Win: Add GetHostNameW fallback for win7 using gethostname #150909

Fulgen301 · 2026-01-10T03:00:07Z

GetHostNameW is only available starting with Windows 8, but gethostname has been available since Windows Vista. Use it as a fallback and convert the result from ANSI to Unicode.

Fixes #150896

This also replaces #150905 - I can rebase it onto the other PR should that one get merged first.

rustbot · 2026-01-10T03:00:12Z

r? @ChrisDenton

rustbot has assigned @ChrisDenton.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

hkBst · 2026-01-11T13:08:55Z

library/std/src/sys/pal/windows/c.rs

+compat_fn_with_fallback! {
+    pub static WS2_32: &CStr = c"ws2_32";
+
+    #[cfg(target_vendor = "win7")]


this seems redundant with the outer cfg attr...

hkBst · 2026-01-11T13:14:29Z

library/std/src/sys/pal/windows/winsock.rs

+    // SAFETY: these parameters specify a valid, writable region of memory.
+    unsafe {
+        if c::gethostname(buffer.as_mut_ptr().cast(), buffer.len() as i32) == c::SOCKET_ERROR {
+            return c::SOCKET_ERROR;
+        }


The safety comment seems to suffice for these lines, but the unsafe block continues. I think it would be better if each unsafe call had its own unsafe block with its own safety comments.

joboet · 2026-01-12T14:15:23Z

library/std/src/sys/pal/windows/winsock.rs

+        if len <= 0 {
+            // GetHostNameW reports WSAEFAULT if the buffer is too small
+            if c::GetLastError() == c::ERROR_INSUFFICIENT_BUFFER {
+                c::SetLastError(c::WSAEFAULT as _);


This should call WSASetLastError.

Technically unnecessary, as WSASetLastError has been the same as SetLastError for the last decades.

joboet · 2026-01-12T14:20:03Z

library/std/src/sys/pal/windows/winsock.rs

+
+#[cfg(target_vendor = "win7")]
+pub unsafe fn hostname_fallback(name: c::PWSTR, namelen: i32) -> i32 {
+    assert!(namelen >= 1);


I think this should return SOCKET_ERROR and set the error to WSAEFAULT, just to be safe.

joboet · 2026-01-12T14:20:46Z

library/std/src/sys/pal/windows/winsock.rs

+            namelen - 1,
+        );
+
+        if len <= 0 {


MultiByteToWideChar always returns zero on error

Suggested change

if len <= 0 {

if len == 0 {

joboet · 2026-01-12T14:21:27Z

library/std/src/sys/pal/windows/winsock.rs

+
+        if len <= 0 {
+            // GetHostNameW reports WSAEFAULT if the buffer is too small
+            if c::GetLastError() == c::ERROR_INSUFFICIENT_BUFFER {


If the error is not ERROR_INSUFFICIENT_BUFFER, this will not set the socket error variable to indicate a reason for SOCKET_ERROR. I'm not sure what error value is correct, through...

rustbot · 2026-01-12T14:23:22Z

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

ChrisDenton · 2026-01-12T21:17:25Z

cc @roblabla @PaulDance it would seem a bit silly to merge #150905 then immediately merge this PR. Do you have any thoughts on this PR.

roblabla · 2026-01-13T10:19:12Z

cc @roblabla @PaulDance it would seem a bit silly to merge #150905 then immediately merge this PR. Do you have any thoughts on this PR.

Yes, this PR obsoletes #150905, and seems like the right approach to me.

roblabla · 2026-01-13T10:21:00Z

library/std/src/sys/pal/windows/winsock.rs

+
+    // SAFETY: these parameters specify a valid, writable region of memory.
+    unsafe {
+        if c::gethostname(buffer.as_mut_ptr().cast(), buffer.len() as i32) == c::SOCKET_ERROR {


It may be worth documenting that we're using this function in the "Underlying system calls" section of library/std/src/net/hostname.rs?

I don't think we document syscalls for the win7 target anywhere(?).

It's technically not a syscall either.

The rust documentation has a habit of using the term "syscall" when it means OS API (technically they're not even syscalls on Linux since we go through libc). But yes, I agree it makes sense to document it. E.g. perhaps similarly to SystemTime.

PaulDance · 2026-01-14T20:16:22Z

it would seem a bit silly to merge #150905 then immediately merge this PR

Yes, it is, but I thought doing so would have the advantage of getting a quick and easy fix known to work merged fast while a more comprehensive but complex one would be implemented with ample time concurrently. Considering the above comments, it kind of confirms that. Either way works in the end, though.

PaulDance · 2026-01-14T20:22:22Z

library/std/src/sys/pal/windows/winsock.rs

+        );
+
+        if len <= 0 {
+            // GetHostNameW reports WSAEFAULT if the buffer is too small


Suggested change

// GetHostNameW reports WSAEFAULT if the buffer is too small

// gethostname reports WSAEFAULT if the buffer is too small

if it indeed still applies here.

That was intentional - if GetHostNameW had a too small buffer, it'd report WSAEFAULT. This provides a fallback implementation, so it matches that error behavior.

PaulDance · 2026-01-14T20:24:23Z

library/std/src/sys/pal/windows/c.rs

+    pub static WS2_32: &CStr = c"ws2_32";
+
+    #[cfg(target_vendor = "win7")]
+    pub fn GetHostNameW(name: PWSTR, namelen: i32) -> i32 {


I guess it could be nice to link the original issue somewhere around here.

Win: Add GetHostNameW fallback for win7 using gethostname

3e2e73c

rustbot assigned ChrisDenton Jan 10, 2026

rustbot added O-windows Operating system: Windows S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. T-libs Relevant to the library team, which will review and decide on the PR/issue. labels Jan 10, 2026

hkBst suggested changes Jan 11, 2026

View reviewed changes

joboet requested changes Jan 12, 2026

View reviewed changes

rustbot added S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels Jan 12, 2026

roblabla reviewed Jan 13, 2026

View reviewed changes

PaulDance reviewed Jan 14, 2026

View reviewed changes

	// GetHostNameW reports WSAEFAULT if the buffer is too small
	// gethostname reports WSAEFAULT if the buffer is too small

Uh oh!

Win: Add GetHostNameW fallback for win7 using gethostname #150909

Are you sure you want to change the base?

Win: Add GetHostNameW fallback for win7 using gethostname #150909

Conversation

Fulgen301 commented Jan 10, 2026 • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

rustbot commented Jan 10, 2026

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

rustbot commented Jan 12, 2026

Uh oh!

ChrisDenton commented Jan 12, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

roblabla commented Jan 13, 2026

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

PaulDance commented Jan 14, 2026

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

8 participants

Fulgen301 commented Jan 10, 2026 •

edited by rustbot

Loading

ChrisDenton commented Jan 12, 2026 •

edited

Loading