You are not logged in.
Hi Steffen,
Well done! But it stay very strange for me because the official glibc is NOT build with -Os, we have that in the glib receipt:
case "$ARCH" in
i386|i486) echo "CFLAGS += -O2 -march=i486 -mtune=i486" > configparms ;;
*) echo "CFLAGS += -O2 -march=$ARCH -mtune=$ARCH" > configparms ;;
esac
And weel done for Cookutils, I didn' see mojo instruction but 3 commands and you can cook: tazdev gen-chroot && tazdev chroot && cook setup --wok
Hope the new toolchain for cooking will solve the bug: Linux 3.2.14, Binutils 2.22, Glibc 2.14.1, GCC 4.6.2
- Christophe
Offline
Hmm, I see you point.
I made a quick test:
Copied the working package to a working folder.
Changed the /etc/slitaz/cook.conf back to -Os
[c]
cook clean-wok
cook glibc-base
ls -la packages/glibc-* working/glibc-*
-rw-r--r-- 1 root root 5192 Apr 12 16:03 packages/glibc-2.13.tazpkg
-rw-r--r-- 1 root root 1200608 Apr 12 16:04 packages/glibc-base-2.13.tazpkg
-rw-r--r-- 1 root root 5188 Apr 12 08:43 working/glibc-2.13.tazpkg
-rw-r--r-- 1 root root 1201020 Apr 12 08:44 working/glibc-base-2.13.tazpkg
[/c]
The working package is a little bit larger.
A search for -Os in the cook output got zero result.
Copied the new package to the CF card, installed it and it still works on both the Geode and the Nehemia boards...
Now, from outside the chroot, I tried to fetch the one distributed with Slitaz 4:
[c]
tazpkg get glibc-base
glibc-base-2.13 already in the cache : /var/cache/tazpkg/4.0/packages
ls -la glibc-base-2.13.tazpkg
-rw-r--r-- 1 root root 1213532 Apr 11 11:38 glibc-base-2.13.tazpkg
[/c]
It is quite different so... what's up with Tank?
Btw. I'm having problems cleaning a single package from the wok:
[c]
cook glibc-base --clean
Setup aufs chroot...
Cleaning: glibc-base [ OK ]
Leaving aufs chroot...
/usr/bin/cook: .: line 77: can't open '/home/slitaz/wok/glibc-base/taz/glibc-base-*/receipt'
[/c]
That is why I clean the entire wok when rebuilding.
@steffen
yep - I confirm, I have a proper X only keyboard and mouse not working
exactly the same as it was, when I replaced glibc-base with glibc-base from slitaz-libre
Offline
I am quite happy - good news - I am writing now from my alix1C Geode CPU - rolling from yesterday
the bug is gone in Slitaz cooking
the cange from "glibc" to "eglibc" rocks see in hg: http://hg.slitaz.org/wok/rev/cb4299f34f9d
thanks pankso for the effort
Thomas
Offline
[ Generated in 0.016 seconds, 7 queries executed - Memory usage: 1.54 MiB (Peak: 1.77 MiB) ]