SynHighlighterPas: Bug with generics

[Pascal]

SynHighlihterPas has some problems with generics (see screenshot of Generics.Collections).

I had a look at the highlighter and was wondering why beginning a FoldBlock after rsAfterEqualOrColon
instead of rsAfterEqual?

Code: Pascal  [Select][+][-]

1. function TSynPasSyn.Func54: TtkTokenKind;
2. begin
3.   if KeyComp('Class') then begin
4.     Result := tkKey;
5.     if (rsAfterEqualOrColon in fRange) and (PasCodeFoldRange.BracketNestLevel = 0)
6.     then begin
7.       fRange := fRange + [rsAtClass] - [rsVarTypeInSpecification];
8.       StartPascalCodeFoldBlock(cfbtClass);
9.     end;
10.   end
11.   else
12.     Result := tkIdentifier;
13. end;
14.  

I actualy don't know of a case where to begin a new foldblock for keyword class after a colon.
So maybe we just can change rsAfterEqualOrColon to rsAfterEqual?

Pascal

[Martin_fr]

I had a look at the highlighter and was wondering why beginning a FoldBlock after rsAfterEqualOrColon
instead of rsAfterEqual?

Probably that code was written before rsAfterEqual was introduced.

But yes it, I think it can be changed.

Please check if the test case still passes. (and add a new test for this   )

[Pascal]

Tests are okay so far. New test is in the works 

[Pascal]

Martin,

could you have a look at the patch. Would this be a suitable test?

[Martin_fr]

looks ok.

Code: Pascal  [Select][+][-]

1.  CheckNode( 3, [], 0,   0,   17, 22,   3, 4,   3, 4,
2.                                   cfbtClass, cfbtClass,  FOLDGROUP_PASCAL,
3.                                   [sfaOpen, sfaFold, sfaFoldFold, sfaMultiLine, sfaMarkup, sfaOpenFold]);
4.  

Because the above checks the x pos, so it checks that the fold is triggered by the correct of the 2 "class" keywords on that line.

Applied the patch.

[Pascal]

I've found an other one 

[Pascal]

Could this patch be a solution? Tests still run fine.

[Martin_fr]

Yes seems ok.

[Pascal]

Okay, i'll also update the test.

[Pascal]

Okay, here is the patch (with changed test).

[Pascal]

... and an other one:

Patch includes last patch!

[Martin_fr]

The 2 latter ones are actually a result of the first patch.

rsAfterEqualOrColon is cleared in function Next.

rsAfterEqual is not cleared.


Same breakage will be for

Code: Pascal  [Select][+][-]

1. foo = object
2.  class function...

I'll have another look if rsAfterEqual  should be cleared in Next too.

IIRC rsAfterEqual  was introduced to handle the ^ symbol.
Either pointer, or char #3 = ^C

[Pascal]

I've tested to clear rsAfterEqual in Next and it seems to work (including new tests for the last two issues).

[Martin_fr]

I fixed it in class/object/record.

I added more tests, the fix in "next" would have broken
var a:string=^a^b;

[Pascal]

Oklay, thanks.