This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [C++ PATCH] Make parser revert digraph "<:"
Jason Merrill <jason@redhat.com> writes:
| On 26 Jan 2004 20:21:19 +0100, Gabriel Dos Reis <gdr@integrable-solutions.net> wrote:
|
| > Jason Merrill <jason@redhat.com> writes:
| >
| > | begin, v.tr.
| > | 2. To cause to come into being; originate.
| > | 3. To come first in: "The numeral 1 begins the sequence."
| > |
| > | In the sense of definition 3, textually, both "<" and "<::" begin the
| > | template-argument-list. To me, this is the obvious interpretation of the
| > | message.
| > |
| > | Are you thinking in terms of definition 2?
| >
| > Yes.
| >
| > When reading a program text, how do a programmer knows that a
| > template-argument list starts? He/she looks for '<' in a pattern like
| > <xxx>
| > (don't tell me you look for '<::' :-/)
|
| Yes, we agree that "'<::' cannot begin a template-argument-list" doesn't
| make sense under definition 2 above. And therefore, when I read that
| message, my brain rejects definition 2 and settles on definition 3, which
| does make sense.
And the issue is which definition is most common, not which we can
choose to make an obscure message somehow meaningful. I rather fix
the obscure message.
| Words are ambiguous. There are usually many possible interpretations, only
| some of which make sense in context. In this case, there is only one
| interpretation which makes sense in context, so that's the one I choose to
| use. Clinging to an interpretation which doesn't make sense, and then
| complaining about it not making sense, seems perverse to me.
That may be perverse, but that is NOT what is happenning here.
-- Gaby