Fix thin archive reading

[bjorn3]

This includes a revert of #105221 to restore fat archive reading with LlvmArchiveBuilder.

Should fix #107162, #107334 and google/shaderc-rs#133

[rustbot]

r? @cjgillot

(rustbot has picked a reviewer for you, use r? to override)

[bjorn3]

@bors try

to provide a toolchain that can be used for testing if this indeed fixed it. This will build a linux toolchain, if necessary I can do a windows toolchain after that.

[bors]

⌛ Trying commit de363d5 with merge 87b60fa8c61c1fc5243bb1bce530da8f3f694d4d...

[bors]

☀️ Try build successful - checks-actions
Build commit: 87b60fa8c61c1fc5243bb1bce530da8f3f694d4d (87b60fa8c61c1fc5243bb1bce530da8f3f694d4d)

[bjorn3]

This indeed fixes reading thin archives. At least on Linux:

$ rustup-toolchain-install-master 87b60fa8c61c1fc5243bb1bce530da8f3f694d4d
[...]
$ echo | rustc - --emit obj --crate-type lib # create dummy object file
$ ar rsT libthin.a rust_out.o # create thin archive
$ echo | rustc +stable -l static=thin - --crate-type staticlib -L .
error: failed to add native library ./libthin.a: Unsupported archive identifier

error: aborting due to previous error

$ echo | rustc +87b60fa8c61c1fc5243bb1bce530da8f3f694d4d -l static=thin - --crate-type staticlib -L .
$ echo $?
0

Going to do a windows try build too as most reports are coming from windows.

@bors try

[bors]

⌛ Trying commit 3020ac1d605aabca87a39d218cdbc67c910f4f22 with merge e65225bb3f98a0001676c0cc139656e7209590c2...

[bors]

☀️ Try build successful - checks-actions
Build commit: e65225bb3f98a0001676c0cc139656e7209590c2 (e65225bb3f98a0001676c0cc139656e7209590c2)

[bjorn3]

@jdm, @sagudev, @jpalus and @kanerogers can you try if this fixes the regression for you? You can use rustup-toolchain-install-master 87b60fa8c61c1fc5243bb1bce530da8f3f694d4d -c cargo to install the linux try build and rustup-toolchain-install-master e65225bb3f98a0001676c0cc139656e7209590c2 -c cargo to install the windows try build if you have rustup-toolchain-install-master installed from crates.io.

Edit: after that you also need to set it as default toolchain using rustup default <commit>.
Edit2: fixed rustup-toolchain-install-master command

[sagudev]

Can confirm it is working for mozjs on windows (apparently we use thin only for windows builds).
Here is nightly build for control test (it is expected to fail).
Here is build using e65225bb3f98a0001676c0cc139656e7209590c2 toolchain.

[jpalus]

@bjorn3 confirmed this PR (specifically de363d5) fixes #107334. Thanks.

[bjorn3]

Thanks for testing!

[kanerogers]

Thanks so much for the quick turnaround @bjorn3!

I can confirm that e65225bb3f98a0001676c0cc139656e7209590c2 fixes google/shaderc-rs#133 on Windows. 👍🏻

[bors]

⌛ Testing commit de363d5 with merge 2527416...

[bors]

☀️ Test successful - checks-actions
Approved by: wesleywiser
Pushing 2527416 to master...

[rust-timer]

Finished benchmarking commit (2527416): comparison URL.

Overall result: ❌ regressions - ACTION NEEDED

Next Steps: If you can justify the regressions found in this perf run, please indicate this with @rustbot label: +perf-regression-triaged along with sufficient written justification. If you cannot justify the regressions please open an issue or create a new PR that fixes the regressions, add a comment linking to the newly created issue or PR, and then add the perf-regression-triaged label to this PR.

@rustbot label: +perf-regression
cc @rust-lang/wg-compiler-performance

Instruction count

This is a highly reliable metric that was used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	0.4%	[0.3%, 0.6%]	13
Regressions ❌ (secondary)	0.5%	[0.4%, 0.5%]	2
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	0.4%	[0.3%, 0.6%]	13

Max RSS (memory usage)

Results

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)	1.4%	[1.4%, 1.4%]	1
Regressions ❌ (secondary)	2.7%	[2.7%, 2.7%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	1.4%	[1.4%, 1.4%]	1

Cycles

This benchmark run did not return any relevant results for this metric.

[pnkfelix]

Discussed in today's T-compiler triage meeting.

@rustbot label: beta-accepted stable-accepted

beta backport is PR #107609

stable backport is PR #107611

(I still would like to try to make regression tests, but that need not block the backports.)