Add and fix tests for {i686, aarch64}-linux-android targets by malbarbo · Pull Request #538 · rust-lang/libc

malbarbo · 2017-02-24T20:18:36Z

I think that these changes do not breaks compatibility.

There are some types and constants changes to i686 and aarch64, but I see these changes as bug fixes instead of breaking changes. Also the type time64_t was remove from aarch64 because it is not defined in this arch.

Fixes#536

- some tests are failing - remove readlink, timegm and sig* functions in favor of the unix/mod.rs definitions - remove time64_t (it is not defined for aarch64) - move some definitions to android/b32.rs and create appropriated definitions in android/b64.rs

rust-highfive · 2017-02-24T20:18:47Z

r? @alexcrichton

(rust_highfive has picked a reviewer for you, use r? to override)

alexcrichton · 2017-02-25T21:48:55Z

+// Android has some weirdness in this definition
+// See weirdness in src/unix/notbsd/android/mod.rs
+#[cfg(any(not(target_os = "android"), // " if " -- appease style checker
+ not(target_arch = "aarch64")))]


I unfortunately regret the block below this, but could we just duplicate these functions in lower modules to avoid this #[cfg] attribute?

alexcrichton · 2017-02-25T21:49:13Z

+ pub type c_char = u8;
+ pub type wchar_t = u32;
+ }
+}


Could this be pushed down further?

alexcrichton · 2017-02-25T21:49:23Z

+ pub const O_DIRECTORY: ::c_int = 0x4000;
+ pub const O_NOFOLLOW: ::c_int = 0x8000;
+ }
+}


Could this also be pushed down further?

alexcrichton · 2017-02-25T21:50:33Z

Looks great to me, thanks so much for taking the charge on this! Overall it'd be great to reduce the #[cfg] introduced here, but other than that looks great to me. (duplication is fine)

- Copy 17 functions definitions from src/unix/mod.rs to src/unix/bsd/mod.rs src/unix/haiku/mod.rs src/unix/notbsd/linux/mod.rs and src/unix/solaris/mod.rs - Add some functions to android that was cfged out - Remove cf* and tc* functions implementations for android (they are available with api >= 12, which was release in 2011)

bors · 2017-02-27T22:02:09Z

☔ The latest upstream changes (presumably #541) made this pull request unmergeable. Please resolve the merge conflicts.

malbarbo · 2017-02-27T23:36:29Z

@alexcrichton All cfg_if! removed.

We got some network issues in the tests.

kamalmarhubi · 2017-02-28T01:40:40Z

@malbarbo maybe try git commit --amend --no-edit && git push -f to force travis to run again.

malbarbo · 2017-02-28T11:18:57Z

@alexcrichton All green now.

@kamalmarhubi Thanks for the tip. It worked.

alexcrichton · 2017-02-28T15:32:39Z

 pub const SIGEV_THREAD_ID: ::c_int = 4;

 f! {
- pub fn sigemptyset(set: *mut sigset_t) -> ::c_int {


It looks like these functions may not have survived a rebase by accident?

These are removed purposely. The old tests used android-9 api, which do not have this functions. The next api version distributed with android ndk is android-12 (release in 2011 https://en.wikipedia.org/wiki/Android_version_history), which define these functions. The definitions used by android are the same of others unix (src/unix/mod.rs).
I can see that this can be a breaking change... Should I return these implementations? (this will require moving definitions out of src/unix/mod.rs)
Keeping our own implementations of these functions seems odd.

Ah ok, in that case sounds great to me!
Android-9 is such an ancient NDK that it seems fine to not require support for that here. If project want that level of compatibility (e.g. libstd) we can vendor the impls ourselves.

alexcrichton · 2017-02-28T20:50:08Z

@bors: r+

bors · 2017-02-28T20:50:09Z

📌 Commit b2791db has been approved by alexcrichton

bors · 2017-02-28T20:50:18Z

⌛ Testing commit b2791db with merge 1fa8e30...

Add and fix tests for {i686, aarch64}-linux-android targets I think that these changes do not breaks compatibility. There are some types and constants changes to i686 and aarch64, but I see these changes as bug fixes instead of breaking changes. Also the type time64_t was remove from aarch64 because it is not defined in this arch. Fixes#536

bors · 2017-03-01T01:49:07Z

💔 Test failed - status-travis

kamalmarhubi · 2017-03-01T14:20:56Z

Failures look related to yesterday's S3 outage:

docker pull failures
falied to download qemu image from S3

kamalmarhubi · 2017-03-01T14:22:44Z

@malbarbo if you get to it before @alexcrichton retries, you could rebase and force push. Push will get Travis to run, rebase will allow homu to use the cached Travis success so long as no other PR is merged in the meantime. Hopefully we can get this out today!

alexcrichton · 2017-03-01T16:24:27Z

I can't even see the failures, but I'll take your word for it!

@bors: retry

bors · 2017-03-01T16:24:43Z

⌛ Testing commit b2791db with merge 612a586...

Add and fix tests for {i686, aarch64}-linux-android targets I think that these changes do not breaks compatibility. There are some types and constants changes to i686 and aarch64, but I see these changes as bug fixes instead of breaking changes. Also the type time64_t was remove from aarch64 because it is not defined in this arch. Fixes#536

bors · 2017-03-01T20:16:18Z

☀️ Test successful - status-appveyor, status-travis
Approved by: alexcrichton
Pushing 612a586 to master...

Even if it is unlikely to happen the VSX instructions can be executed in 32bit mode just as well.

malbarbo added 4 commits February 23, 2017 16:08

Add test support for {aarch64, i686}-linux-android
599eec3

Fix i686-linux-android failing tests
c532518

Fix aarch64-linux-android failing tests
f2f1129

rust-highfive assigned alexcrichton Feb 24, 2017

Enable travis tests for {i686, aarch64}-linux-android
d500a45

malbarbo mentioned this pull request Feb 25, 2017
Make a release :-) #539
Closed

alexcrichton reviewed Feb 25, 2017
View reviewed changes

malbarbo added 3 commits February 27, 2017 16:41

Remove cfg_if from src/unix/notbsd/android/
1175c66

Merge branch 'master' into ci-android
bfeb360

malbarbo force-pushed the ci-android branch from 3e0afd2 to d9ccad7Compare February 27, 2017 20:26

Fix missing prefix ::
a3759d7

malbarbo force-pushed the ci-android branch from d9ccad7 to a3759d7Compare February 27, 2017 20:41

Fix style
9a20f93

kamalmarhubi mentioned this pull request Feb 28, 2017
Any plans for 0.8.0? nix-rust/nix#519
Closed

Merge branch 'master' into ci-android
b2791db

malbarbo force-pushed the ci-android branch from b0efa03 to b2791dbCompare February 28, 2017 02:11

alexcrichton reviewed Feb 28, 2017
View reviewed changes

bors merged commit b2791db into rust-lang:masterMar 1, 2017

bors mentioned this pull request Mar 1, 2017
Fix android SA_* constants #511
Merged

malbarbo deleted the ci-android branch March 2, 2017 10:01

malbarbo restored the ci-android branch March 2, 2017 11:07

malbarbo mentioned this pull request May 4, 2017
Fix stat definition for android aarch64 rust-lang/rust#41752
Closed

danielverkamp pushed a commit to danielverkamp/libc that referenced this pull request Apr 28, 2020
Move the vsx code to powerpc (rust-lang#538)
3e2efe9
Even if it is unlikely to happen the VSX instructions can be executed in 32bit mode just as well.

Uh oh!

Conversation

malbarbo commented Feb 24, 2017• edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

rust-highfive commented Feb 24, 2017

Uh oh!

alexcrichtonFeb 25, 2017

Choose a reason for hiding this comment

Uh oh!

alexcrichtonFeb 25, 2017

Choose a reason for hiding this comment

Uh oh!

alexcrichtonFeb 25, 2017

Choose a reason for hiding this comment

Uh oh!

alexcrichton commented Feb 25, 2017

Uh oh!

bors commented Feb 27, 2017

Uh oh!

malbarbo commented Feb 27, 2017

Uh oh!

kamalmarhubi commented Feb 28, 2017

Uh oh!

malbarbo commented Feb 28, 2017

Uh oh!

alexcrichtonFeb 28, 2017

Choose a reason for hiding this comment

Uh oh!

malbarboFeb 28, 2017

Choose a reason for hiding this comment

Uh oh!

alexcrichtonFeb 28, 2017

Choose a reason for hiding this comment

Uh oh!

alexcrichton commented Feb 28, 2017

Uh oh!

bors commented Feb 28, 2017

Uh oh!

bors commented Feb 28, 2017

Uh oh!

bors commented Mar 1, 2017

Uh oh!

kamalmarhubi commented Mar 1, 2017

Uh oh!

kamalmarhubi commented Mar 1, 2017

Uh oh!

alexcrichton commented Mar 1, 2017

Uh oh!

bors commented Mar 1, 2017

Uh oh!

bors commented Mar 1, 2017

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

malbarbo commented Feb 24, 2017•
edited
Loading