fix: ensuring no segfault when c library function returns oversized string

[Angus-Bethke-Bachmann]

Changed

• Ensuring no segfault when c library function returns oversized string
• Added a new field to string data type information that specifies if user supplied a size in source

Testing

• Added tests for specific case in call lowerer
• Added lit test to verify the truncation case
• Added lit test to verify the c library case

[github-actions]

Build Artifacts

🐧 Linux

Artifact	Link	Size
deb-x86_64	Download	38.3 MB
schema	Download	0.0 MB
stdlib	Download	32.3 MB
plc-x86_64	Download	43.4 MB
deb-aarch64	Download	30.7 MB
plc-aarch64	Download	43.2 MB

From workflow run

🪟 Windows

Artifact	Link	Size
stdlib.lib	Download	4.0 MB
stdlib.dll	Download	0.1 MB
plc.exe	Download	38.2 MB

From workflow run

[mhasel]

The logic in this match arm seems like it duplicates the _ arm of is_output_assignment_and_type_cast_needed. Can we extract this into a function and use them in all applicable predicates here?

[Angus-Bethke-Bachmann]

With more changes to address the above these have diverged far enough apart that this is no longer necessary. But I'm happy to spend more time trying to streamline this file if you'd like?

[mhasel]

Previous comments have all been addressed, however a regression slipped in with the guessing heuristic approach

[mhasel]

Instead of this guessing heuristic, can't we thread the correct index through here? I think the lowerer should deterministically know it (either aggregates or fn-pointers, right?)

I think this also caused a regression where it now scans if there is a VAR_OUTPUT either left or right, which can cause mis-fires on arguments that are not VAR_OUTPUTs at all. If you have an aggregate VAR_INPUT that is larger than a neighboring VAR_OUTPUT, it will get a temp variable instead, leading to incorrect behaviour.

Regression example:

FUNCTION Config_GetString : DINT
      VAR_INPUT  Key   : STRING[1023]; END_VAR
      VAR_OUTPUT Value : STRING[28];   END_VAR
      Value := 'hi';
  END_FUNCTION

  PROGRAM mainProg
      VAR keyVar : STRING[10]; outVar : STRING[28]; r : DINT; END_VAR
      keyVar := 'my-key';
      r := Config_GetString(keyVar, outVar);
      PRINTF('key=[%s]$N', keyVar); // will print 'key=[]', `keyVar` passed to VAR_INPUT `Key` was replaced by a temp
  END_PROGRAM

  FUNCTION main
      mainProg();
  END_FUNCTION

[Angus-Bethke-Bachmann]

Yeah... that was a terrible idea 😞 I've switch it to look at the annotation instead to get the correct parameter index. Only snag was that the lowerer could sometimes change the parameter indexes if a new parameter was inserted, so I added an explicit flag for that. This should cover all the bases (for now anyway).