Sergi Siso

Isn't this overwriting DIRF(ITH), so #423 would also fix this? (maybe you already knew since I see you also commented on that issue)

@arporter This is ready for a first review, it continues generalising the remaining NEMO transformations. ArrayAccess2Loop was mostly used in the SIR transformation scripts. I have not run the generated...

@arporter This is ready for another review. I also removed a couple of configuration utility functions that Codecov said were not used anymore (and was planning to remove in an...

See a branch with a method to retrieve the original source code line: https://github.com/stfc/PSyclone/tree/2062_origin_string As you can see a successful retrieval needs to pass: ``` if self._ast: if hasattr(self._ast, 'item'):...

@hbrunie your usage is correct. I fixed some issues, if you pull the branch now it should return the line number from everything inside the loop statements. However, the loops...

@hbrunie the filename should also be in the string now (for statements). I have created a PR to put this functionality into the master branch even if it doesn't support...

#2511 Fixes the issue when the array.datatype is an array of character type. But I will not close this issue yet because it is still a problem when the type...

As we talk about, these issues are also present in the arrayrange2loop transformation, although we correctly refuse them in the validate because they contain non-elemental intrinsics. If we could convert...

@arporter Having a second thought on the previous comment, wouldn't the array to loop conversion from ``` a(:) = b(:) * sum(c, dim=2 ) ``` be achieved without a temporary...

PR #2372 fixes some bugs on the WHERE parsing and adds the construct into a CodeBlock when there is a reduction that we can not safely canonicalise. This prevents the...