fix(profiling): allow to build on Alpine Linux #1391
Conversation
jd
force-pushed
the
profiling/alpine-support
branch
from
b3cb8f8
to
1a3a3db
Compare
jd
changed the title
| @@ -938,6 +949,7 @@ workflows: | |||
| - requestsgevent | |||
| - sqlalchemy | |||
| - sqlite3 | |||
| - test_build_alpine | |||
There was a problem hiding this comment.
I'd consider adding all versions we support on alpine, but this is fine.
There was a problem hiding this comment.
Indeed, the compatibility issue is with their libc, not with any Python version specifically, so I don't think it's worth it.
This hides the pthread_t type usage and force the cast to unsigned long thread no matter what. This is fine with glibc and musl libc at least. Fixes DataDog#1388
jd
force-pushed
the
profiling/alpine-support
branch
from
1a3a3db
to
f7c098c
Compare
Do you mean for Alpine, or for every Linux distro? |
Yes? It should mostly just be "you need python headers and a c++ compiler", example for probably don't need to be exhaustive... ? or maybe not worth it? |
|
I don't think it has been a problem so far. But if we ever get support/issues opened on this, it might be worth it. (don't misread me, I don't want to sound like the guy who doesn't want to write documentation. 😆) |
| # ctypedef unsigned long pthread_t | ||
| # but e.g. musl libc defines pthread_t as a struct __pthread * which breaks the arithmetic Cython | ||
| # wants to do. | ||
| # We pay this with a warning at compilation time, but it works anyhow. |
This hides the pthread_t type usage and force the cast to unsigned long thread no matter what. This is fine with glibc and musl libc at least. Fixes #1388
This hides the pthread_t type usage and force the cast to unsigned long thread
no matter what.
This is fine with glibc and musl libc at least.
Fixes #1388