into from
Oct 1, 2024

Conversation

ejgallego
Copy link
Contributor

Let's see how far this gets, likely we will need some extra setup but depexts got much better in Opam 2.2.

@ejgallego

This comment was marked as outdated.

@ejgallego

This comment was marked as outdated.

@hhugo
Copy link
Member

hhugo commented Sep 30, 2024

You can rebase and should be able to drop the revert of targetint

@hhugo
Copy link
Member

hhugo commented Sep 30, 2024

Note that installing graphics locally with the correct multi-arch setup doesn't work either, so this is not due to missing i386 packages in the Ubuntu version on CI.

What is the error ?

@ejgallego
Copy link
Contributor Author

What is the error ?

Actually sorry, I had installed the wrong X11 package for 32bit , so many packages to take care of.

With this update, I think things are close to be ready, except for the test suite and the compilation failure on splay.ml

Feel free to push to the branch directly as I'm not sure how to fix these issues myself. The compilation failure is:

File "benchmarks/sources/ml/splay.ml", line 192, characters 47-57:
192 |     let s = (s + 0x7ed55d16 + (s lsl 12)) land 0xffffffff in
                                                     ^^^^^^^^^^
Error: Integer literal exceeds the range of representable integers of type int

Test suite I'm waiting for the full CI results as the failures might not be the same in the expect tests.

@ejgallego
Copy link
Contributor Author

ejgallego commented Sep 30, 2024

CI setup is IMO acceptable.

Test suite requires work on 2/3 cases, but beyond my expertise I'm afraid.

@hhugo
Copy link
Member

hhugo commented Sep 30, 2024

Why not add a job for OCaml 5.2 32bit ?

@ejgallego
Copy link
Contributor Author

ejgallego commented Sep 30, 2024

Why not add a job for OCaml 5.2 32bit ?

I was actually thinking of that, it is a good idea IMO.

I kind of refrained last minute as I have no JSOO + OCaml 5.x experience, so I thought it could be wise to make the 4.12+32bit CI work first.

I am myself still on 4.14.x due to lack of memprof-limits in 5.x series, which is critical to Coq coq-lsp users.

That doesn't matter for the JSOO version of Coq, however I still use 4.14 as my base dev branch. But I'm really looking forward to try JSOO 5.x with coq / coq-lsp / jsCoq.

@hhugo hhugo merged commit ef42a56 into ocsigen:master Oct 1, 2024
18 of 19 checks passed
@hhugo
Copy link
Member

hhugo commented Oct 1, 2024

I'm merging this one, feel free to open a new one to add 5.2 32bit to the ci

@ejgallego ejgallego deleted the 32ci branch October 1, 2024 12:05
@ejgallego
Copy link
Contributor Author

A problem with 5.2 + 32bit is that it is a byte only distribution, so JSOO will be a bytecode build, and thus quite slow.

Should we still go ahead with a byte-only CI?

I managed to compile coq-lsp with 5.2 + 32bit tho, but seems like this will become a dead end for us soon as coq-lsp requires ppx stuff, and for example, ppx_inline_test doesn't support 32bits anymore (time_now.0.17.0 fails to build).

But at least it works now!

@hhugo
Copy link
Member

hhugo commented Oct 1, 2024

Testing byte only with 5.2 seems to be a good addition

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants