This is the mail archive of the
mailing list for the glibc project.
Re: building only libm from glibc source
- From: "Carlos O'Donell" <carlos at systemhalted dot org>
- To: "Martin Koeppe" <mkoeppe at gmx dot de>
- Cc: libc-help at sourceware dot org
- Date: Tue, 17 Jun 2008 08:41:50 -0400
- Subject: Re: building only libm from glibc source
- Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:received:received:message-id:date:from:sender :to:subject:cc:in-reply-to:mime-version:content-type :content-transfer-encoding:content-disposition:references :x-google-sender-auth; bh=v1+yFyqOFAm4CH1oa9XV32JTmrmxC/tThg+180gynFA=; b=gKLuwSf+N5G+K6wUpgm7IeUlBseAYzI321HSWlAGF1ZI+0BvxRlmFmmt5qsF73ghMW E21CXJmpgXP6ItbRyl21SbO3SaB2NGcvXzoClt5wzUJJ4YA88GNKY4UGYuDoH2CySlby BicgvAatRGQiAa/Y+DFVJ4s2inOFEsnw/C+5c=
- Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=message-id:date:from:sender:to:subject:cc:in-reply-to:mime-version :content-type:content-transfer-encoding:content-disposition :references:x-google-sender-auth; b=cUeK6eegz4UOa5IFYb3fYm0Yuv66D33mlvXtuB9ZXAcDOeabbNs6DZrdX0iZg3Bprd g8mATbQtHGz507rD8w2e2p1Guxq3yg60zWj8FkwCuBIZ8/3mFXQbWVLxvr7spkg5IAP/ xdyj+87vO7xLQHaC1mv18VrgNRfEhcaddVpYs=
- References: <Pine.LNX.firstname.lastname@example.org>
On Tue, Jun 17, 2008 at 4:48 AM, Martin Koeppe <email@example.com> wrote:
> On Interix the system libm is buggy, so I need a replacement. My idea is to
> only build the libm part out of the glibc sources, but unfortunately this
> doesn't seem that easy. There seem to be many dependencies to the libc part.
> I'm only interested in the libm.a and libm.so files, I currently don't need
> to build/install the glibc headers. Any help here in how to get that working
> would be appreciated.
I *think* you should be able to build just libm.so with this:
make -r PARALLELMFLAGS="" CVSOPTS="" -C ../../src/libc/math objdir=`pwd` all
It may need to build other required dependencies.
> To overcome the system libm bugs I currently use libm from newlib, but
> unfortunately this one has other bugs, too, so I now think it's necessary to
> try glibc's libm.