[x264-devel] Have configure script define libm variable if present

[Brad Smith]

On Mon, Sep 03, 2012 at 05:21:34PM -0400, Brad Smith wrote:
> On Sun, Aug 26, 2012 at 10:47:08AM -0400, Brad Smith wrote:
> > On Sat, Aug 25, 2012 at 06:23:35AM -0700, Jason Garrett-Glaser wrote:
> > > On Sat, Aug 25, 2012 at 6:14 AM, Brad Smith <brad at comstyle.com> wrote:
> > > > On Tue, Aug 21, 2012 at 12:09:43PM +0300, Georgi Chorbadzhiyski wrote:
> > > >> On 8/21/12 10:30 AM, Brad Smith wrote:
> > > >> >Have the configure script set libm if the OS has libm. Idea
> > > >> >copied from libpthread and just move setting LDFLAGS to below
> > > >> >the OS case check. I also did this to be able to use $libm in
> > > >> >another diff I'll send after this.
> > > >> >
> > > >> >
> > > >> >diff --git a/configure b/configure
> > > >> >index 086b399..5b18b1b 100755
> > > >> >--- a/configure
> > > >> >+++ b/configure
> > > >> >@@ -448,6 +448,7 @@ else
> > > >> >      fi
> > > >> >  fi
> > > >> >
> > > >> >+libm=""
> > > >> >  case $host_os in
> > > >> >      beos*)
> > > >> >          SYS="BEOS"
> > > >> >@@ -456,37 +457,37 @@ case $host_os in
> > > >> >      darwin*)
> > > >> >          SYS="MACOSX"
> > > >> >          CFLAGS="$CFLAGS -falign-loops=16"
> > > >> >-        LDFLAGS="$LDFLAGS -lm"
> > > >> >+        libm="-lm"
> > > >> >          if [ "$pic" = "no" ]; then
> > > >> >              cc_check "" -mdynamic-no-pic && CFLAGS="$CFLAGS -mdynamic-no-pic"
> > > >> >          fi
> > > >> >          ;;
> > > >> >      freebsd*)
> > > >> >          SYS="FREEBSD"
> > > >> >-        LDFLAGS="$LDFLAGS -lm"
> > > >> >+        libm="-lm"
> > > >> >          ;;
> > > >> >      kfreebsd*-gnu)
> > > >> >          SYS="FREEBSD"
> > > >> >          define HAVE_MALLOC_H
> > > >> >-        LDFLAGS="$LDFLAGS -lm"
> > > >> >+        libm="-lm"
> > > >> >          ;;
> > > >> >      netbsd*)
> > > >> >          SYS="NETBSD"
> > > >> >-        LDFLAGS="$LDFLAGS -lm"
> > > >> >+        libm="-lm"
> > > >> >          ;;
> > > >> >      openbsd*)
> > > >> >          SYS="OPENBSD"
> > > >> >-        LDFLAGS="$LDFLAGS -lm"
> > > >> >+        libm="-lm"
> > > >> >          ;;
> > > >> >      *linux*)
> > > >> >          SYS="LINUX"
> > > >> >          define HAVE_MALLOC_H
> > > >> >-        LDFLAGS="$LDFLAGS -lm"
> > > >> >+        libm="-lm"
> > > >> >          ;;
> > > >> >      gnu*)
> > > >> >          SYS="HURD"
> > > >> >          define HAVE_MALLOC_H
> > > >> >-        LDFLAGS="$LDFLAGS -lm"
> > > >> >+        libm="-lm"
> > > >> >          ;;
> > > >> >      cygwin*)
> > > >> >          EXE=".exe"
> > > >> >@@ -512,7 +513,7 @@ case $host_os in
> > > >> >      sunos*|solaris*)
> > > >> >          SYS="SunOS"
> > > >> >          define HAVE_MALLOC_H
> > > >> >-        LDFLAGS="$LDFLAGS -lm"
> > > >> >+        libm="-lm"
> > > >> >          if cc_check "" /usr/lib/64/values-xpg6.o; then
> > > >> >              LDFLAGS="$LDFLAGS /usr/lib/64/values-xpg6.o"
> > > >> >          else
> > > >> >@@ -525,6 +526,8 @@ case $host_os in
> > > >> >          ;;
> > > >> >  esac
> > > >> >
> > > >> >+LDFLAGS="$LDFLAGS $libm"
> > > >> >+
> > > >> >  case $host_cpu in
> > > >> >      i*86)
> > > >> >          ARCH="X86"
> > > >> >
> > > >>
> > > >> It would be nicer if libm is set by default to -lm and then only
> > > >> platforms that have named this library differently, would need
> > > >> to change it?
> > > >
> > > > Either way would be fine with me and I can understand the appeal
> > > > of your suggestion as you can just clear the libm variable for
> > > > the OS's without a libm. OS's without libm are the exception.
> > > >
> > > > Jason, any comments?
> > > 
> > > I don't have the experience/expertise to say anything meaningful about
> > > this kind of configure stuff, so I'll let other people decide.
> > 
> > OK can we just go with my initial patch since it follows as closely to
> > the original behavior of the code that it was replacing and just filling
> > in the libm variable as necessary? If anyone wants to change it later
> > that's fine.
> 
> ping.

Hello? Anyone?

-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.