Added the waitid interface #418
Conversation
…n individual threads, thread groups, or every thread on the system.
|
(rust_highfive has picked a reviewer for you, use r? to override) |
|
Invalid pointer type attempting fix. |
| P_ALL, | ||
| P_PID, | ||
| P_PGID | ||
| } |
There was a problem hiding this comment.
Right now we actually tend to avoid literal enum types as they're not extensible (while the C side is), so perhaps these could just be pub const P_ALL: c_int = i32?
There was a problem hiding this comment.
That should work. I wasn't sure which was more idiomatic. Should I go thought the whole song of dance of pub type id_t = c_int; pub const P_ALL: ::id_t = i32; ?
Oh nevermind id_t is defined.
There was a problem hiding this comment.
Yeah let's take the route of listing all the constants explicitly.
There was a problem hiding this comment.
I lifted the the constant definition to linux/mod.rs from /linux/other/mod.rs because sys/wait.h should be standard cross all GNU/Linux platforms.
The type+explicitly declared constants is added. Waiting for Travis-ci to report.
…across all *GNU* platforms. So I think putting at this high of level is okay, but if that is true I should lift waitid up to apply to all the various linux flavors
| pub enum fpos64_t {} // TODO: fill this out with a struct | ||
|
|
||
| pub const P_PID: ::id_t = 1; | ||
| pub const P_PGID: ::id_t = 2; | ||
| pub const P_ALL: ::id_t = 3; |
There was a problem hiding this comment.
Update.
Also while I have your attention. How does one go about adding the linux/*.h sources for linking?
I can see where to add them for the test suite/crate. This was the main issue why I couldn't merge the 2 remaining for PTRACE yesterday/last night.
There was a problem hiding this comment.
Unfortunately those headers are a little difficult include, but if the uapi business works then I think it'd be relatively easy to add!
|
Looks like CI may still be failing? |
|
Style error I'll submit a commit in the morning got distracted today On Saturday, October 8, 2016, Alex Crichton [email protected]
|
|
I'm not sure how to explain this failing. On a lot of platforms the are running tests for |
|
It looks like there's failures across various platforms, for examples mips-musl has |
Expand on release procedure
Waitid allows for a non-blocking way to check the status of individual threads, groups of threads, or every thread on the system.
Not 100% sure about how the linking is handled. Also can't confirm that it exists for every platform currently. I'm more or less gonna use travis for that.