emscripten: deprecate file offset types

[dybucc]

Description

This PR deprecates all LFS64 types and routines in emscripten. This target follows quite closely the musl upstream. The changes they make are not concerned with file offset types. The only exception to this is the blkcnt_t type. This type is always 32-bits wide no matter the target machine.

All other types have received the same treatment as in #5170.

Sources

• Upstream emscripten sources showing both the final alltypes.h header with the off_t definition and the input to the sed script that generates it.

  https://github.com/emscripten-core/emscripten/blob/1637a13961ab1d6f9595db4db26bfc16a68e6ad8/system/lib/libc/musl/arch/emscripten/bits/alltypes.h#L221
  https://github.com/emscripten-core/musl/blob/85e0e3519655220688e757b9d5bfd314923548bd/include/alltypes.h.in#L29

• Upstream emscripten fork of musl showing what has been assumed to be the "main" definition of off64_t and related types when both the _GNU_SOURCE and _LARGEFILE64_SOURCE feature test
  macros are present.

  https://github.com/emscripten-core/musl/blob/85e0e3519655220688e757b9d5bfd314923548bd/include/sys/types.h#L79

Checklist

• Relevant tests in libc-test/semver have been updated
• No placeholder or unstable values like *LAST or *MAX are included (see #3131)
• Tested locally (cd libc-test && cargo test --target mytarget); especially relevant for platforms that may not be checked in CI

[dybucc]

Style check fails for reasons outside the changes introduced in this PR. The formatting checks seem
to run on a post-macro expanded crate output. When annotating the item produced by the extern_ty
macro with an attribute, this expands with a wrongly indented function block, even though the actual
source code is correctly indented. The output log of the CI job shows this.

[tgross35]

@rustbot blocked
@SnoozeThis wait 2 months -> remove label S-blocked, add label S-waiting-on-review

Holding off until there is a good alternative to off64_t

[SnoozeThis]

(https://snoozeth.is/FoBvqh9wkR8) I will wait until Wed, 19 Aug 2026 08:32:43 UTC and then add label S-waiting-on-review and remove label S-blocked.

@rustbot claim.

[tgross35]

Actually I think I'm missing something - what does this PR fix? I only see it adding deprecation warnings and deleting ino64_t (was that intentional?)

[dybucc]

The ino_t removal was not intentional. PRs are getting frozen because there's yet to be a solution for 32-bit platforms. This should be no issue in emscripten because the unsuffixed types were already fixed-precision 64-bit wide types. Anybody who was using the LFS64 types can switch to the unsuffixed types with no regressions with a nice warning advising them of it.

The PR description got updated with also mentioning that the blkcnt_t type is an exception. Its bit width does not change so that's the same as with the other file offset types. But it is always defined as a 32-bit and not as a 64-bit signed integer.

And then this transitively applies to the affected records that use these types.

[rustbot]

This PR was rebased onto a different main commit. Here's a range-diff highlighting what actually changed.

Rebasing is a normal part of keeping PRs up to date, so no action is needed—this note is just to help reviewers.