Add macos-14 (arm64) CI (and made it pass finally)

[cielavenir]

• Changed macos-latest CI to macos-13 because macos-latest points 12
• Added macos-14 CI, which is arm64
• Fixed arm64 mach compilation again (cf https://github.com/intel/isa-l/pull/226)
• Fixed isal_deflate_icf_finish_lvl1 dispatcher (sorry) (why no one noticed that...)
• Avoid using x18 register

CI passed according to https://github.com/cielavenir/isa-l/actions/runs/8114714856

[cielavenir]

Extended tests passed as well https://github.com/cielavenir/isa-l/actions/runs/8130749368/job/22219331109

[pablodelara]

HI @cielavenir. Thanks for your PR! Can you clean it up a little bit? There are a bunch of "merge" and "revert" patches that are polluting the PR. Also, since you have integrated the fix in PR #226, can you close that PR?

[cielavenir]

@pablodelara Squashed

[cielavenir]

by the way, not sure I'm noob at github, but how can I run CI without patching on: field? I'm asking because this caused pushing back and forth.

[pablodelara]

Thanks @cielavenir! What I meant is to keep only the commits that are adding value overall, and remove the "merge" commits. Looking at the PR, you could have 4-5 commits, each one adding/fixing some functionality. E.g.:

• Add macos14
• Fix mach compilation on MacOS
• Avoid x18 (register?)
• Fixed isal_deflate_icf_finish_lvl1 dispatcher

[cielavenir]

Actually is it possible to merge #226 and #280 then rebase this pull request again?

[pablodelara]

@cielavenir Well, those PR's are closed. We could reopen them, but still this PR should be split into several commits.

[cielavenir]

right please reopen and merge them

[cielavenir]

@pablodelara @liuqinfei Now rebased again. The thing is, #226 and #280 are obvious but avoiding x18 register should be reviewed carefully as I rely only on CI result.

[pablodelara]

@liuqinfei can you review this PR?

[liuqinfei]

yes.

---- Replied Message ---- | From | Pablo de @.> | | Date | 03/14/2024 21:19 | | To | @.> | | Cc | @.>@.> | | Subject | Re: [intel/isa-l] Add macos-14 (arm64) CI (and made it pass finally) (PR #278) |

@liuqinfei can you review this PR?

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.Message ID: @.***>

[liuqinfei]

I think the change is ok for Aarch64 platform. And, i verified this PR on kunpeng 920 (Aarch64), all the test cases pass as following.

[cielavenir]

rebased again.

now let's see if macOS M1 extended tests pass.

[pablodelara]

This is merged now, thanks!