support Errno formatting in no_std #43
Conversation
|
Hi @A1-Triard, sorry for the delay! Any reason why this was closed? This looks pretty good to me. |
|
Hi! I had finally decided it will be better for me to stick with my own fork (errno-no-std, name is misleading now, but I don't want to pollute crates.io with multiply renamed copies). (This is not because the response delay, I have my own reasons.) That's the only reason, why it was closed. In any case, I've reopened the request now. Please feel free to merge it if you want. |
There was a problem hiding this comment.
LGTM modulo comments
Also, there's this error in CI:
error[E0433]: failed to resolve: use of undeclared type `CStr`
Error: --> src/unix.rs:44:26
|
44 | let c_str = unsafe { CStr::from_ptr(buf.as_ptr()) };
| ^^^^ use of undeclared type `CStr`
error[E0412]: cannot find type `c_char` in this scope
Error: --> src/unix.rs:35:25
|
35 | let mut buf = [0 as c_char; 1024];
| ^^^^^^
|
help: a builtin type with a similar name exists
|
35 | let mut buf = [0 as char; 1024];
| ~~~~
help: consider importing this type alias
|
17 | use libc::c_char;
|
error[E0425]: cannot find function `strerror_r` in this scope
Error: --> src/unix.rs:37:12
|
37 | if strerror_r(err.0, buf.as_mut_ptr(), buf.len() as libc::size_t) < 0 {
| ^^^^^^^^^^ not found in this scope
|
help: consider importing this function
|
17 | use libc::strerror_r;
|
| for (c, i) in UStr::from_slice(input).char_indices_lossy().take_while(|&(x, _)| x != '\0') { | ||
| let c_len = c.len_utf8(); | ||
| if c_len > output.len() { | ||
| return (i, unsafe { output.as_ptr().offset_from(output_start) } as usize); |
There was a problem hiding this comment.
This unsafe looks unnecessary. Can we rewrite it using safe code? Say, by keeping track of how many bytes are written so far.
| use winapi::um::winbase::{FORMAT_MESSAGE_FROM_SYSTEM, FORMAT_MESSAGE_IGNORE_INSERTS}; | ||
|
|
||
| use Errno; | ||
|
|
||
| #[cfg(feature = "std")] | ||
| fn from_utf16_lossy(input: &[u16], mut output: &mut [u8]) -> (usize, usize) { |
There was a problem hiding this comment.
This function guarantees that the written bytes are valid UTF-8. To localize this invariant, can we have the function return a &str directly?
fn from_utf16_lossy<'a>(input: &[u16], output: &'a mut [u8]) -> &'a str
|
Because I deleted the fork, I have to create a new pull request (#44) instead of this one. Please re-review it. |
No description provided.