This is the mail archive of the archer@sourceware.org mailing list for the Archer project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Meeting Agenda 2008-08-27


>>>>> "Sami" == Sami Wagiaalla <swagiaal@redhat.com> writes:

Sami> ie the lexer will create tokens NAME COLONCOLON NAME, and not specify
Sami> that name is a TYPENAME or a symbol NAME. Then in the parser have a
Sami> rule like this

Just a bit of background here...

In C this lexer trick is called the "typedef hack".  Basically the C
grammar is ambiguous.  E.g., consider:

    (x) -y

If x is a typedef, then this is a cast of a unary negation.
If x is a variable, then this is a subtraction.

So, it became common practice in C to check names while lexing.

There are various options for fixing this... but just removing this
hack from the lexer is not enough.

Tom


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]