Re: Some experiments with cpplib's output end

[Neil Booth <neil at daikokuya dot demon dot co dot uk>]

Zack Weinberg wrote:-

> We've squelched fputs_unlocked, but now cpp_output_token has to do
> more work, which means we don't win as much as we might like.
> Nonetheless this is definitely better than what we had.
> 
> It's possible to squeeze out another four seconds by using
> fwrite_unlocked instead of fwrite, but this involves more autoconf
> gook.  I am particularly unfond of it in this case, since the default
> chosen by glibc is so completely and obviously wrong.

Thanks for the analysis, Zack.  I wonder if we're better something
simple and braindead, like our own scratch buffer of size roughly 8K
say, that we manage ourselves and only dump (probably bypassing stdio)
when it's about to overflow?

I might look at this in the future.  Meanwhile I'm still looking at
token handling; we can squeeze more out of that.

Neil.