Dead code (was: Re: Good numbers from Ritter's new string allocator)

[Paolo Carlini]

Alexandre Oliva wrote:

> On Nov 27, 2001, Paolo Carlini <pcarlini@unitus.it> wrote:
>
> > As regards __pagesize, I volunteer to take care of it in the next few
> > weeks, of course, but someone should introduce me to the configury
> > machinary needed.
>
> FWIW, the machinery is already there in libiberty.  In fact, you may
> even use libiberty's getpagesize.c in case you find
> AC_CHECK_FUNCS doesn't find getpagesize.
>
> > For instance, something that troubles me, is that the
> > pagesize is a TARGET parameter, so presently I have no idea how I could
> > test for it at build time.
>
> When building a target library, --host=target and
> --with-cross-host=host.

Thanks for your feedback.
Right now, as you may have read on the libstdc++ list, with Loren we are very
involved in the exponential shaping part of the work (as per Nathan Myers and
Benjamin's invitation) and honestly, given my inexperience of configure issues,
I cannot anticipate, how much time it will take me to make progress, even taking
advantage of your hints.

So: do you think it would be feasible for you to contribute to the project? I'm
following your postings and I know that for sure for you it is just a matter of
minutes. Even __pagesize alone would be a big improvement.

In case you are interested, I briefly recall you that the code containing the
"magic numbers" (mostly Loren's work) is already on the mainline, in
libstdc++v3/include/bits/basic_string.cc. Loren has provided detailed comments.
This one was the corresponding patch:

    http://gcc.gnu.org/ml/libstdc++/2001-11/msg00280.html

Thanks,
Paolo.