LGTM. Thanks for the fix.

Thanks a lot for the feedback, it's easy to miss such silly mistakes.

Thanks a lot for the feedback! I've changed accordingly to use -o now. Hopefully I got the header guards right this time.

Thanks for the feedback. Updated accordingly.

Thanks for the feedback! I am still working on some of the things you've pointed out, however I have managed to add tests.
Unfortunately, since the plugin outputs a file with the same name every time, this causes some of the tests to fail because they're ran in parallel.
I will have a patch which adds the source file's name to the one generated by the plugin so that there\' a unique file each time.

@Meinersbur is this one ready to be merged also?

@Meinersbur is this one ready to be merged?

Rebase to mollify buildbots.

Rebased to mollify buildbots.

Rebased to stop a Clang test failure from some other change in main.

Re-enabling some tests.

Re-enabling tests on Windows.

Rebased to add support for new tests and also see if "resolve67.f90" passes the test.

Commenting so that my inline comment shows.

This fixed the issue, I checked and none of the tests fail now. Thanks a ton.
LGTM.

Rebased to include newly added tests.

Using braced initialisation for the lambda.

Updating to use braced initialisation, removed parentheses in the return statement and converted the captures into lambda parameters.

@Meinersbur would it be fine to also disable these failings tests on Windows? I have filed a bug report on bugzilla[1] for D107956. I could include them there or file a new one.

Updated to disable the 2 failing tests on Windows and also filed a bug report in bugzilla.

Specified encoding to fix a failing test on Windows.

In D107041#2953476, @ashe wrote:

Thanks, that fixes the weird diff formatting. But why are all of the spaces removed?

@Meinersbur is this ready to be merged?

Rebased and added support for 3 new tests.

Hopefully fixes the few failing tests due to a regex issue.

Would it be fine for me to disable to failing tests for now and have them re-enabled with the fix?

"folding21.f90" was added a few days ago and fails in all cases. It's not been set as "XFAIL" so I am unsure if that's something on my end.

Updated to use tempfile.TemporaryDirectory. Addressed comments by Andrzej.

In D107954#2941348, @awarzynski wrote:

Makes a lot of sense, but I would rephrase the summary a bit. Basically, you are making sure here that you are comparing the files line by line rather than the whole thing at a time. By making sure that the comparison is done line-by-line, you also make the output much nicer. Perhaps:

In D107575#2944131, @Meinersbur wrote:

The evaluation order is indeed undefined for operators except || and &&. I suggest to store the results in a variables. Them being part of a if-else will require some restructuring.

In D107956#2947399, @Meinersbur wrote:

Would you like me to look at why "modfile18.f90" and "modfile20.f90" are failing?

Updated to use join with unified_diff.

Fixed modfile09.f90 failing. Updated to use the tempfile library. Using Python-style docstrings now.

In D107953#2941417, @Meinersbur wrote:

Is there a reason to use lit's suggestion for a temporary directory instead using tempfile.TemporaryDirectory or tempfile.mkdtemp?

In D107956#2941438, @Meinersbur wrote:

Is the failure of modfile09.f90 related to D107953?

"resolve12.f90" and "resolve26.f90" have been fixed. Addressed feedback from Andrzej.

Changed to be less intrusive.

Fixed a bug in a test file. And also updated the scripts to use docstrings instead of normal comments.

In D107575#2929026, @Meinersbur wrote:

There seems to be a problem in how import common not finding common.py in a sibling directory.

Incorporating feedback from Kiran and Andrzej.

Changes for more idiomatic python.

Incorporated Michael's feedback into a small patch.