Add flock64 to linux_like platforms

[elichai]

Hi,
I'm not sure about what to make of the ifdef in glibc __USE_LARGEFILE64 that covers it.
but I copied the impl from glibc's headers.
everything is the same except for sparc which has a reserved short int.
and in musl it's just #define flock64 flock

[rust-highfive]

Thanks for the pull request, and welcome! The Rust team is excited to review your changes, and you should hear from @gnzlbg (or someone else) soon.

If any changes to this PR are deemed necessary, please add them as extra commits. This ensures that the reviewer can see what has changed since they last reviewed the code. Due to the way GitHub handles out-of-date commits, this should also make it reasonably obvious what issues have or haven't been addressed. Large or tricky changes may require several passes of review and changes.

Please see the contribution instructions for more information.

[gnzlbg]

@bors: r+

[bors]

📌 Commit 0da3a5f has been approved by gnzlbg

[glaubitz]

If this gets merged before my sparc PR, I should probably add flock64 to sparc as well.

[bors]

⌛ Testing commit 0da3a5f with merge 1d883da...

[bors]

💔 Test failed - status-azure

[gnzlbg]

From the error messages it appears that the type is defined differently on different targets. You need to fix those.

[elichai]

From the error messages it appears that the type is defined differently on different targets. You need to fix those.

First of all awesome test suite :) I need to read how it works.

Second, seems that the problem is the musl targets, which makes sense. i'll fix it.

[gnzlbg]

Notice that there are generic files for linux/gnu/mod.rs and linux/musl/mod.rs, so you might get away with just having two structs (I'm not sure).

[elichai]

Notice that there are generic files for linux/gnu/mod.rs and linux/musl/mod.rs, so you might get away with just having two structs (I'm not sure).

the glibc once are fine (they match glibc correctly and pass the tests) on the other hand, musl have
#define flock64 flock and even that it's under #if defined(_LARGEFILE64_SOURCE) || defined(_GNU_SOURCE) so maybe there's just no flock64 in the test framework, although the errors are weird. I'm trying to reproduce them locally

https://git.musl-libc.org/cgit/musl/tree/include/fcntl.h#n202

[gnzlbg]

The test framework always defines _GNU_SOURCE for all linux targets: https://github.com/rust-lang/libc/blob/master/libc-test/build.rs#L2074

[glaubitz]

Be careful though as flock64 is different on sparc* targets so you need arch-specific definitions.

[elichai]

Be careful though as flock64 is different on sparc* targets so you need arch-specific definitions.

I know, did that already, the weird thing is actually musl targets which should be flock==flock64

[elichai]

@gnzlbg Ok, I got it down to this example: https://ideone.com/rgk7OZ
the problem seems to be line 22:

  typedef struct {
    unsigned char c;
    flock v;
  } type;

it should be struct flock v;

The question is why this happens

[elichai]

Update, that's definitely because I defined pub type flock64 = ::flock;. it seems like the translation unit to C doesn't understand it properly.
i'll try to reproduce on ctest and open an issue there.

Should I replace the typedef with a duplicate struct?

[gnzlbg]

There are methods in ctest to configure how to name the type aliases, you can use cfg.type_name to tell ctest to expand flock64 on that particular target to struct flock64 - there should be many examples of this in the libc-test/build.rs.

[gnzlbg]

Thanks! @bors: r+

[bors]

📌 Commit 490e073 has been approved by gnzlbg

[bors]

⌛ Testing commit 490e073 with merge 7f33bcb...

[bors]

☀️ Test successful - checks-cirrus-freebsd-10, checks-cirrus-freebsd-11, checks-cirrus-freebsd-12, status-azure
Approved by: gnzlbg
Pushing 7f33bcb to master...

[glaubitz]

Oh, is the CI working again? I just re-pushed my PR.