Landed in LLVM commit llvmorg-13-init-3801-ge5d958c45629.

• adjust the test to work on Windows as well

@jasonliu Please see my inline comments for the corner case and the left FIXME. Thanks for your review.

1: add more comment

LGTM if the nit comment is addressed
There are two items that we need to follow up after this patch:

1. The corner case/bug in https://reviews.llvm.org/D95518#inline-919030
2. The possibility of addressing the FIXME for https://reviews.llvm.org/D95518#inline-918884

In D82363#2603093, @thakis wrote:

This fails on bots that have a non-x86 default triple: http://45.33.8.238/macm1/4580/step_10.txt

Probably just needs an explicit triple.

Please take a look, and revert for now if it takes a while to fix.

This fails on bots that have a non-x86 default triple: http://45.33.8.238/macm1/4580/step_10.txt

• adding a test

Added a test case for DIBuilder

1: address @jasonliu comments

Rebased onto latest master, numerous small cleanups.

Rebased onto latest master, address review comments.

@StephenTozer Thanks!

1: address comments

The fundamentals of the patch look solid; minor point about the HTML templating, and also I think there should be a couple of tests for the llvm-original-di-preservation script - nothing too complicated, just verifying that it produces the correct HTML for a given input, as with the opt-viewer tests.

Thank you @rnk.

some simplify to the test case

address @jasonliu comments:
1: handle explicit section
2: add more testcases

1: format fixing

This is probably more integration testing than I think optnone needs, but I don't see any reason to say no to it.

typo fix

1: handle neighbouring .loc

thanks for your review @hubert.reinterpretcast . See my inline reply.

address @hubert.reinterpretcast comments

Ping :)

Ping :)

Here's an example that seems to work:

In D97311#2588856, @aprantl wrote:

Have you considered instead introducing a bool isConstant field in DILocalVariable? This way we could reuse all the machinery for local/global variables and only need to switch out the TAG in AsmPrinter.

gentle ping...@hubert.reinterpretcast

1: make sure what we end is the .text section

In D97311#2588877, @aprantl wrote:

Guess it gets a bit complicated if it has to support non-local constants too.

We can already represent optimized out global variables with constant values as !DIGlobalVariableExpression(!DIGlobalVariable(...), !DIExpression(DW_OP_constu, ...)).
We would have to also add a bit to DIGlobalVariable, but other than that this seems perfectly feasible?

Please run tests before committing.

Guess it gets a bit complicated if it has to support non-local constants too.

In D97311#2588856, @aprantl wrote:

Have you considered instead introducing a bool isConstant field in DILocalVariable? This way we could reuse all the machinery for local/global variables and only need to switch out the TAG in AsmPrinter.

Have you considered instead introducing a bool isConstant field in DILocalVariable? This way we could reuse all the machinery for local/global variables and only need to switch out the TAG in AsmPrinter.

LGTM, but pls2leave it 24 hours or so in case @rnk chimes in.

Rebase.

Thanks @rnk! I will land this shortly.

Resolved the previous mentioned issue. Ready for review.

• Rebase on top of trunk

• Rebase on top of trunk

Also fails locally for me in a normal cmake build. Reverted for now in f14a14dd2564703db02f80c00db8ae492b594f77.

(breaks check-llvm more on windows: http://45.33.8.238/win/33808/step_11.txt)

This breaks check-llvm: http://45.33.8.238/linux/40166/step_12.txt

@ikudrin Hi, I start to commit these NFC changes as now all of them are approved. Welcome your post-commit comment if there is any. Thank you again for your review!

lgtm

A few typos. Nice to have these!

LGTM. Thanks!

1: update due to change in D96409 and D95998

1: use more meaningful variable name.

1: put "_" in the function body to avoid strange string parameters