merging cccp and cpplib

[Michael Meissner]

On Mon, Dec 07, 1998 at 06:42:44PM -0800, Paul Eggert wrote:
>    Date: Mon, 07 Dec 1998 16:09:30 -0800
>    From: Per Bothner <bothner@cygnus.com>
> 
>    What we do care about is how standalone cpp plus cc1/cc1plus
>    compares with cc1/cc1plus with integrated cpplib.
> 
> Yes, that's correct.
> 
> It may be tricky to compare them in general, though.  For example,
> multiprocessor hosts might favor a separate cpp process since it's
> more parallelizable.

This is a myth, because cccp.c has always buffered the entire output internally
within memory (at least in the 9+ years I've been working on GCC since 1.2x or
1.3x), until it finally does the writes at the very end of processing.

> Perhaps we could modify cpplib so that it used threads (if available).
> That should make it dominate cpp even on multiprocessor hosts.
> It'd be more work though....

Please no.

-- 
Michael Meissner, Cygnus Solutions (Massachusetts office)
4th floor, 955 Massachusetts Avenue, Cambridge, MA 02139, USA
meissner@cygnus.com,	617-354-5416 (office),	617-354-7161 (fax)