std: merge the unix-like io::error modules into one file#158870

valentynkit · 2026-07-06T17:20:34Z

library/std/src/sys/io/error/ had a separate decode_error_kind (errno -> io::ErrorKind) for each POSIX-like target: unix.rs, wasi.rs, and teeos.rs. The three were near-identical hand copies and had drifted apart, each in its own way (#158490).

Per the discussion there, this folds wasi.rs and teeos.rs into unix.rs. decode_error_kind and is_interrupted are now shared; errno, set_errno, and error_string stay per-target behind cfg, next to the target-specific cases the file already had. A new mapping is written once and applies to all three.

The shared table is unix's existing one, so the merge also picks up the arms wasi and teeos were missing:

wasi now maps ENOTEMPTY -> DirectoryNotEmpty (was Uncategorized).
teeos now maps EINPROGRESS -> InProgress, EMFILE/ENFILE -> TooManyOpenFiles, and EOPNOTSUPP -> Unsupported (were Uncategorized).

All of those errnos are defined on the respective targets in libc, so the arms compile everywhere the file does. wasi's EWOULDBLOCK arm becomes the x == EAGAIN || x == EWOULDBLOCK guard the other two already use; that is a no-op on wasi, where EWOULDBLOCK == EAGAIN.

No test: with one shared decode_error_kind the copies can't drift again, and the decode arms are not tested today. (Let me know if it worth covering with tests)

Addresses #158490

r? @joboet

rustbot · 2026-07-06T17:20:40Z

joboet is currently at their maximum review capacity.
They may take a while to respond.

joboet · 2026-07-06T17:26:23Z

+
+#[cfg(target_os = "teeos")]
+pub fn errno() -> i32 {
+    unsafe { (*libc::__errno_location()) as i32 }


I think it'd be good to use the extern "C" redefinition of __errno_location for TEEOS as well; that way it'll benefit from the #[ffi_const] annotation.

View changes since the review

joboet · 2026-07-06T17:27:31Z

+#[cfg(target_os = "wasi")]
+pub fn error_string(errno: i32) -> String {
+    let mut buf = [0 as libc::c_char; 1024];
+
+    let p = buf.as_mut_ptr();
+    unsafe {
+        if libc::strerror_r(errno as libc::c_int, p, buf.len()) < 0 {
+            panic!("strerror_r failure");
+        }
+        str::from_utf8(CStr::from_ptr(p).to_bytes()).unwrap().to_owned()
+    }
+}


I'd merge this with the UNIX implementation. Keeping the larger buffer size for WASI seems less risky to me, perhaps there's a good reason for it being larger? Not sure...

View changes since the review

Merged, kept the 1024 buffer for them.
I couldn't find the reason in the history either. But bigger is probably safer, will truncate longer messages.
The merge also switches wasi from unwrap to from_utf8_lossy.

rustbot · 2026-07-06T17:28:33Z

Reminder, once the PR becomes ready for a review, use @rustbot ready.

valentynkit · 2026-07-06T18:06:39Z

@rustbot ready

joboet · 2026-07-06T18:34:05Z

+#[cfg(any(target_family = "unix", target_os = "teeos"))]
 unsafe extern "C" {
    #[cfg(not(any(target_os = "dragonfly", target_os = "vxworks", target_os = "rtems")))]


I'd just add WASI to these exceptions.

Suggested change

#[cfg(any(target_family = "unix", target_os = "teeos"))]

unsafe extern "C" {

#[cfg(not(any(target_os = "dragonfly", target_os = "vxworks", target_os = "rtems")))]

unsafe extern "C" {

#[cfg(not(any(target_os = "dragonfly", target_os = "vxworks", target_os = "rtems", target_os = "wasi")))]

View changes since the review

joboet · 2026-07-06T18:34:27Z

@@ -1,6 +1,9 @@
-use crate::ffi::{CStr, c_char, c_int};
+use crate::ffi::c_int;
+#[cfg(any(target_family = "unix", target_os = "wasi"))]


Suggested change

#[cfg(any(target_family = "unix", target_os = "wasi"))]

#[cfg(not(target_os = "teeos"))]

View changes since the review

joboet · 2026-07-06T18:35:01Z

+#[cfg(all(
+    any(target_family = "unix", target_os = "teeos"),
+    not(any(target_os = "dragonfly", target_os = "vxworks", target_os = "rtems"))
+))]


Likewise...

View changes since the review

joboet · 2026-07-06T18:38:02Z

+#[cfg(any(target_family = "unix", target_os = "wasi"))]
 pub fn error_string(errno: i32) -> String {
-    const TMPBUF_SZ: usize = 128;
+    const TMPBUF_SZ: usize = 1024;


I'd keep the 128-byte buffer for UNIX – some smaller UNIXes are very mindful about stack usage. Try using cfg! perhaps.

View changes since the review

joboet · 2026-07-07T17:26:03Z

+#[cfg(all(
+    target_family = "unix",
+    not(target_os = "dragonfly"),
+    not(target_os = "vxworks"),
+    not(target_os = "rtems")
+))]


It's fine to leave this unused on TEEOS...

Suggested change

#[cfg(all(

target_family = "unix",

not(target_os = "dragonfly"),

not(target_os = "vxworks"),

not(target_os = "rtems")

))]

#[cfg(not(any(

target_os = "dragonfly",

target_os = "vxworks",

target_os = "rtems",

target_os = "wasi",

)))]

View changes since the review

Thanks for you guidance on this, applied your suggestions.
Sorry that haven't noticed it in first attempts.

No worries 😄!

joboet

This looks fine, thank you! Could you squash your commits, please?

View changes since this review

The POSIX-like `decode_error_kind` copies under `sys/io/error/` (unix, wasi, teeos) had drifted apart, each missing different arms. Fold wasi and teeos into unix: `decode_error_kind` and `is_interrupted` are shared, while `errno`, `set_errno`, `error_string`, and the `errno_location` extern stay per-target behind `cfg`, expressed as exceptions to the unix/wasi/teeos set rather than enumerating the matching targets, matching the platform `cfg_attr`s the file already had. teeos now goes through the shared `errno_location` extern instead of calling `libc::__errno_location()` directly, so it gets the same `#[ffi_const]` annotation as the other targets. `error_string` is shared between unix and wasi, with the buffer sized per target: 128 bytes on unix, where some targets are mindful of stack usage, and 1024 on wasi. The shared decode table is unix's existing one, so wasi picks up `ENOTEMPTY` -> `DirectoryNotEmpty` and teeos picks up `EINPROGRESS` -> `InProgress`, `EMFILE`/`ENFILE` -> `TooManyOpenFiles`, and `EOPNOTSUPP` -> `Unsupported`, all previously `Uncategorized` on those targets. No test: with one shared `decode_error_kind` the copies can't drift again, and no test asserts the errno to `ErrorKind` mappings today.

valentynkit · 2026-07-07T17:47:30Z

This looks fine, thank you! Could you squash your commits, please?

Sure, squashed it, the commit body is pretty detailed after squash, so if you would prefer to make it more concise, please let me know.

joboet · 2026-07-08T22:38:52Z

No, that's fine. Though I should point out that the merge commit will include the PR description anyway, so commit messages don't need to be that detailed...

@bors r+

rust-bors · 2026-07-08T22:38:55Z

📌 Commit 865fb57 has been approved by joboet

It is now in the queue for this repository.

@joboet

…file, r=joboet std: merge the unix-like io::error modules into one file `library/std/src/sys/io/error/` had a separate `decode_error_kind` (errno -> `io::ErrorKind`) for each POSIX-like target: `unix.rs`, `wasi.rs`, and `teeos.rs`. The three were near-identical hand copies and had drifted apart, each in its own way (rust-lang#158490). Per the discussion there, this folds `wasi.rs` and `teeos.rs` into `unix.rs`. `decode_error_kind` and `is_interrupted` are now shared; `errno`, `set_errno`, and `error_string` stay per-target behind `cfg`, next to the target-specific cases the file already had. A new mapping is written once and applies to all three. The shared table is unix's existing one, so the merge also picks up the arms wasi and teeos were missing: - wasi now maps `ENOTEMPTY` -> `DirectoryNotEmpty` (was `Uncategorized`). - teeos now maps `EINPROGRESS` -> `InProgress`, `EMFILE`/`ENFILE` -> `TooManyOpenFiles`, and `EOPNOTSUPP` -> `Unsupported` (were `Uncategorized`). All of those errnos are defined on the respective targets in libc, so the arms compile everywhere the file does. wasi's `EWOULDBLOCK` arm becomes the `x == EAGAIN || x == EWOULDBLOCK` guard the other two already use; that is a no-op on wasi, where `EWOULDBLOCK == EAGAIN`. No test: with one shared `decode_error_kind` the copies can't drift again, and the decode arms are not tested today. (Let me know if it worth covering with tests) Addresses rust-lang#158490 r? @joboet

Rollup of 25 pull requests Successful merges: - #158871 (add relnotes for 1.97.0) - #158968 (stdarch subtree update) - #157690 (codegen_ssa: pack small const aggregates into immediate stores) - #158541 (Move `std::io::Write` to `core::io`) - #154445 (rustdoc: Represent `--output-format=json` coverage and ir differently) - #156370 (Reject linked dylib EII default overrides) - #157153 (allow `Allocator`s to be used as `#[global_allocator]`s) - #158495 (Rename HAS_CT_PROJECTION to HAS_CONST_ALIAS) - #158617 (allow mGCA const arguments to fall back to anon consts) - #158645 (Fix splat ICEs and ban it in closures) - #158655 (Fix coroutine MIR saved local remapping) - #158666 (Carry the `b_offset` inside `BackendRepr::ScalarPair`) - #158870 (std: merge the unix-like io::error modules into one file) - #158920 (Update wasm-component-ld to 0.5.26) - #158926 (wrapping_sh* methods: clarify underspecified reference) - #158927 (add core test run with `-Zforce-intrinsic-fallback`) - #158932 (Do not build the compiler when invoking `x perf compare`) - #158937 (Emit the emscripten entry point as `__main_argc_argv`) - #151379 (Stabilize `VecDeque::retain_back` from `truncate_front`) - #156144 (Better docs for PartialEq (includes macro rename)) - #156548 ( Library support for aarch64-unknown-linux-pauthtest target) - #158307 (CI job for parallel frontend ui tests) - #158347 (Improve generic parameters handling for #[diagnostic::on_const]) - #158722 (delegation: do not always inherit `ConstArgHasType` predicates) - #158741 (Simplify `Option::into_flat_iter` signature)

@joboet

…file, r=joboet std: merge the unix-like io::error modules into one file `library/std/src/sys/io/error/` had a separate `decode_error_kind` (errno -> `io::ErrorKind`) for each POSIX-like target: `unix.rs`, `wasi.rs`, and `teeos.rs`. The three were near-identical hand copies and had drifted apart, each in its own way (rust-lang#158490). Per the discussion there, this folds `wasi.rs` and `teeos.rs` into `unix.rs`. `decode_error_kind` and `is_interrupted` are now shared; `errno`, `set_errno`, and `error_string` stay per-target behind `cfg`, next to the target-specific cases the file already had. A new mapping is written once and applies to all three. The shared table is unix's existing one, so the merge also picks up the arms wasi and teeos were missing: - wasi now maps `ENOTEMPTY` -> `DirectoryNotEmpty` (was `Uncategorized`). - teeos now maps `EINPROGRESS` -> `InProgress`, `EMFILE`/`ENFILE` -> `TooManyOpenFiles`, and `EOPNOTSUPP` -> `Unsupported` (were `Uncategorized`). All of those errnos are defined on the respective targets in libc, so the arms compile everywhere the file does. wasi's `EWOULDBLOCK` arm becomes the `x == EAGAIN || x == EWOULDBLOCK` guard the other two already use; that is a no-op on wasi, where `EWOULDBLOCK == EAGAIN`. No test: with one shared `decode_error_kind` the copies can't drift again, and the decode arms are not tested today. (Let me know if it worth covering with tests) Addresses rust-lang#158490 r? @joboet

Rollup of 23 pull requests Successful merges: - #158968 (stdarch subtree update) - #154445 (rustdoc: Represent `--output-format=json` coverage and ir differently) - #158495 (Rename HAS_CT_PROJECTION to HAS_CONST_ALIAS) - #158666 (Carry the `b_offset` inside `BackendRepr::ScalarPair`) - #158870 (std: merge the unix-like io::error modules into one file) - #158920 (Update wasm-component-ld to 0.5.26) - #158926 (wrapping_sh* methods: clarify underspecified reference) - #158927 (add core test run with `-Zforce-intrinsic-fallback`) - #158932 (Do not build the compiler when invoking `x perf compare`) - #158937 (Emit the emscripten entry point as `__main_argc_argv`) - #151379 (Stabilize `VecDeque::retain_back` from `truncate_front`) - #156144 (Better docs for PartialEq (includes macro rename)) - #156548 ( Library support for aarch64-unknown-linux-pauthtest target) - #157995 (`Vec::dedup_by` docs explicit function argument order) - #158307 (CI job for parallel frontend ui tests) - #158741 (Simplify `Option::into_flat_iter` signature) - #158807 (Add regression test for CString::clone_into unwind safety) - #158862 (Fix the span for parameter suggestion ) - #158894 (Make the ordering of non-terminal binds in ambiguity error messages deterministic) - #158902 (add codegen test for range length bound propagation) - #158913 (Update `browser-ui-test` version to `0.24.1`) - #158935 (std: support real fd methods on Emscripten) - #158978 (Add regression test for too-big by-value ABI args)

@joboet

Rollup merge of #158870 - valentynkit:posix-io-error-single-file, r=joboet std: merge the unix-like io::error modules into one file `library/std/src/sys/io/error/` had a separate `decode_error_kind` (errno -> `io::ErrorKind`) for each POSIX-like target: `unix.rs`, `wasi.rs`, and `teeos.rs`. The three were near-identical hand copies and had drifted apart, each in its own way (#158490). Per the discussion there, this folds `wasi.rs` and `teeos.rs` into `unix.rs`. `decode_error_kind` and `is_interrupted` are now shared; `errno`, `set_errno`, and `error_string` stay per-target behind `cfg`, next to the target-specific cases the file already had. A new mapping is written once and applies to all three. The shared table is unix's existing one, so the merge also picks up the arms wasi and teeos were missing: - wasi now maps `ENOTEMPTY` -> `DirectoryNotEmpty` (was `Uncategorized`). - teeos now maps `EINPROGRESS` -> `InProgress`, `EMFILE`/`ENFILE` -> `TooManyOpenFiles`, and `EOPNOTSUPP` -> `Unsupported` (were `Uncategorized`). All of those errnos are defined on the respective targets in libc, so the arms compile everywhere the file does. wasi's `EWOULDBLOCK` arm becomes the `x == EAGAIN || x == EWOULDBLOCK` guard the other two already use; that is a no-op on wasi, where `EWOULDBLOCK == EAGAIN`. No test: with one shared `decode_error_kind` the copies can't drift again, and the decode arms are not tested today. (Let me know if it worth covering with tests) Addresses #158490 r? @joboet

Rollup of 23 pull requests Successful merges: - rust-lang/rust#158968 (stdarch subtree update) - rust-lang/rust#154445 (rustdoc: Represent `--output-format=json` coverage and ir differently) - rust-lang/rust#158495 (Rename HAS_CT_PROJECTION to HAS_CONST_ALIAS) - rust-lang/rust#158666 (Carry the `b_offset` inside `BackendRepr::ScalarPair`) - rust-lang/rust#158870 (std: merge the unix-like io::error modules into one file) - rust-lang/rust#158920 (Update wasm-component-ld to 0.5.26) - rust-lang/rust#158926 (wrapping_sh* methods: clarify underspecified reference) - rust-lang/rust#158927 (add core test run with `-Zforce-intrinsic-fallback`) - rust-lang/rust#158932 (Do not build the compiler when invoking `x perf compare`) - rust-lang/rust#158937 (Emit the emscripten entry point as `__main_argc_argv`) - rust-lang/rust#151379 (Stabilize `VecDeque::retain_back` from `truncate_front`) - rust-lang/rust#156144 (Better docs for PartialEq (includes macro rename)) - rust-lang/rust#156548 ( Library support for aarch64-unknown-linux-pauthtest target) - rust-lang/rust#157995 (`Vec::dedup_by` docs explicit function argument order) - rust-lang/rust#158307 (CI job for parallel frontend ui tests) - rust-lang/rust#158741 (Simplify `Option::into_flat_iter` signature) - rust-lang/rust#158807 (Add regression test for CString::clone_into unwind safety) - rust-lang/rust#158862 (Fix the span for parameter suggestion ) - rust-lang/rust#158894 (Make the ordering of non-terminal binds in ambiguity error messages deterministic) - rust-lang/rust#158902 (add codegen test for range length bound propagation) - rust-lang/rust#158913 (Update `browser-ui-test` version to `0.24.1`) - rust-lang/rust#158935 (std: support real fd methods on Emscripten) - rust-lang/rust#158978 (Add regression test for too-big by-value ABI args)

rustbot assigned joboet Jul 6, 2026

rustbot added S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. T-libs Relevant to the library team, which will review and decide on the PR/issue. labels Jul 6, 2026

joboet requested changes Jul 6, 2026

View reviewed changes

valentynkit marked this pull request as ready for review July 6, 2026 18:03

rustbot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. and removed S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. labels Jul 6, 2026

valentynkit requested a review from joboet July 6, 2026 18:14

joboet reviewed Jul 6, 2026

View reviewed changes

joboet reviewed Jul 7, 2026

View reviewed changes

joboet approved these changes Jul 7, 2026

View reviewed changes

valentynkit force-pushed the posix-io-error-single-file branch from cf885f7 to 7b789b7 Compare July 7, 2026 17:43

This comment has been minimized.

Sign in to view

valentynkit force-pushed the posix-io-error-single-file branch from 7b789b7 to 865fb57 Compare July 7, 2026 17:45

rust-bors Bot added S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels Jul 8, 2026

jhpratt mentioned this pull request Jul 8, 2026

Rollup of 25 pull requests #158986

Closed

This was referenced Jul 9, 2026

Rollup of 25 pull requests #158989

Closed

Rollup of 23 pull requests #158991

Merged

rust-bors Bot merged commit 376f430 into rust-lang:main Jul 9, 2026
13 checks passed

rustbot added this to the 1.99.0 milestone Jul 9, 2026

valentynkit mentioned this pull request Jul 13, 2026

std: per-platform decode_error_kind copies have drifted (wasi/teeos missing arms) #158490

Closed

	#[cfg(any(target_family = "unix", target_os = "wasi"))]
	#[cfg(not(target_os = "teeos"))]

Uh oh!

Uh oh!

Conversation

valentynkit commented Jul 6, 2026 • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

rustbot commented Jul 6, 2026

Uh oh!

joboet Jul 6, 2026 • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

valentynkit Jul 6, 2026

Choose a reason for hiding this comment

Uh oh!

joboet Jul 6, 2026 • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

valentynkit Jul 6, 2026

Choose a reason for hiding this comment

Uh oh!

rustbot commented Jul 6, 2026

Uh oh!

valentynkit commented Jul 6, 2026

Uh oh!

joboet Jul 6, 2026 • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

joboet Jul 6, 2026 • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

joboet Jul 6, 2026 • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

joboet Jul 6, 2026 • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

joboet Jul 7, 2026 • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

valentynkit Jul 7, 2026

Choose a reason for hiding this comment

Uh oh!

joboet Jul 7, 2026

Choose a reason for hiding this comment

Uh oh!

joboet left a comment • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

This comment has been minimized.

valentynkit commented Jul 7, 2026

Uh oh!

joboet commented Jul 8, 2026

Uh oh!

rust-bors Bot commented Jul 8, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

valentynkit commented Jul 6, 2026 •

edited by rustbot

Loading

joboet Jul 6, 2026 •

edited by rustbot

Loading

joboet Jul 6, 2026 •

edited by rustbot

Loading

joboet Jul 6, 2026 •

edited by rustbot

Loading

joboet Jul 6, 2026 •

edited by rustbot

Loading

joboet Jul 6, 2026 •

edited by rustbot

Loading

joboet Jul 6, 2026 •

edited by rustbot

Loading

joboet Jul 7, 2026 •

edited by rustbot

Loading

joboet left a comment •

edited by rustbot

Loading