| View previous topic :: View next topic |
| Author |
Message |
discostu Guru


Joined: 01 Nov 2002 Posts: 333
|
Posted: Sun Dec 08, 2002 12:59 am Post subject: bootstrap.sh texinfo error |
|
|
I currently have Gentoo 1.1a. I'm trying to install Gentoo 1.4 from the livecd and when I run bootstrap.sh (progressing from stage 1 to stage 2), I get an error.
| Code: | checking for C compiler default output... configure: error: C compiler cannot create executables
!!! ERROR: sys-apps/texinfo-4.2-r5 failed.
!!! Function src_compile, Line 11, Exitcode 77
!!! (no error messages) |
Any Ideas?
In my make.conf, the only thing I changed was to uncomment/edit it to say | Code: | | CFLAGS="-mcpu=athlon -03 -pipe" |
thanks, stett |
|
| Back to top |
|
 |
discostu Guru


Joined: 01 Nov 2002 Posts: 333
|
Posted: Sun Dec 08, 2002 7:33 am Post subject: |
|
|
I got the bootstrap.sh to work by commenting out the CFLAGS line in "make.conf"
Now I'm doing a "emerge system" and I'm getting an error:
| Code: | /usr/sbin/ebuild.sh: sed: command not found
/usr/sbin/ebuild.sh: cd /var/tmp/portage/sh-utils-2.0.15/work/sh-utils-2.0.15: No such file or directory
can't find file to patch at input line 3
Perhaps you used the wrong -p or --strip option?
The text leading to this was:
------------------------------------
|--- sh-utils-2.0.15/src/uname.c Thu Jul 18
|+++ sh-utils-2.0.15-carlos/src/uname.c Wed Sep 4
File to patch:
Skip this patch? [y]
Skipping patch.
5 out of 5 hunks ignored
!!! ERROR: sys-apps/sh-utls-2.0.15 failed |
I have been copying /bin/sed to the bin where I did the chroot (/mnt/gentoo/), but it is removed everytime I do "emerge system" |
|
| Back to top |
|
 |
rac Bodhisattva


Joined: 30 May 2002 Posts: 6553 Location: Japanifornia
|
Posted: Sun Dec 08, 2002 9:27 am Post subject: |
|
|
Did you do a fresh install with clean partitions, or are you trying to somehow overlay and upgrade an existing 1.1 installation? _________________ For every higher wall, there is a taller ladder |
|
| Back to top |
|
 |
discostu Guru


Joined: 01 Nov 2002 Posts: 333
|
Posted: Sun Dec 08, 2002 10:32 pm Post subject: |
|
|
| rac wrote: | | Did you do a fresh install with clean partitions, or are you trying to somehow overlay and upgrade an existing 1.1 installation? |
That was my problem. I formatted my /boot and root partitions, and started the install from scratch everything worked. Thanks! |
|
| Back to top |
|
 |
Koon Retired Dev


Joined: 10 Dec 2002 Posts: 518
|
Posted: Wed Jan 22, 2003 4:09 pm Post subject: |
|
|
Same problem here with 1.4rc2 :
Error while compiling texinfo during bootstrap using :
-march=athlon-xp -O3 -pipe
So the problem (bootstrap fails using default compile options) is still there.
I solved it by starting from stage2. |
|
| Back to top |
|
 |
Koon Retired Dev


Joined: 10 Dec 2002 Posts: 518
|
Posted: Thu Jan 23, 2003 9:07 am Post subject: |
|
|
Bugzilla search found #10327 :
-march=athlon-xp breaks texinfo on a fresh install
This is a texinfo / gcc compatibility bug that prevents people using -march=athlon-xp from bootstrapping successfully.
Solution (worked for me [TM]):
- Starting from stage2, emerge system
- When the emerge texinfo fails, change compile options to for example -mcpu=i686 (which is what people do when they comment the CFLAGS out).
- emerge texinfo
- restore your original make.conf options
- emerge system again
-K |
|
| Back to top |
|
 |
piranha n00b

Joined: 10 Jun 2002 Posts: 40
|
Posted: Wed Feb 05, 2003 8:11 am Post subject: |
|
|
| Koon wrote: | Same problem here with 1.4rc2 :
Error while compiling texinfo during bootstrap using :
-march=athlon-xp -O3 -pipe
So the problem (bootstrap fails using default compile options) is still there.
I solved it by starting from stage2. |
Same problem here. |
|
| Back to top |
|
 |
bludger Guru

Joined: 09 Apr 2003 Posts: 389
|
Posted: Tue Apr 22, 2003 3:57 am Post subject: |
|
|
| I have the same problem with an athlon-tbird. Is this bug entered in the bug tracking database? Does anyone have a workaround for starting at stage 1? |
|
| Back to top |
|
 |
bludger Guru

Joined: 09 Apr 2003 Posts: 389
|
Posted: Tue Apr 22, 2003 4:11 am Post subject: |
|
|
| bludger wrote: | | I have the same problem with an athlon-tbird. Is this bug entered in the bug tracking database? Does anyone have a workaround for starting at stage 1? |
More completely, I have the following flags -march=athlon-tbird -O3 -pipe -fomit-frame-pointer -mmmx -msse -m3dnow.
When I replaced athlon-tbird with i686, the problem did not go away. When I also removed mmx, sse and 3dnow it compiled ok. |
|
| Back to top |
|
 |
Koon Retired Dev


Joined: 10 Dec 2002 Posts: 518
|
|
| Back to top |
|
 |
bludger Guru

Joined: 09 Apr 2003 Posts: 389
|
Posted: Tue Apr 22, 2003 11:13 am Post subject: |
|
|
Thanks for the pointer. There was a request to enter a bug in gcc.gnu.org, but there was no cross reference to that entry (if it was ever made). The bug was then set to "resolved". I find it strange that a bug can be set to "resolved" when the problem still exists and there is no indication that it has been duplicated somewhere else successfully. For a distribution such as gentoo, where most components are handled externally, it seems to me (as a humble noob) that rather than just closing such bugs, another category (eg. "external") with a cross-reference to the external bug would make more sense.
Anyway, I am still looking for a workaround for a stage1 installation. Currently I think that I will have to complete the stage1 with march=i686 and then do an emerge -u world with march=athlon-tbird (and texinfo excluded. Is there a way of specifying that a particular application should have a particular arch flag during bootstrap? |
|
| Back to top |
|
 |
bludger Guru

Joined: 09 Apr 2003 Posts: 389
|
Posted: Wed Apr 23, 2003 8:42 am Post subject: |
|
|
Check this link:
https://forums.gentoo.org/viewtopic.php?t=42459&highlight=illegal+instruction
It seems that this is a kernel bug in that the kernel is not turning on the sse flag for the newer athlons. Check your /proc/cpuinfo when running 2.4.20 versus 2.4.19. It should be fixed in 2.4.21.
My problem is that I have an older athlon (model 4), which doesn't seem to support this flag), but I stupidly turned on the sse flag anyway. This means that such actions as building texinfo and running top try to use this flag, but cannot. In my case, I need to re-emerge without sse. For those with an athlon-xp, they need to get the kernel patch, recompile everything without sse, or wait for the newer kernel.
I think my comments about closing bugs prematurely still apply, however. |
|
| Back to top |
|
 |
Koon Retired Dev


Joined: 10 Dec 2002 Posts: 518
|
Posted: Wed Apr 23, 2003 10:35 am Post subject: |
|
|
Nice ! I didn't find the solution on my own research...
I will test the kernel patch ASAP.
-K |
|
| Back to top |
|
 |
dilandau Guru


Joined: 24 May 2003 Posts: 485 Location: germany
|
Posted: Mon Jan 05, 2004 11:23 pm Post subject: |
|
|
-mno-sse did solve it for me _________________ gentoo linux - amd duron stalebred 1600 - elsa gladiac 311 (nvidia) - elitegroup k7s5a (sis, lan) - 256mb ram - wintv pci fm - airstar2 dvb-t pci - ide cdwriter - hp psc 1110 |
|
| Back to top |
|
 |
CoolAJ86 n00b


Joined: 24 May 2004 Posts: 61 Location: Shelburne, VT
|
Posted: Sat Apr 02, 2005 3:06 pm Post subject: Same Problem, but different |
|
|
I too had this error.
I learned two things:
double check your spelling ("march=athron-xp" isn't an actual flag )
try one additional flag at a time until it breaks -- then you've found the problem!
| Code: |
# This failed
#CFLAGS="-Os -march=athlon-xp -pipe -msse -mfpmath=sse -falign-functions=64 -fforce-addr -D_FILE_OFFSET_BITS=64 -visibility-inlines-hidden"
# This worked
CFLAGS="-Os -march=athlon-xp -pipe -msse -mfpmath=sse -falign-functions=64 -fforce-addr -D_FILE_OFFSET_BITS=64"
|
I see a lot of people using "big optimizations" and I was under the impression that this was a "good thing". However, I did some reading on a few sites, and if you're on I desktop system I don't see how the faster run time really outways the 30% larger binary size for browsing and checking e-mail. So I've said to florida with the optimizations - I just want it small because I've got CPU to burn, while memory is still a commodity. _________________ "May the source be with you."
Laterz-
~CoolAJ86
http://CoolAJ86.Havenite.net
http://www.uvm.org/vague - LUG VT |
|
| Back to top |
|
 |
|