🚀 go-pugleaf

In article <44mk1m$jop@newshost.lanl.gov>,
Tanmoy Bhattacharya <tanmoy@qcd.lanl.gov> wrote:
>In article <DANIELV.95Oct1122701@crt0.crt>, danielv@crt0.crt (Daniel
>Villeneuve) writes: 
>|> As I'm reading the lexical rules of the standard, it appears to me that, as
>|> early as translation phase 3, the file is decomposed into a stream of
>|> preprocessing tokens and the `sizeof' string is then converted into the
>|> `operator' token as listed in section 3.1.5.

I misread this before, and I apologise. There is indeed a problem.

> I guess everything hold together if one assumes that the `sizeof' in the
> definition of operator is the result of reduction of `sizeof' keyword token,
> and not the sequence of source code characters `s' `i' .. `f'.

No, I think this is stretching it too far, though it points at a
sensible way to correct the defect.

> Where in the standard does it specify this? In fact, where does the standard
> state which pp-token becomes which token?

With this exception, there is only one token that a given pp-token can
be converted to, though of course that token sometimes occurs in more
than one place within 6.1. This could all do with tidying up.

-- 
Clive D.W. Feather                       | If you lie to the compiler,
clive@demon.net        (work, preferred) | it will get its revenge.
clive@stdc.demon.co.uk (home)            |   - Henry Spencer