Changes requested in review.

In D94052#2478970, @ftynse wrote:

I have two requests:

1. use UnitAttr instead of BoolAttr
2. add tests for any user-visible error messages

(Potentially for a follow-up, we should check attributes in the op verifier, not in the translation, but I see there are other attributes checked in the translation).

LGTM.

LGTM.

Looks good. Thanks for doing this.

In D88909#2314781, @SouraVX wrote:

Thanks for the patch!
NIT:
A more reasonable commit message could be:
[NFCI][flang] Introduce new helper TODO for further development
Or something similar(aligned to intent) :)

fir.alloca ops should be hoisted to the entry block. Because Fortran is pass-by-reference, correctness will often simply require stack allocations. However, that said, in cases where alloca ops can be promoted to registers, they will be although that is disabled at the moment.

In D86305#2230794, @SouraVX wrote:

In D86305#2230724, @aprantl wrote:

I'm fine with whatever numbering makes your development the easiest. My only ask is that if we still have a gap after this, please document what nodes are expected to go there, so there 1) won't be any future conflicts and 2) we don't pointlessly reserve space.

I'll have to re-confirm with our team. Making sure we don't break anything is priority for us.
@schweitz do you have any reservations WRT this ?

In D86305#2230837, @SouraVX wrote:

In D86305#2230794, @SouraVX wrote:

In D86305#2230724, @aprantl wrote:

I'm fine with whatever numbering makes your development the easiest. My only ask is that if we still have a gap after this, please document what nodes are expected to go there, so there 1) won't be any future conflicts and 2) we don't pointlessly reserve space.

I'll have to re-confirm with our team. Making sure we don't break anything is priority for us.
@schweitz do you have any reservations WRT this ?

We don't have any reservations with bit 42 and 43. If @schweitz also don't have any reservations we can mark/document them to be usable by others.

You might want to add "[flang]" to the subject line as this change supports Fortran. Thanx.

Thanks, Sourabh. That will help sync these patches with our internal branch. Glad to see this get upstreamed.

Thanks for adding this.

In D85828#2213768, @sameeranjoshi wrote:

+1 @DavidTruby - For pointing the reason for MLIR to use .md files.
It's correct that LLVM uses Sphinx infrastructure to build docs for any sub-project except MLIR.

In D83522#2155799, @jurahul wrote:

Fix build failure

In D83522#2152131, @jurahul wrote:

Ping. Not sure if flang is accepting such changes, if not please let me know and I will abandon this.

In D83397#2139202, @echristo wrote:

You also still haven't replied with "what bot/compiler/etc is going to break by making this change". Would you please do that?

I've reverted these models from the upstreamed code for now. It will avoid the warnings.

In D83397#2139131, @echristo wrote:

In D83397#2139111, @schweitz wrote:

Hi Eric,

There is an active development branch for the flang middle end. https://github.com/flang-compiler/f18-llvm-project/tree/fir-dev

That's not part of the llvm project.

That code base is being upstreamed piecemeal. Not all of the code is upstreamed at this point. It is simply a false impression that code in the middle of being upstreamed is "unused" or "unnecessary". Since it not all of it is upstreamed, changing interfaces and support code in llvm-project directly is going to cause problems that can become hard to track and resolve while the upstreaming is ongoing.

It very much is unused and unnecessary as there are no pieces of that code in the repository.

Hi Eric,

This will not compile in our builds.

Thanks for pointing out those comments. Yes, the I/O runtime is much further along now. I'll update them.

Commit 5c02a2421b951a14d1294598d0082b32659c7082

Committed as bd2c3014e120949c4acc1d107ac71c4091f00a3d

In D82693#2120288, @jdoerfert wrote:

In the passes files there are 6 pass creators declared. I failed to find all the definitions in this patch.
It seems sensible to split this into multiple commits. I would also recommend tests for new functionality.

[EDIT: Took too long to write my comments, seems abandoned now]

Did you remove the test_any.sh script in one of these?