From diy-linux-dev@diy-linux.org Mon Aug 18 23:36:17 2008 From: diy-linux-dev@diy-linux.org (ROGER BLAKE) Date: Tue, 19 Aug 2008 00:36:17 +0100 (BST) Subject: Minor fix for mktime check hanging/failing during bash configure with gcc 4.3 Message-ID: <626712.81943.qm@web86507.mail.ukl.yahoo.com> Hi All, Has anyone else noticed mktime hanging during the bash configure script ? I've had a look at this and it seems to be a known problem with configure scripts generated by autoconf earlier than 2.62 used together with a recent toolchain. As far as I can see only bash is affected in diy-linux - coreutils and findutils have the macro but I assume were packaged with the corrected code. *** Bash requires autoconf (2.62) to be run prior to configure to fix the bug. autoconf gets the new checking code from typically /var/share/autoconf/autoconf/functions.m4. *** [My limited understanding of the process is that aclocal should also be run prior to autoconf to generate an updated aclocal.m4 file, but that does not work with bash since the original aclocal.m4 contains hacked macros not present in standard autoconf - i.e. SHELL_CHECK_DECL] If my understanding needs correcting please let me know ! Fixing the bash configure hang should save several minutes of build time once the whole build moves to gcc 4.3. (my starting host is now OpenSuse 11 which already has gcc 4.3.1) Roger From diy-linux-dev@diy-linux.org Tue Aug 19 00:40:27 2008 From: diy-linux-dev@diy-linux.org (Greg Schafer) Date: Tue, 19 Aug 2008 10:40:27 +1000 Subject: Minor fix for mktime check hanging/failing during bash configure with gcc 4.3 In-Reply-To: <626712.81943.qm@web86507.mail.ukl.yahoo.com> References: <626712.81943.qm@web86507.mail.ukl.yahoo.com> Message-ID: <20080819004027.GA25177@eyo32.local> On Tue, Aug 19, 2008 at 12:36:17AM +0100, ROGER BLAKE wrote: > Has anyone else noticed mktime hanging during the bash configure script ? > I've had a look at this and it seems to be a known problem with configure > scripts generated Yes. The issue came up earlier: http://www.diy-linux.org/pipermail/diy-linux-dev/2008-February/001200.html But thanks for reporting anyway. I *will* get GCC 4.3 into the Refbuild soon.. I promise :-) Regards Greg From diy-linux-dev@diy-linux.org Wed Aug 20 12:25:36 2008 From: diy-linux-dev@diy-linux.org (James W. Morris) Date: Wed, 20 Aug 2008 13:25:36 +0100 Subject: touch problem Message-ID: Hi=2C I'm following the reference build instructions and have come unstuck at bas= h pass2 compilation. Shortly after configure (after compilation of eval.c): touch: settings times of `parser_built': No such process and make exits. I decided to upgrade to coreutils-2.6.12=2C could that be an issue? I am also unsure if at this stage I should be continuing to use /temptools/bash-pass1 or /temptools/bash ?? BTW=2C the build is on amd_64 (BIARCH=3DYES) within 64Studio (debian etch b= ased). BTW2=2C I have noticed other packages compiled earlier in the build have al= so shown the same message but have not caused make to exit. Regards + Cheers=2C James Morris. ~(sirromseventyfive)~ www.jwm-art.net This message has been scanned by radar and an old x-ray machine. No viruses= were found=2C but other ailments were detected. _________________________________________________________________ Win New York holidays with Kellogg=92s & Live Search=20 http://clk.atdmt.com/UKM/go/107571440/direct/01/= From diy-linux-dev@diy-linux.org Wed Aug 20 14:04:51 2008 From: diy-linux-dev@diy-linux.org (james jwm-art net) Date: Wed, 20 Aug 2008 14:04:51 +0000 Subject: touch problem Message-ID: Hi, I've made bash-3.2 compile without producing the touch error by uncommenting the appropriate 'experimental new rules' in the Makefile.in and commenting out the regular rules. I've absolutely no idea if there will be any consequences. Cheers, James. (Patch if you need, at bottom of msg) ---------------------- Hi=3D2C I'm following the reference build instructions and have come unstuck at bas=3D h pass2 compilation. Shortly after configure (after compilation of eval.c): touch: settings times of `parser_built': No such process and make exits. I decided to upgrade to coreutils-2.6.12=3D2C could that be an issue? I am also unsure if at this stage I should be continuing to use /temptools/bash-pass1 or /temptools/bash ?? BTW=3D2C the build is on amd_64 (BIARCH=3D3DYES) within 64Studio (debian etch b=3D ased). BTW2=3D2C I have noticed other packages compiled earlier in the build have al=3D so shown the same message but have not caused make to exit. Regards + Cheers=3D2C James Morris. ****************************************** --- bash-3.2/Makefile.in=092006-08-17 18:03:35.000000000 +0000 +++ bash-3.2-newrules/Makefile.in=092008-08-20 15:49:17.000000000 +0000 @@ -560,24 +560,24 @@ =09$(CC_FOR_BUILD) $(CCFLAGS_FOR_BUILD) -DBUILDTOOL -c -o $@ $(srcdir)/version.c # old rules -GRAM_H =3D parser-built -y.tab.o: y.tab.c ${GRAM_H} command.h ${BASHINCDIR}/stdc.h input.h -${GRAM_H}:=09y.tab.h -=09@-if test -f y.tab.h ; then \ -=09=09cmp -s $@ y.tab.h 2>/dev/null || cp -p y.tab.h $@; \ -=09fi -y.tab.c y.tab.h: parse.y +#GRAM_H =3D parser-built +#y.tab.o: y.tab.c ${GRAM_H} command.h ${BASHINCDIR}/stdc.h input.h +#${GRAM_H}:=09y.tab.h +#=09@-if test -f y.tab.h ; then \ +#=09=09cmp -s $@ y.tab.h 2>/dev/null || cp -p y.tab.h $@; \ +#=09fi +#y.tab.c y.tab.h: parse.y #=09-if test -f y.tab.h; then mv -f y.tab.h old-y.tab.h; fi -=09$(YACC) -d $(srcdir)/parse.y -=09touch parser-built +#=09$(YACC) -d $(srcdir)/parse.y +#=09touch parser-built #=09-if cmp -s old-y.tab.h y.tab.h; then mv old-y.tab.h y.tab.h; else cp -p y.tab.h ${GRAM_H}; fi # experimental new rules - work with GNU make but not BSD (or OSF) make -#y.tab.o: y.tab.c y.tab.h -#y.tab.c y.tab.h: parse.y command.h ${BASHINCDIR}/stdc.h input.h -#=09-if test -f y.tab.h; then mv -f y.tab.h old-y.tab.h; fi -#=09$(YACC) -d $(srcdir)/parse.y -#=09-if cmp -s old-y.tab.h y.tab.h; then mv old-y.tab.h y.tab.h; fi +y.tab.o: y.tab.c y.tab.h +y.tab.c y.tab.h: parse.y command.h ${BASHINCDIR}/stdc.h input.h +=09-if test -f y.tab.h; then mv -f y.tab.h old-y.tab.h; fi +=09$(YACC) -d $(srcdir)/parse.y +=09-if cmp -s old-y.tab.h y.tab.h; then mv old-y.tab.h y.tab.h; fi $(READLINE_LIBRARY): config.h $(READLINE_SOURCE) =09@echo making $@ in ${RL_LIBDIR} From diy-linux-dev@diy-linux.org Wed Aug 20 16:59:31 2008 From: diy-linux-dev@diy-linux.org (james jwm-art net) Date: Wed, 20 Aug 2008 16:59:31 +0000 Subject: touch problem SOLVED In-Reply-To: Message-ID: Ok Ok, I should've listened to you: latest and greatest is not always best ;-) >I decided to upgrade to coreutils-2.6.12=3D2C could that be an issue? From diy-linux-dev@diy-linux.org Thu Aug 21 02:19:50 2008 From: diy-linux-dev@diy-linux.org (James W. Morris) Date: Thu, 21 Aug 2008 03:19:50 +0100 Subject: newb @ EOF Message-ID: Hi list=2C Sorry to be a pain. I was really excited about this=2C thinking everything would be there... just follow the instructions - not thinking in other words. Anyway=2C I got to the end=2C using recommended pkg versions=2C ignoring two perl test failures: at ../ext/Sys/Syslog/t/syslog.t line 204 at io/fs.t line 269 # Failed at io/fs.t line 270 And now suddenly realise I don't actually have a clue about getting it all out from /temptools (--> /home/build/diy_linux ) and into the partition I setup for the LFS I quickly aborted (at gcc stage 1). Tried copying the files across to the partition and then changing $SYSROOT to point to that=2C but then $TEMPTOOLS/bin/env is not found. I expected to discover that merely copying the files would not work anyway=2C but can't even prove that to myself. A LFS hint http://www.linuxfromscratch.org/hints/downloads/files/lfs_next_to_existing_= systems.txt suggests mounting a loop device (sorry only scanned it so far) but I'm not convinced it's right for this. I don't understand the chroot command. RTFM right? Why do the files in /etc magically appear when I enter the chroot env but not from outside=2C even as root proper? That's the problem with copying... Can't quite remember=2C but something put me off from mounting the partition within the chroot environment and copying everything to it. Gotta sleep=2C Cheers=2C james ~(sirromseventyfive)~ www.jwm-art.net This message has been scanned by radar and an old x-ray machine. No viruses= were found=2C but other ailments were detected. amd64=2C 64studio (debian etch + bits and bobs)=2C stuff. _________________________________________________________________ Win New York holidays with Kellogg=92s & Live Search=20 http://clk.atdmt.com/UKM/go/107571440/direct/01/= From diy-linux-dev@diy-linux.org Thu Aug 21 05:53:01 2008 From: diy-linux-dev@diy-linux.org (Greg Schafer) Date: Thu, 21 Aug 2008 15:53:01 +1000 Subject: touch problem In-Reply-To: References: Message-ID: <20080821055301.GA22478@eyo32.local> On Wed, Aug 20, 2008 at 01:25:36PM +0100, James W. Morris wrote: > I'm following the reference build instructions and have come unstuck at > bash pass2 compilation. Shortly after configure (after compilation of > eval.c): > > touch: settings times of `parser_built': No such process > > and make exits. > > I decided to upgrade to coreutils-2.6.12, could that be an issue? I think so, but I'm still behind on testing latest package updates. I recall seeing a thread on the Coreutils list which I presume explains what you are seeing. Thread starts here: http://lists.gnu.org/archive/html/bug-coreutils/2008-06/msg00022.html Regards Greg From diy-linux-dev@diy-linux.org Thu Aug 21 06:21:10 2008 From: diy-linux-dev@diy-linux.org (Greg Schafer) Date: Thu, 21 Aug 2008 16:21:10 +1000 Subject: newb @ EOF In-Reply-To: References: Message-ID: <20080821062110.GA22653@eyo32.local> On Thu, Aug 21, 2008 at 03:19:50AM +0100, James W. Morris wrote: > And now suddenly realise I don't actually have a clue about getting > it all out from /temptools (--> /home/build/diy_linux ) and into the > partition I setup for the LFS I quickly aborted (at gcc stage 1). > > Tried copying the files across to the partition and then changing > $SYSROOT to point to that, but then $TEMPTOOLS/bin/env is > not found. I expected to discover that merely copying the files > would not work anyway, but can't even prove that to myself. Umm, I think you're confused :-/ Once it's built, it's built. Forget about the env vars. From outside the chroot, just copy everything (minus the temptools) to the target partition, install a kernel, setup some bootscripts and other misc config files (/etc/{fstab,resolv.conf} etc) then configure the bootlaoder and you're away. Well, that was over-simplified but you get the picture.. If needed, you can chroot into the target partition eg: (as root on the host) chroot /path/to/target/area Regards Greg From diy-linux-dev@diy-linux.org Thu Aug 21 22:06:22 2008 From: diy-linux-dev@diy-linux.org (james jwm-art net) Date: Thu, 21 Aug 2008 22:06:22 +0000 Subject: newb @ EOF In-Reply-To: <20080821062110.GA22653@eyo32.local> Message-ID: On 1/1/1970, "Greg Schafer" wrote: >Umm, I think you're confused :-/ Once it's built, it's built. Forget about >the env vars. From outside the chroot, just copy everything (minus the >temptools) to the target partition, install a kernel, setup some bootscripts >and other misc config files (/etc/{fstab,resolv.conf} etc) then configure >the bootlaoder and you're away. Well, that was over-simplified but you get >the picture.. If needed, you can chroot into the target partition eg: The confusion was worse than you thought: `chroot /temptools' ... Now I better go look into the LFS init scripts (for want of a better place to look) now that I can (very nearly) boot the system. Thanks, James.