Re: (patch) Workaround for gcc's float complex bug.

[Mumit Khan <khan at NanoTech dot Wisc dot EDU>]

On Mon, 22 Nov 1999, Chris Faylor wrote:

> On Mon, Nov 22, 1999 at 04:52:02PM -0800, Benjamin Kosnik wrote:
> >Mumit, all your changes should be in now, including the acinclude.m4
> >bit I missed earlier. I guess the only things left are:
> >
> >1) WEOF
> >
> >2) getting the cygwin DLL to run mknumeric_limits: Chris, I've cc'd you
> >in the hopes that you will work with Mumit to get Vadim's patch in--I
> >talked to you about this briefly a couple of weeks ago when you were
> >in Sunnyvale.
> 
> Are you sure that you talked to me?  I don't recall this.
> 
> I have vast reservations about complicating Cygwin's signal handling
> structure with Vadim's changes.

I know, and that's why I pointed out that it wasn't sanctioned by you.
However, I need the patch to build C++ std limits file, which 
unfortunately stresses signal handling. I'm going to have prebuilt
versions of the generated files in question so others can build Cygwin
libstdc++-v3 in the meantime. 

> I think I'm coming in in the middle here.  I don't know what macros
> you're talking about or why you hate macros.

You mean to tell me that you don't every posting in every mailing
list ;-) Yeah, you're coming in the middle, sorry. See the archive
(http://sourceware.cygnus.com/ml/libstdc++/), and search for my name
in the last 3 days or so. It's only a handful of messages.

The wchar_t specializations require WEOF, which doesn't exist for Cygwin
yet, and I had to add a WEOF macro, which is a gross hack. I'd rather
get this done via autoconf, but there are a few other issues to consider
(eg., what is the appropriate value?). I'm hoping for some free time to 
spend some time looking at the wide character support for Cygwin, and 
hopefully we can get rid of it then. Chris, do you have any plans in this 
area?

Benjamin, would it make more sense to autoconf the wchar_t specializations
and only do it when the target supports well enough for libstdc++-v3?

Regards,
Mumit