feat(tablegen): add MLIR TableGen grammar

[RIvance]

https://llvm.org/docs/TableGen
https://llvm.org/docs/TableGen/ProgRef.html

[kaby76]

Perhaps you could remove the extra parentheses here and in a few other places?

$ bash /c/Users/Kenne/Documents/GitHub/g4-scripts/find-useless-parentheses.sh
Using built-in parser.
CSharp 0 TableGen.g4 success 0.0477444
Computing useless parentheses in TableGen.g4
Using built-in parser.
CSharp 0 TableGen.g4 success 0.0477226
Found useless parentheses in grammars...
TableGen.g4:L100:     :   positionalArgValueList (',')? namedArgValueList
                                                 ^
TableGen.g4:L333:     :   '(' dagArg (dagArgList)? ')'
                                     ^
TableGen.g4:L460: DecimalInteger  :  ('+' | '-')? (UDigit)+;
                                                  ^
TableGen.g4:L464: TokIdentifier   :  (UDigit)* UAlpha (UAlpha | UDigit)*;
                                     ^
TableGen.g4:L455: TokString       : '"' (~('"') | Escape)* '"';
                                          ^
TableGen.g4:L457: TokCode         : '[{' (~(']') | ']' ~('}') | Escape)*? '}]';
                                           ^
TableGen.g4:L457: TokCode         : '[{' (~(']') | ']' ~('}') | Escape)*? '}]';
                                                        ^
05/25-06:29:01 ~/issues/g4-4107/tablegen/Generated-CSharp

[kaby76]

Field (and fragment EOL below) are unused. Is something missing?

$ bash /c/Users/Kenne/Documents/GitHub/g4-scripts/find-unused-lexer-symbols.sh
No arguments were provided.
Finding unused lexer symbols in grammars...
./desc.xml
1st pass: fold string literals in parser rules to make explicit lexer rule used.
2nd pass: find unreferenced lexer rule symbols.
Field
EOL
05/25-06:32:18 ~/issues/g4-4107/tablegen/Generated-CSharp

[kaby76]

The coverage for this one test is ~50%, but perhaps you could add additional tests to cover more of the grammar?

$ trcover.exe ../examples/ToyOps.td
Parsing grammar C:/msys64/home/Kenne/issues/g4-4107/tablegen/Generated-CSharp/TableGen.g4
CSharp 0 ../examples/ToyOps.td success 0.0336701
Analyzing...
Analysis 0 ../examples/ToyOps.td 0.0115427
05/25-06:39:13 ~/issues/g4-4107/tablegen/Generated-CSharp

Output: cover.html.txt

[teverett]

Looking at the build results, @kaby76 , is this an bug in the Antlr Python generator?

Traceback (most recent call last):
  File "/Users/runner/work/grammars-v4/grammars-v4/tablegen/Generated-Python3/Test.py", line 8, in <module>
    from TableGenParser import TableGenParser;
  File "/Users/runner/work/grammars-v4/grammars-v4/tablegen/Generated-Python3/TableGenParser.py", line 3[80](https://github.com/antlr/grammars-v4/actions/runs/9233773832/job/25415776942?pr=4107#step:24:81)6
    return self.getToken(TableGenParser.True, 0)
                                        ^^^^

[kaby76]

The problem is rule True. It's a symbol conflict. 'True' needs to be renamed with a trailing '_', or with some other naming scheme, e.g. 'KW_' Probably False has a problem as well.

[teverett]

@kaby76 thanks! Hey @RIvance could you make the fix in your PR?

[spannella]

Hey this would still be a really nice addition, let me know I can easily help.