This is the mail archive of the
cygwin-developers
mailing list for the Cygwin project.
Re: [RFC] Cygwin libstdc++ plan (operator new/delete replacement)
- From: Charles Wilson <cygwin at cwilson dot fastmail dot fm>
- To: cygwin-developers at cygwin dot com
- Date: Fri, 02 Oct 2009 09:43:26 -0400
- Subject: Re: [RFC] Cygwin libstdc++ plan (operator new/delete replacement)
Corinna wrote:
> On Oct 2 08:36, Andy Koppe wrote:
>> 2009/10/2 Dave Korn
>> > Right now however head isn't working for me at all, for what I think must
>> > be unrelated reasons, given that the snapshot works but clean head doesn't.
>>
>> Same here. Bash just exits immediately, with or without the remaining
>> locale patches.
>>
>> Do I need to do more than just copy cygwin0.dll to /bin/cygwin1.dll? I
>> did update gcc since that last worked.
>
> Personally I'm using a Linux cross gcc "i686-pc-cygwin-gcc (GCC) 4.3.2
> 20080827 (alpha-testing) 1" which is based on Dave's first gcc-4 release
> plus a small patch from cgf necessary for the cross build. I have no
> problems at all running my self-built Cygwin DLLs. FWIW, I'm usually
> not building with optimization for easier debugging.
I'm having problems with a self-built cygwin1.dll too. Built from head
last night with -ggdb -O0, using Dave's latest gcc-4.3.4-1. Same
symptoms as Andy.
(I'm actually kinda glad you guys are seeing the problem, because my
source has one change: I replaced
winsup/cygwin/lib/pseudo-reloc.c
with a copy of
winsup/mingw/pseudo-reloc.c
which supports v2 relocs as well as v1. I wanted to see if it would
"just work". If *I* was having the problem and you guys weren't, I'd
have blamed pseudo-reloc but now it looks like the problem is elsewhere.
pseudo-reloc might STILL be borked for cygwin, but...)
--
Chuck