Add Location::file_with_nul #135054

cramertj · 2025-01-03T01:53:45Z

This is useful for C/C++ APIs which expect the const char* returned from FILE or std::source_location::file_name.

rustbot · 2025-01-03T01:53:53Z

rustbot has assigned @cuviper.
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

rustbot · 2025-01-03T01:53:55Z

Some changes occurred to the CTFE / Miri interpreter

cc @rust-lang/miri

Some changes occurred to the CTFE machinery

cc @rust-lang/wg-const-eval

Noratrieb · 2025-01-03T08:58:51Z

see rust-lang/libs-team#466

compiler/rustc_const_eval/src/interpret/place.rs

compiler/rustc_const_eval/src/util/caller_location.rs

RalfJung · 2025-01-03T09:20:42Z

library/core/src/panic/location.rs

+    /// Although `file` returns an `&str`, the characters of `file` are guaranteed to be followed
+    /// by a nul-terminator. This allows for greater interoperabilty with C and C++ code using
+    /// `__FILE__` or `std::source_location::file_name`, both of which return nul-terminated
+    /// `const char*`.


Note that under Stacked Borrows, it is Undefined Behavior to access the trailing nul byte since it is outside the range of this reference. So it seems ill-advised to write documentation that encourages people to perform such accesses.

Removed. At one point I actually had a test for this and removed it when I realized this issue, but I forgot to remove the docs. Thanks for catching!

cramertj · 2025-01-03T17:21:10Z

see rust-lang/libs-team#466

Renamed to file_with_nul as per that proposal.

This is useful for C/C++ APIs which expect the const char* returned from __FILE__ or std::source_location::file_name.

bjorn3 · 2025-01-05T10:36:30Z

How does this compare with #131828?

cramertj · 2025-01-06T21:34:58Z

@bjorn3 This is essentially the same feature, although this implementation doesn't require an increase to the size of the Location struct itself (and could actually be used to decrease it by dropping the size of the slice if we cared to do so).

Darksonn · 2025-01-07T14:58:26Z

The other PR does not change the size of the Location struct either.

oli-obk · 2025-01-07T15:28:39Z

library/core/src/panic/location.rs

+            let str_len = self.file_bytes_with_nul.len() - 1;
+            // SAFETY: `file_bytes_with_nul` without the trailing nul byte is guaranteed to be
+            // valid UTF8.
+            unsafe { crate::str::from_raw_parts(self.file_bytes_with_nul.as_ptr(), str_len) }


str is allowed to have null bytes within. So you could keep the type at &'a str and use &self.file[..self.file.len() -1] here without unsafe code

I could! I think that just moves the unsafety into a call to from_bytes_with_nul_unchecked for the CStr, though, so I think it's roughly the same either way.

That unsafety exists either way unless you just directly encode it as CStr, tho we're still holding out for that becoming a thin pointer, at which point it would be expensive to get the str out of it.

Maybe it's simpler to just store a raw pointer and length without the nul-terminator included. That way, Location hold the integer we want most of the time without triggering provenance concerns.

That way, Location hold the integer we want most of the time without triggering provenance concerns.

What integer do you mean?

I wasn't aware of any provenance concerns, was that discussed in #131828?

The provenance discussion is in a hidden subthread above: #135054 (comment)

I mean that storing a raw pointer / length instead of a slice would let us store the length without the nul-terminator included, which makes more sense to me.

compiler/rustc_const_eval/src/util/caller_location.rs

cramertj · 2025-01-07T16:19:41Z

The other PR does not change the size of the Location struct either.

Thanks for the correction! I must've misread it when I saw it. In either case, I don't have a strong preference and am happy to remove this one if it makes life easier. I didn't know about the first PR when I wrote this one.

Darksonn · 2025-01-08T09:14:21Z

I closed my original PR in favor of this PR, and opened a new PR #135240 that shows how the alternative approach of adding a compiler flag would look.

Noratrieb · 2025-01-08T17:58:09Z

Let's do a perf run to get a concrete binary size overhead (which will of course be so tiny that it's negligible, but having proof of that seems to be necessary in this discussion!).
@bors try @rust-timer queue

bors · 2025-01-08T17:59:23Z

⌛ Trying commit 73810aa with merge 35c365c...

Add Location::file_with_nul This is useful for C/C++ APIs which expect the const char* returned from __FILE__ or std::source_location::file_name.

bors · 2025-01-08T19:50:53Z

☀️ Try build successful - checks-actions
Build commit: 35c365c (35c365c6fce43dba6f1e28794102c6f406ba02d9)

rust-timer · 2025-01-08T22:35:35Z

Finished benchmarking commit (35c365c): comparison URL.

Overall result: ❌ regressions - no action needed

Benchmarking this pull request likely means that it is perf-sensitive, so we're automatically marking it as not fit for rolling up. While you can manually mark this PR as fit for rollup, we strongly recommend not doing so since this PR may lead to changes in compiler perf.

@bors rollup=never
@rustbot label: -S-waiting-on-perf -perf-regression

Instruction count

This is the most reliable metric that we have; it was used to determine the overall result at the top of this comment. However, even this metric can sometimes exhibit noise.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	0.2%	[0.2%, 0.3%]	3
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-	-	0

Max RSS (memory usage)

Results (primary 9.1%)

This is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	9.1%	[9.1%, 9.1%]	1
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	9.1%	[9.1%, 9.1%]	1

Cycles

Results (primary -1.7%, secondary -2.6%)

This is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-1.7%	[-3.3%, -0.8%]	14
Improvements ✅ (secondary)	-2.6%	[-3.8%, -1.3%]	7
All ❌✅ (primary)	-1.7%	[-3.3%, -0.8%]	14

Binary size

Results (primary 0.0%, secondary 0.0%)

This is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	0.0%	[0.0%, 0.0%]	22
Regressions ❌ (secondary)	0.0%	[0.0%, 0.1%]	22
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	0.0%	[0.0%, 0.0%]	22

Bootstrap: 763.722s -> 763.688s (-0.00%)
Artifact size: 325.68 MiB -> 325.70 MiB (0.01%)

rustbot assigned cuviper Jan 3, 2025

rustbot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. T-compiler Relevant to the compiler team, which will review and decide on the PR/issue. T-libs Relevant to the library team, which will review and decide on the PR/issue. labels Jan 3, 2025

RalfJung reviewed Jan 3, 2025

View reviewed changes

cramertj force-pushed the file-cstr branch from f1b6278 to d070bb7 Compare January 3, 2025 17:20

cramertj changed the title ~~Add Location::file_cstr~~ Add Location::file_with_nul Jan 3, 2025

cramertj force-pushed the file-cstr branch from d070bb7 to d28ac01 Compare January 3, 2025 17:21

Add Location::file_witH_nul

73810aa

This is useful for C/C++ APIs which expect the const char* returned from __FILE__ or std::source_location::file_name.

cramertj force-pushed the file-cstr branch from d28ac01 to 73810aa Compare January 3, 2025 17:24

cramertj mentioned this pull request Jan 3, 2025

ACP: Add nul-terminated version of core::panic::Location::file rust-lang/libs-team#466

Open

ojeda mentioned this pull request Jan 3, 2025

core wanted features & bugfixes Rust-for-Linux/linux#514

Open

13 tasks

oli-obk reviewed Jan 7, 2025

View reviewed changes

compiler/rustc_const_eval/src/util/caller_location.rs Show resolved Hide resolved

This was referenced Jan 7, 2025

Add nul-terminated filename for #[track_caller] #131828

Closed

Add unstable option to nul-terminate location strings #135240

Open

This comment has been minimized.

Sign in to view

rustbot added the S-waiting-on-perf Status: Waiting on a perf run to be completed. label Jan 8, 2025

This comment has been minimized.

Sign in to view

rustbot removed the S-waiting-on-perf Status: Waiting on a perf run to be completed. label Jan 8, 2025

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Add Location::file_with_nul #135054

Add Location::file_with_nul #135054

cramertj commented Jan 3, 2025

rustbot commented Jan 3, 2025

rustbot commented Jan 3, 2025

Noratrieb commented Jan 3, 2025

RalfJung Jan 3, 2025 •

edited

Loading

cramertj Jan 3, 2025

cramertj commented Jan 3, 2025

bjorn3 commented Jan 5, 2025

cramertj commented Jan 6, 2025

Darksonn commented Jan 7, 2025

oli-obk Jan 7, 2025

cramertj Jan 7, 2025

oli-obk Jan 7, 2025

Darksonn Jan 7, 2025

oli-obk Jan 8, 2025

RalfJung Jan 8, 2025

Darksonn Jan 8, 2025

cramertj commented Jan 7, 2025

Darksonn commented Jan 8, 2025 •

edited

Loading

Noratrieb commented Jan 8, 2025 •

edited

Loading

This comment has been minimized.

bors commented Jan 8, 2025

bors commented Jan 8, 2025

This comment has been minimized.

rust-timer commented Jan 8, 2025

Add Location::file_with_nul #135054

Are you sure you want to change the base?

Add Location::file_with_nul #135054

Conversation

cramertj commented Jan 3, 2025

rustbot commented Jan 3, 2025

rustbot commented Jan 3, 2025

Noratrieb commented Jan 3, 2025

RalfJung Jan 3, 2025 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

cramertj commented Jan 3, 2025

bjorn3 commented Jan 5, 2025

cramertj commented Jan 6, 2025

Darksonn commented Jan 7, 2025

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

cramertj commented Jan 7, 2025

Darksonn commented Jan 8, 2025 • edited Loading

Noratrieb commented Jan 8, 2025 • edited Loading

This comment has been minimized.

bors commented Jan 8, 2025

bors commented Jan 8, 2025

This comment has been minimized.

rust-timer commented Jan 8, 2025

Overall result: ❌ regressions - no action needed

RalfJung Jan 3, 2025 •

edited

Loading

Darksonn commented Jan 8, 2025 •

edited

Loading

Noratrieb commented Jan 8, 2025 •

edited

Loading