Refactoring CI regarding ARM binary publish #2136
Conversation
curquiza
force-pushed
the
refacto-arm-ci
branch
from
444986f
to
78cf8f1
Compare
| ## Environment variable is not passed using env: | ||
| ## LD gold won't work with MUSL | ||
| # env: | ||
| # JEMALLOC_SYS_WITH_LG_PAGE: 16 |
There was a problem hiding this comment.
Removing that line for the ARM build won’t be a problem?
There was a problem hiding this comment.
Which lines? These lines were already commented in the other file
There was a problem hiding this comment.
I am not sure but I would like that first command that returns something different than zero should make the script abort or at least we should check that curl doesn't return something different than 0. It would help in not having Not Found in the final meilisearch file as it can be dangerous, we chmod u+x it and a simple bash script can be returned by a middleware or something... We should abort before chmoding the file.
Co-authored-by: Clément Renault <clement@meilisearch.com>
This is expected, when using a M1, the downloaded binary should be x86_64. Maybe I'm wrong, but @tamo confirmed the aarch64 binary did not work on M1. Because, if aarch64 works on M1, I can make the script simpler. Can you test aarch64 binary on your mac @Kerollmops? |
|
Indeed, aarch64 can't work on macOS as it is an ELF binary file, not a Mach-O one. If running the |
curquiza
force-pushed
the
refacto-arm-ci
branch
from
33ebadc
to
c4a2d70
Compare
| arch: aarch64 # aka ARMv8 | ||
| distro: ubuntu18.04 | ||
| env: | | ||
| JEMALLOC_SYS_WITH_LG_PAGE: 16 |
There was a problem hiding this comment.
Which lines? These lines were already commented in the other file
It was not commented here, for the arm binary
There was a problem hiding this comment.
This CI is the old one to generate armv8 binary.
In the new CI (to generate aarch64 binary), line 87, we can see the env variable is still present if I'm correct
There was a problem hiding this comment.
Ah yes sorry I missed it my bad!
| arch: aarch64 # aka ARMv8 | ||
| distro: ubuntu18.04 | ||
| env: | | ||
| JEMALLOC_SYS_WITH_LG_PAGE: 16 |
There was a problem hiding this comment.
Ah yes sorry I missed it my bad!
|
bors merge |
Fixes #1909
publish-binary.yml