Mail Archives: cygwin/2000/10/11/11:34:53
On Wed, Oct 11, 2000 at 12:33:07PM +0100, Fifer, Eric wrote:
>>Chris Faylor wrote:
>>>On Tue, Oct 10, 2000 at 01:50:37PM -0400, Larry Hall (RFK Partners, Inc)
>wrote:
>>>At 01:08 PM 10/10/2000, Fifer, Eric wrote:
>>>>+ Restructuring of libm.a, libc.a and libcygwin.a:
>>>> Should be mentioned in the changes list?
>>>> Also, __infinity seems to be multiply-defined:
>>>> /usr/lib/libcygwin.a(ds00023.o)(.text+0x0): multiple definition
>of
>>>>`__infinity'
>>>> /usr/lib/libm.a(s_infconst.o)(.text+0x0): first defined here
>>>> Or I might be doing something wrong. I'm still struggling
>>>> to get Perl to compile.
>>>
>>>libm.a and libc.a are symlinks to libcygwin.a. Right now, gcc doesn't
>>>understand this so your best bet is to use only one of these in your
>>>link stream. If you don't, the resulting executable will crash.
>>
>>Also, this is unchanged since 1.1.4, AFAIK.
>
>Really? This is what I see in the latest snapshot.
>
>$ bunzip2 -c cygwin-inst-20001010.tar.bz2 | tar vtf - | egrep
>'lib(c|g|m|cygwin)\.a'
>-rw-r--r-- dj/games 246456 2000-10-11 05:35 usr/lib/libg.a
>-rw-r--r-- dj/games 215962 2000-10-11 05:36 usr/lib/libm.a
>-rw-r--r-- dj/games 0 2000-10-11 05:35 usr/lib/libc.a link to
>usr/lib/libg.a
>-rw-r--r-- dj/games 762638 2000-10-11 05:35 usr/lib/libcygwin.a
In that case, the snapshot is screwed up. libm.a should be linked
to libcygwin.a.
This is not an issue for the 1.1.5 release. I *always* symbolically
link libcygwin.a to libm.a. Here is the text from my release script:
ln -sf libcygwin.a libc.a
ln -sf libcygwin.a libm.a
ln -sf libcygwin.a libg.a
cgf
--
Want to unsubscribe from this list?
Send a message to cygwin-unsubscribe AT sourceware DOT cygnus DOT com
- Raw text -