Rejected : Re: [cdt-patch] K&R C patch

[John Camelon <jcamelon@xxxxxxxxxx>]

Almost perfect ... except for the following
few points: 

1.  GNUCSourceParser#declarator()
you should defer your call/handling of countKnRCParms() until you are at
the point where you see that you are ready to consume function parameters
(i.e. right before the label "parameterDeclarationLoop". 

2.  Inside GNUCSourceParser#countKnRCParms()
you need to make sure that if you are optionally consuming a token that
it is represented somehow in the physical AST.  

3.  I would avoid returning early
from GNUCSourceParser#declarator() as the current strategy allows for things
to be added in only one place.  If I provided another additional language
variant construct that I needed represented in the IASTFunctionDeclarator
I would have to add it in two places.  

Andrew can perhaps give you some feedback
on the CVisitor changes.
Thanks again!

JohnC
www.eclipse.org/cdt

cdt-patch-admin@xxxxxxxxxxx wrote on 01/19/2005 01:27:30
PM:

> 
> This patch adds basic K&R C support to the GNUCSourceParser. 
> - added support for K&R C syntax 
> - added support for ASTProblemDeclaration with faulty K&R C syntax
> - added simple tests 
> 
> Although K&R C still isn't fully supported and the patch needs
the 
> following (and will be submitted as a separate patch): 
> - getDeclarations(IBinding) support for K&R C 
> - check types for K&R C with pointers, etc 
> - add test code from GCC Torture tests (mainly the 
> snavigator/demo/monop project) 
> - support semantic problems such as K&R C parameter declaration
that
> has a different type than the prototype 
> 
> Devin Steffler
> IBM's Eclipse CDT
> Ottawa (Palladium), Ontario, Canada 
> [attachment "cdt_core_tests_patch.txt" deleted by John 
> Camelon/Ottawa/IBM] [attachment "cdt_core_patch.txt" deleted
by John
> Camelon/Ottawa/IBM]