A Rafael D Teixeira
Wed, 01 Aug 2001 13:00:32 -0300
Ravi Pratap M <firstname.lastname@example.org> wrote:
>>I think weŽll have a lot of work on two fronts:
>>1 - adapt your parser mappings to the new sintax of VB.NET
>Indeed. It turns out that the current VB6 grammar that we have in
>GB is rather neat so it should be quite easy to map it to VB.NET :-)
I looked at your Bison grammar files (for GB) and them compared with the Jay
grammar that Miguel wrote for mcs.
VB6 is organized over multiple types of files (.BAS, .CLS, .FRM, ...) and
your grammar has support for each one, VB.NET uses just .VB files with a
single sintax as C#, besides I donŽt know if Bison works well with Unicode
and C# lex code, therefore I propose to use Jay as Miguel and, if he
concedes, to start out with its C# grammar and adapt it having your Bison
Grammar as a guide. I have a two monitor setup that comes on handy on those
>>2 - adapt the parser to generate the same trees of MiguelŽs mcs so >> that
>>weŽll use the same backend to generate IL.
>>Yes, this makes sense :-)
Seeing what MiguelŽs jay file is already doing IŽm quite amused with the
fact that Microsoft made VB.NET more similar to C# than to VB6, I think the
trees and even the grammar files will be very similar. Not that VB.NET is
just a diferent 'wording' of C#: there are support for some constructs that
are specific, but the number is small.
Kunde, is interested in supporting VB6 syntax on Mono, I think that he can
start directly from what GB has achieved...
IŽve gained a second workstation at work, for .NET research and development
and so IŽm planning to effectively contribute some code to Mono soon.
If Miguel allows IŽll get his Jay file and mix with your Bison files as
outlined above. Send me a go-ahead when you can Miguel.
One question, what Unicode encoding our compilers will support? I think that
as C# has 2-byte chars the parser is already supporting Unicode, but I
didnŽt check if the lexer is reading it correctly from the files.
Another question, this one for Miguel, your lexer is suporting identifiers
with Unicode escaping? Things like 'my\u00ABclass.my\U00010BFFmethod()' that
are valid in C#?
Get your FREE download of MSN Explorer at http://explorer.msn.com/intl.asp