In my latest attempt to build and test this, .../test/Semantics/call19.f90 is failing because it's producing extra messages regarding entities that are "not a procedure". But I trust you to make sure that everything works after integration with the latest sources.

For what it's worth, here's my cmake command that results in a successful static full build:

In D104305#2822589, @klausler wrote:

The non-shared-library build of the current llvm-project/main tree fails on this new test; it looks as if gfortran is being invoked and passed an option that it does not recognize. The shared library build passes all tests.

All builds, tests, and looks good.

Things build OK, but I when I run the tests, I get failures in implicit07.f90 and resolve29.f90.

All builds, tests, and looks good.

All builds, tests, and looks good.

All builds, tests, and looks good.

Looks great! Thanks for doing this.

Stay tuned a little, I think I spotted the problem ...

Thanks for doing this. These changes look good, but I can't get them to build successfully. I'm using the GNU 9.3 C++ compiler. My build fails to link the bin/flang-new executable. Here's an excerpt of the error messages:

Responding to Peter's comments.

Responding to comments from Peter.

All builds, tests, and looks good.

@awarzynski, I'm not sure what a good transition period would be. Actually, I don't see the benefit of removing any functionality from f18. This might be a good topic to bring up at the Wednesday f18 meeting.

All builds, tests, and looks good.

All builds, tests, and looks good.

All builds, tests, and looks goodl.

All builds, tests, and looks good.

All builds, tests, and looks good.

All builds, tests, and looks good.

All builds, tests, and looks good.

All builds, tests, and looks good.

All builds, tests, and looks good.

All builds, tests, and looks good.

In D103177#2782459, @klausler wrote:

I'm not clear why you plan to remove code from f18.cpp.

All builds, tests, and looks good.

All builds, tests, and looks good.

All builds, tests, and looks good.

All builds, tests, and looks good.

I removed changes to flang/lib/Optimizer/CodeGen/PreCGRewrite.cpp, since it looks like they will cause problems for Doxygen, and I can't currently test for them.

Thanks for your comments, @awarzynski. The version of clang-format that I build from the latest sources is 13.0.0. I've been in the habit of the last many months to always use the latest version of clang-format, but I'm not sure if there's general agreement on that policy. I'll post a message on the mailing list about this.

In D102424#2758592, @sscalpone wrote:

@PeteSteinfeld Doc to build docs https://github.com/llvm/llvm-project/tree/main/flang#how-to-generate-documentation

All builds, tests, and looks good.

Actually, I got a diff when I ran clang-format on check-io.cpp.

All builds, tests, and looks good.

In D102259#2752174, @klausler wrote:

Instead of imposing an arbitrary limit on the size of an expression it would be better to change the algorithm to handle larger trees with more use of iteration and tail recursion, I think.

Thanks for the feedback, Jean!

I have tested full shared and non-shared builds with this change, and both look good.

@Renaud-K, I'm not sure what the use-case is that this fixes. Can you please describe it in more detail, and I'll try a build with that exact use case to verify that this fixes it.

All builds, tests, and looks good.

All builds, tests, and looks good to me.

Responding to Peter's comments.

Responding to Peter's comments.

Responding to Peter's comments. I totally changed my implementation to be as
described in the current commit description.