mflatt
2020-7-29 11:02:21

I think CI will not break, but I patched things only minimally by removing the ChezScheme checkout.


mflatt
2020-7-29 11:03:25

Of course, the CI files for ChezScheme-as-subtree are in the subtree, so that part won’t run until it’s moved and updated to work with a subtree.


mflatt
2020-7-29 11:05:08

By “I think CI will not break”, I really mean “I’ve been taking advantage of the existing CI setup to check my changes.”


mflatt
2020-7-29 11:09:59

They way pb boot-file management turned out, I think it’s useful to CI and worth testing. Use make fetch-pb to download pb to the racket/src/ChezScheme/boot directory, and drop the --enable-racket= when running configure for CS. Meanwhile, it would be good to have one build that still uses --enable-racket= to check that build path; the supplied Racket just needs to be v7.1 or later, but it could be a just-built Racket (CS or BC).


samth
2020-7-29 13:45:07

Do you want to test a deliberately-older Racket?


pocmatos
2020-7-29 14:18:51

Lets move all the builds except one to use pb instead of --enable-racket. We can use --enable-racket on one for one of the fastest linux builds just to make sure it works. I will port all the others to using pb .


pocmatos
2020-7-29 14:19:04

It’s great because they don’t need to depend anymore on the cgc build.


pocmatos
2020-7-29 14:19:32

I am experimenting with CI behind vagrant to test on FreeBSD.


pocmatos
2020-7-29 14:20:49

Waiting for a bug in docker to be fixed so we can get armv7 builds working.


pocmatos
2020-7-29 14:21:15

The other thing that’s ongoing is my experimentation with scribble/text to generate the yml files.


pocmatos
2020-7-29 14:21:40

I have been busy with work but I finished a big task today so things are relatively calm. Will use this time to push these changes forward.


pocmatos
2020-7-29 14:36:02

@mflatt Are you bringing in the subtree of Chez with --squash ?


pocmatos
2020-7-29 14:37:02

If Chez’s dev continues in racket/ChezScheme, I assume we don’t need the whole history of Chez inside Racket.


mflatt
2020-7-29 14:37:02

No — preserving history, but pruning boot files


pocmatos
2020-7-29 14:37:49

Is there any advantage in keeping the history of Chez inside the racket repo?


mflatt
2020-7-29 14:38:30

I think Chez Scheme development for Racket will continue in the racket repo. Probably with occasional exports to ChezScheme, but only because I’m reluctant to give that repo up.


pocmatos
2020-7-29 14:38:47

ah - perfect.


pocmatos
2020-7-29 14:38:54

Well - that’s even better then. :slightly_smiling_face:


pocmatos
2020-7-29 14:39:05

Thanks!!!


pocmatos
2020-7-29 14:41:17

So we also need to bring in the Chez workflows to the racket repo.


pocmatos
2020-7-29 14:41:35

So that we keep testing Chez separately from racket.


mflatt
2020-7-29 14:43:40

Yes. At first, there will still be a “.github” in “racket/src/ChezScheme” whose content needs to be moved up and triggered by changes to the subtree.


pocmatos
2020-7-29 14:45:27

Yes - I will sort that out as soon as you do the subtree creation