This is the mail archive of the
mailing list for the Archer project.
Re: [RFC] Koenig lookup patch 3
- From: Tom Tromey <tromey at redhat dot com>
- To: Sami Wagiaalla <swagiaal at redhat dot com>
- Cc: Project Archer <archer at sourceware dot org>
- Date: Mon, 09 Nov 2009 12:29:04 -0700
- Subject: Re: [RFC] Koenig lookup patch 3
- References: <49BABABE.firstname.lastname@example.org> <email@example.com><49F87751.firstname.lastname@example.org> <email@example.com><4A969900.firstname.lastname@example.org> <email@example.com><4AD6340F.firstname.lastname@example.org> <4AD63BA4.email@example.com><firstname.lastname@example.org> <4AEF2E3C.email@example.com>
- Reply-to: Tom Tromey <tromey at redhat dot com>
>>>>> "Sami" == Sami Wagiaalla <firstname.lastname@example.org> writes:
Tom> It seems like these additions would introduce a parser ambiguity,
Tom> because UNKNOWN_NAME could be parsed as adl_func or as name_not_typename
Tom> (and hence variable).
Tom> Is bison silent about this?
Sami> Good point. Same goes for everything that is common between name and
Sami> name_no_typename. Anyways I have changed this to:
It still seems like the new productions are ambiguous, because
UNKNOWN_NAME alone is an `exp'. If bison doesn't give more warnings
after the patch, though, then I wouldn't worry about it.
Sami> * c-exp.y: Created token UNKOWN_NAME.
Sami> Created grammar rules for UNKOWN_NAME.
Typo: should be `UNKNOWN'.
Other than that this looks good to me.
I was curious to know what happens when the current language is C, not
C++. Do we still attempt ADL?