Move back to the main nightly channel #723
Conversation
Now that the upstream wasm32-wasi target has been fixed, we should be good to go!
We don't know what version of the wasi api it's compiled against, so let's avoid it.
|
I believe the one failure is #721, but otherwise should be good to go! |
| let stat = match wasi_unstable::fd_prestat_get(i) { | ||
| Ok(s) => s, | ||
| Err(_) => break, | ||
| }; | ||
| if stat.pr_type != wasi::PREOPENTYPE_DIR { | ||
| continue; | ||
| } | ||
| let mut dst = Vec::with_capacity(stat.u.dir.pr_name_len); | ||
| if wasi::fd_prestat_dir_name(i, dst.as_mut_ptr(), dst.capacity()).is_err() { |
There was a problem hiding this comment.
Hmm, is it OK to mix wasi_unstable and wasi for the syscalls here? I mean, first, we call wasi_unstable::fd_prestat_get just to later on call wasi::fd_prestat_dir_name. It might be OK, but definitely looks confusing!
There was a problem hiding this comment.
Oh oops this was a mistake! I'll switch this as well to match the above
| } | ||
| dst.set_len(stat.u.dir.pr_name_len); | ||
| if dst == path.as_bytes() { | ||
| return Ok(wasi::path_open(i, 0, ".", wasi::OFLAGS_DIRECTORY, 0, 0, 0) |
There was a problem hiding this comment.
Oh, and here as well, we call wasi::path_open instead of wasi_unstable::path_open.
|
Updated! |
|
Looks like we're running into a slew of bugs related to duplicate symbol names in LLVM IR, which I've posted a fix for at rust-lang/rust#67363. I'll try to think of a way to work around that in the meantime. |
Well, that's what the tests and CIs are for, right? ;-) |
|
Ok I'll close this for now while the rustc PR bakes. |
Now that the upstream wasm32-wasi target has been fixed, we should be
good to go!