Skip to content

Conversation

@sergisiso
Copy link
Collaborator

Change Ref2ArrayRange behaviour by letting it pass if the provided node is already in the expanded form, but making it fail if for a expression we cannot guarantee if it could be expanded or not. This will bring validations that now need to be done every time its used to inside the transformation.

@codecov
Copy link

codecov bot commented Dec 15, 2025

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 99.91%. Comparing base (97dd02d) to head (c350586).

Additional details and impacted files
@@            Coverage Diff             @@
##           master    #3260      +/-   ##
==========================================
- Coverage   99.91%   99.91%   -0.01%     
==========================================
  Files         376      376              
  Lines       53529    53522       -7     
==========================================
- Hits        53484    53477       -7     
  Misses         45       45              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:
  • ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

@sergisiso sergisiso self-assigned this Dec 17, 2025
@sergisiso sergisiso changed the title (closes #2884) Change Ref2ArrayRange behaviour (closes #2884) Allow WHERE with imported symbols by changing Ref2ArrayRange behaviour to do the validations Dec 17, 2025
@sergisiso
Copy link
Collaborator Author

sergisiso commented Dec 17, 2025

@arporter @LonelyCat124 This is ready for review. It changes the Reference2ArrayReference to fail if the outcome cannot be guaranteed, so thing like WHERE can rely on its validation.

This will be followed by #1858 (structures) and #2722 (dependencies), which can be done in a single location after this PR, but it was to big to do here.

Copy link
Member

@arporter arporter left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Really nice Sergi, it's good to see the simplification that this has achieved.
I'm worried that we are no longer raising an exception when we see a StructureReference - given that we can't handle them this needs to be flagged?
Apart from that, it's just minor tidying. I'll run the integration tests next time (or you could fire them off perhaps once you're done).

# TODO issue #1858. Add support for structures containing arrays.
if node and node.parent and isinstance(node.parent, Call):
if node.position == 0:
return
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Comment please - I didn't readily understand at first but presumably this is checking that the node isn't the Reference associated with the routine target of a Call?

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Exactly. Added comment

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

You probably don't need the and node.parent part of L120 as, if it's None, it won't pass the isinstance(node.parent, Call)?

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Removed

trans.validate(reference)
assert ("node is passed as an argument to a Call to a non-elemental "
"routine (DEALLOCATE(a)" in str(info.value))
trans.validate(assign.lhs)
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The docstring says we test that a StructureReference raises an exception but clearly it doesn't any longer. Given that we don't support them yet, I would argue we should raise an exception as we don't know what to do with them.

Copy link
Collaborator Author

@sergisiso sergisiso Dec 19, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It felt bad to have the comment that are not supported and then just return the validation.

The reason I did it is because both the WHERE and the ArrayAssignment2Loop already implement and have several tests for structure references (which they happily take at face-value, without checking if they need to be added ranges).
Adding this restriction here, without implementing proper support for letting valid cases pass, and array types extended, makes all of those fail.

Also this is not worse than master, in master it had an exception but it was "except:pass" everywhere it was used.

I suggest I tackle this in a follow up (I attempted starting it but it quickly grows very large), or we can block this for now. Let me know what you prefer.

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I implemented it in #3266

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks Sergi. As you're in the process of doing #3266, please could you just update this test with TODO to that PR (or would you like to get that PR on first?).

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

or would you like to get that PR on first?

I based that PR after this one's HEAD, so I can't change the order. The other PR is ready to go once this one is merge.

I added the TODO just before the test to avoid conflicts as I already modified this test significantly in the follow up.

@sergisiso sergisiso force-pushed the 2884_change_ref2arrayrange_behaviour branch from 99ab3e0 to 3e1c6b8 Compare December 19, 2025 13:33
@sergisiso
Copy link
Collaborator Author

@arporter This is ready for another look, I just fired the integration test.

Copy link
Member

@arporter arporter left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks Sergi. Just a bit more tidying to do and we have some indirect coverage changes:
image
I think you can probably just remove the try...except from that bit of code?

to do so (e.g. it won't convert call arguments because it would change the
bounds values).
Note that if the provided node does not need to be modified is provided (
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

s/is provided// but I can do this if there's nothing else.

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Done

# TODO issue #1858. Add support for structures containing arrays.
if node and node.parent and isinstance(node.parent, Call):
if node.position == 0:
return
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

You probably don't need the and node.parent part of L120 as, if it's None, it won't pass the isinstance(node.parent, Call)?

'''Test that a reference in a PointerAssignment raises an exception. '''
def test_pointer_assignment(fortran_reader, fortran_writer):
''' Test that a reference in a PointerAssignment raises an exception
Pointer and target attributes (currently represented by partial_datatype
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pls close the parentheses after "partial_datatype"

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Done

trans.validate(reference)
assert ("node is passed as an argument to a Call to a non-elemental "
"routine (DEALLOCATE(a)" in str(info.value))
trans.validate(assign.lhs)
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks Sergi. As you're in the process of doing #3266, please could you just update this test with TODO to that PR (or would you like to get that PR on first?).

@sergisiso
Copy link
Collaborator Author

I think you can probably just remove the try...except from that bit of code?

I already tried this but it didn't work. Unfortunately psyad is unsafe in master as it considers any imported symbol is scalar or a non-elemental call. This happens to be true but it is unsafe. Ultimately we need to remove the except as the transformation guarantees this but if I do it now there is a big regression because most of the psyad real code has imports that are unchecked. We need capabilities to follow imports in psyad and its build system. I modified a test to show this and created an issue and TODOs.

@sergisiso
Copy link
Collaborator Author

@arporter This is ready for another look

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants