Prev: Printing problem
Next: 64 bit question
From: Juan Mª Santiago Burruchaga on 4 Mar 2010 06:22 The packcage gnuplot-4.2.6-x86_64-1 is in the last ChangeLof for Slackware-current 32 bits.
From: Mikhail Zotov on 4 Mar 2010 13:32 On Thu, 04 Mar 2010 12:22:44 +0100 Juan M� Santiago Burruchaga <juanmsantiago(a)telefonica.net> wrote: > The packcage gnuplot-4.2.6-x86_64-1 is in the last ChangeLof for > Slackware-current 32 bits. > There's a typo in the build script. -- Mikhail
From: Sylvain Robitaille on 5 Mar 2010 00:49 Mikhail Zotov wrote: > There's a typo in the build script. This seems like a good time to suggest again that the target architecture is likely best auto-detected at build time, at least in the default cases, in Slackware's build scripts. Perhaps someone with some influence on those scripts (Eric? Robby?) could propose it at a higher level? In case it helps, in my own build scripts I use the following to determine the architecture I'm building on: ARCH=${ARCH:-`uname -m`} Nice, simple, and portable. It's followed shortly after with: case "$ARCH" in alpha) SLKCFLAGS="-O2 -m${MOPT}=ev4" LIBDIRSUFFIX="64" SLKLDFLAGS="-L/usr/lib${LIBDIRSUFFIX}"; ;; i?86) SLKCFLAGS="-O2 -march=${ARCH} -m${MOPT}=${ARCH}" SLKLDFLAGS=""; LIBDIRSUFFIX="" ;; s390) SLKCFLAGS="-O2" SLKLDFLAGS=""; LIBDIRSUFFIX="" ;; powerpc) SLKCFLAGS="-O2" SLKLDFLAGS=""; LIBDIRSUFFIX="" ;; x86_64) SLKCFLAGS="-O2 -fPIC" LIBDIRSUFFIX="64" SLKLDFLAGS="-L/usr/lib${LIBDIRSUFFIX}"; ;; athlon-xp) SLKCFLAGS="-march=athlon-xp -O3 -pipe -fomit-frame-pointer" SLKLDFLAGS=""; LIBDIRSUFFIX="" ;; esac Which should look more or less familiar to those who have worked with these scripts from multiple sources (I didn't make the above up, I simply expanded it from other sources). This makes it *very* easy to create Slackware packages for various architectures without risking typos ... -- ---------------------------------------------------------------------- Sylvain Robitaille syl(a)encs.concordia.ca Systems analyst / AITS Concordia University Faculty of Engineering and Computer Science Montreal, Quebec, Canada ----------------------------------------------------------------------
From: Mikhail Zotov on 5 Mar 2010 03:52 On Fri, 5 Mar 2010 05:49:31 +0000 (UTC) Sylvain Robitaille <syl(a)alcor.concordia.ca> wrote: > Mikhail Zotov wrote: > > > There's a typo in the build script. > > This seems like a good time to suggest again that the target > architecture is likely best auto-detected at build time, at least in > the default cases, in Slackware's build scripts. Perhaps someone > with some influence on those scripts (Eric? Robby?) could propose it > at a higher level? In case it helps, in my own build scripts I use > the following to determine the architecture I'm building on: > > ARCH=${ARCH:-`uname -m`} This is already implemented in more than 200 of the build scripts. Check, for example, bash.SlackBuild: # Automatically determine the architecture we're building on: if [ -z "$ARCH" ]; then case "$( uname -m )" in i?86) export ARCH=i486 ;; arm*) export ARCH=arm ;; # Unless $ARCH is already set, use uname -m for all other archs: *) export ARCH=$( uname -m ) ;; esac fi I guess guys just don't have time to fix them all at once. -- Mikhail
From: Loki Harfagr on 5 Mar 2010 05:50
Fri, 05 Mar 2010 11:52:24 +0300, Mikhail Zotov did cat : > On Fri, 5 Mar 2010 05:49:31 +0000 (UTC) Sylvain Robitaille > <syl(a)alcor.concordia.ca> wrote: > >> Mikhail Zotov wrote: >> >> > There's a typo in the build script. >> >> This seems like a good time to suggest again that the target >> architecture is likely best auto-detected at build time, at least in >> the default cases, in Slackware's build scripts. Perhaps someone with >> some influence on those scripts (Eric? Robby?) could propose it at a >> higher level? In case it helps, in my own build scripts I use the >> following to determine the architecture I'm building on: >> >> ARCH=${ARCH:-`uname -m`} > > This is already implemented in more than 200 of the build scripts. > Check, for example, bash.SlackBuild: > > # Automatically determine the architecture we're building on: if [ -z > "$ARCH" ]; then > case "$( uname -m )" in > i?86) export ARCH=i486 ;; > arm*) export ARCH=arm ;; > # Unless $ARCH is already set, use uname -m for all other archs: > *) export ARCH=$( uname -m ) ;; > esac > fi > > I guess guys just don't have time to fix them all at once. That's quite what I thought ;-) Now, I'm wondering if this would be risky or not: ----- x86_64) SLKCFLAGS="-O2 -fPIC" uname -p|grep -qw 'i[57]' && SLKCFLAGS="-O2 -fPIC -march=core2 -mtune=core2 " LIBDIRSUFFIX="64" SLKLDFLAGS="-L/usr/lib${LIBDIRSUFFIX}"; ----- at least for 'jack' related stuff I had the feeling it gave some more room, now I'm not a HW guy the least then better check it with someone who knows where he talks from :-) |