Gentoo Forums
Gentoo Forums
Gentoo Forums
Quick Search: in
GCC 3.2 is here: Now is the time to rebuild everything. How?
View unanswered posts
View posts from last 24 hours

Goto page Previous  1, 2, 3, 4, 5, 6, 7, 8  Next  
Reply to topic    Gentoo Forums Forum Index Portage & Programming
View previous topic :: View next topic  
Author Message
Aspholyte
n00b
n00b


Joined: 16 Aug 2002
Posts: 6

PostPosted: Sun Aug 18, 2002 10:18 am    Post subject: Reply with quote

I got a problem. I did rmerge -f tonight, everything worked fine until this:

>>> original instance of package unmerged safely.
>>> Regenerating /etc/ld.so.cache...
>>> sys-apps/sed-3.02.80-r3 merged.

* Regenerating GNU info directory index...
* Processed 85 info files.
* IMPORTANT: 29 config files in /etc need updating.
* Type emerge --help config to learn how to update config files.

[[1;31mFile '' does not exist![[0m
[[1;31mMerge failed. Using backup CFLAGS[[0m
[[1;31mFile '' does not exist![[0m
[[1;31mFile '' does not exist![[0m
[[1;31mMerge failed. Using backup CFLAGS[[0m
[[1;31mFile '' does not exist![[0m
[[1;31mFile '' does not exist![[0m
[[1;31mMerge failed. Using backup CFLAGS[[0m
...
and so on
...
[[1;31mFile '' does not exist![[0m
[[1;31m
Failed to emerge package sys-apps/modutils-2.4.16-r3!
Failed to emerge package sys-devel/make-3.79.1-r4!
Failed to emerge package dev-libs/libpcre-3.9-r1!
...
then there are a whole lot of packages it didn't emerge
...
Failed to emerge package app-editors/bluefish-0.7-r2!
Failed to emerge package net-irc/xchat-1.8.9-r3!
[[0m


this is what tee wrote to my file. I'm kinda confused now. I did an emerge rsync right before to be sure to be up to date

This are my make.conf settings:

CHOST="i686-pc-linux-gnu"
CFLAGS="-march=athlon-xp -O3 -pipe -fomit-frame-pointer -fforce-addr"
CXXFLAGS="-march=athlon-xp -O3 -pipe -fomit-frame-pointer -fforce-addr"
BCFLAGS="-march=athlon-xp -O3 -pipe"
BCXXFLAGS="-march=athlon-xp -O3 -pipe"

any idea what i did wrong?
Back to top
View user's profile Send private message
_puck_
Tux's lil' helper
Tux's lil' helper


Joined: 12 May 2002
Posts: 111
Location: Kaiserslautern / Germany

PostPosted: Sun Aug 18, 2002 1:40 pm    Post subject: Reply with quote

Hello there. I just had a great idea (as always it turned out to have some minor disatvantages) ... I use a second partition to try newest gentoo things and do not boot from a boot cd, instead I am using my working system ... After chroot nano says:
Code:

coruscant profiles # nano
Error opening terminal: rxvt.

Any clues?

But thats not why I am posting here. I read a whole lot of things about compiler options using with gcc3.2. I have a Athlon 1400 (C) and want to know, which ones i can use, which ones i better avoid and which ones are completly useless. ;) ... I searched for that before, but didn't find anything. Thank you for helping.
_________________
regards
puck
___________________________
#There is more than one way to do it.
Back to top
View user's profile Send private message
Tuxisuau
Apprentice
Apprentice


Joined: 04 May 2002
Posts: 213
Location: Catalunya (Europe)

PostPosted: Sun Aug 18, 2002 6:45 pm    Post subject: Reply with quote

Before doing anything, check:

http://lists.gentoo.org/pipermail/gentoo-dev/2002-August/014324.html

Or you expose yourself to have not nice breakages as I had...
_________________
IM me at tuxisuau@jabber.7a69ezine.org
Back to top
View user's profile Send private message
Dolio
l33t
l33t


Joined: 17 Jun 2002
Posts: 650

PostPosted: Sun Aug 18, 2002 7:57 pm    Post subject: Reply with quote

Aspholyte: Hrm...

I don't think you did anything wrong.

Those error messages come from the code where rmerge calculates the absolute path of the ebuild file...

Maybe after sed got re-emerged, something in the sed package got borked... I'll look into it.

Edit: I'm, so far, unable to reproduce this problem. I tried doing an 'rmerge -f sed man' which remerges sed and then 3 other packages related to man, and it worked without problems... The only other thing I could try is doing a plain 'rmerge -f', but it would be at least 3 hours before I'd know if the bug was reproduced...

I've also tried downgrading sed, and then rmerging up to the newer version. That worked fine.

Has anyone else encountered this bug? I suppose its possible that its not a problem with sed at all (although then it'd have to be a problem with an argument passed to a function in the script...). I'm kind of stumped.

Aspholyte: If you have the courage to try again, you could try running rmerge with the -vv option (two verboses). You'll get lots of debug information, however, before each package time you get a "File '' does not exist" error, you should get a "Merging: <PACKAGE>" message. If that <PACKAGE> is just blank, then its a problem with the script I have to look at. If not, then something went screwy with sed while you were rmerging.

Good luck.
_________________
They don't have a good bathroom to do coke in.
Back to top
View user's profile Send private message
_puck_
Tux's lil' helper
Tux's lil' helper


Joined: 12 May 2002
Posts: 111
Location: Kaiserslautern / Germany

PostPosted: Sun Aug 18, 2002 8:28 pm    Post subject: Reply with quote

Ok, i read a bit in the gcc.gnu.org online documentation and in all threats including "the mother of all CFLAGS threads". Here is what I would like to use an a Athlon (C) 1400 ... does anyone know, if some of teh options work or do not work?
Code:

-march=athlon-4 -O3 -loop -pipe -fforce-addr -malign-double -fomit-frame-pointer -ffast-math -fno-exceptions -fexpensive-optimizations -falign-functions -falign-loops -frerun-loop-opt -frerun-cse-after-loop -floop-optimize -funroll-loops -mfpmath=sse -msse2 -mno-see2 -3dnow -mno-3dnow -mmmx -mno-mmx -msse -mno-sse

_________________
regards
puck
___________________________
#There is more than one way to do it.
Back to top
View user's profile Send private message
Dolio
l33t
l33t


Joined: 17 Jun 2002
Posts: 650

PostPosted: Sun Aug 18, 2002 9:19 pm    Post subject: Reply with quote

_puck_: Why do you have -msse and then -mno-sse? In fact, you do that with a lot of flags. That's just turning stuff on and then back off. You might as well just do -mno-sse.

Back when I was compiling with gcc3.1 (I have an Athlon Thuderbird 1400 MHz), xfree failed on one directory when I used -fforce-addr, so you might want to turn that off for X. X also needs -mno-3dnow -mno-mmx for at least one directory (as of 3.2_pre) with an Athlon-tbird, so you might want to try that if you're getting errors (You'll probably want -mmmx and -m3dnow on most other stuff, though, because, well, why not?).

I'm not sure this all applies to you, though (You have a palomino, right?).
_________________
They don't have a good bathroom to do coke in.
Back to top
View user's profile Send private message
phong
Bodhisattva
Bodhisattva


Joined: 16 Jul 2002
Posts: 778
Location: Michigan - 15 & Ryan

PostPosted: Mon Aug 19, 2002 2:42 am    Post subject: Reply with quote

_puck_:

I agree with the things Dolio mentioned, and in addtion, -malign-double breaks Python, Perl and probably others as well, and is IMNSHO a Very Bad Idea(tm). I'm pretty sure I can blame some seg faults I was having with gcc on -malign-double, though I havn't proven it. -ffast-math is OK sometimes, not OK in general. IIRC -fexpensive-optimizations should be turned on already by -O3, as should -frerun-loop-opt and -frerun-cse-after-loop. -floop-optimize is implied by -frerun-loop-opt (which basically means do loop-optimize twice). Also, I don't recall seeing -fno-exceptions in the list of optimizations in the gcc manual, so I don't know what it does or if its a good idea.
_________________
"An empty head is not really empty; it is stuffed with rubbish. Hence the difficulty of forcing anything into an empty head."
-- Eric Hoffer
Back to top
View user's profile Send private message
Aspholyte
n00b
n00b


Joined: 16 Aug 2002
Posts: 6

PostPosted: Mon Aug 19, 2002 10:48 am    Post subject: Reply with quote

I wasn't able to reproduce that bug, either.
Well, I'm recompiling everything the 5th time now, cause everytime something else just hangs (it doesn't fail to merge a package, it just hangs somewhere and doesn't continue :/ ). Strangely, rmerge -p won't resume the previous rmerge, either.
Well, one good thing is that my X compiled with -march=athlon-xp -03 -pipe -fforce-addr -fomit-frame-pointer ;)

I'll just hope it works this time.

btw, Thx for your script, Dolio! :)
Back to top
View user's profile Send private message
trythil
Tux's lil' helper
Tux's lil' helper


Joined: 06 Jun 2002
Posts: 123
Location: RHIT, Terre Haute, IN, USA

PostPosted: Mon Aug 19, 2002 4:47 pm    Post subject: Reply with quote

phong wrote:
also, I don't recall seeing -fno-exceptions in the list of optimizations in the gcc manual, so I don't know what it does or if its a good idea.


-fno-exceptions turns off support for C++ exception handling, which can break code that uses it.
Back to top
View user's profile Send private message
_puck_
Tux's lil' helper
Tux's lil' helper


Joined: 12 May 2002
Posts: 111
Location: Kaiserslautern / Germany

PostPosted: Mon Aug 19, 2002 4:57 pm    Post subject: Reply with quote

phong and Dolio: Thank you very much for the comments, I will use it with my next rebuilt, starting tonight ;) ... I have a 1.4 cvs snapshot2 0020818-2037 and it did work already with no options and with the standard options in the make.conf for athlon only to bootstrap and emerge system.

The other options did not even compile the first gcc-3.2 properly. If I have more results, I will let you know.
_________________
regards
puck
___________________________
#There is more than one way to do it.
Back to top
View user's profile Send private message
Dolio
l33t
l33t


Joined: 17 Jun 2002
Posts: 650

PostPosted: Mon Aug 19, 2002 6:47 pm    Post subject: Reply with quote

Aspholyte: Wow. You're having weird problems. :)

Do you get any kind of errors when -p fails to resume the previous merge? The script should save the previous merge data before doing any actual emerging, so I'm not sure why -p would be broken by one of the emerges hanging (note, if you have to Ctrl-C out of rmerge, you can try using the --fix option. It should recover the proper data and resume from where you were at the time when you had to kill it. Not sure if you knew this).

Thanks for helping me get the kinks worked out of this thing.
_________________
They don't have a good bathroom to do coke in.
Back to top
View user's profile Send private message
_puck_
Tux's lil' helper
Tux's lil' helper


Joined: 12 May 2002
Posts: 111
Location: Kaiserslautern / Germany

PostPosted: Mon Aug 19, 2002 9:38 pm    Post subject: Reply with quote

Dolio: Thats what I thought first. I bought it as a Athlon "C"-series ... A friend of mine thinks palomino is the same than the xp, I don't believe so, but he is very into the stuff. Here is my /proc/cpuinfo, maybe someone can tell if I can use -march=athlon-4 or just plain athlon, and of course IF it is a palomino or a thorowbred (I don't know how to spell, I googled and din't find a hint) ...

Code:

processor       : 0
vendor_id       : AuthenticAMD
cpu family      : 6
model           : 4
model name      : AMD Athlon(tm) Processor
stepping        : 4
cpu MHz         : 1399.924
cache size      : 256 KB
fdiv_bug        : no
hlt_bug         : no
f00f_bug        : no
coma_bug        : no
fpu             : yes
fpu_exception   : yes
cpuid level     : 1
wp              : yes
flags           : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 mmx fxsr syscall mmxext 3dnowext 3dnow
bogomips        : 2752.51

Thats the CFLAGS I am using for the bootstrap and the system at the moment (normal optimizations already made a compile) ... better than my first attempt mesa thinks. ;)

Code:

CFLAGS="-march=athlon -loop -fforce-addr -frerun-loop-opt -fomit-frame-pointer -funroll-loops -O3 -pipe -mmmx -m3dnow"


Waiting and hoping.
_________________
regards
puck
___________________________
#There is more than one way to do it.
Back to top
View user's profile Send private message
El_Presidente_Pufferfish
Veteran
Veteran


Joined: 11 Jul 2002
Posts: 1179
Location: Seattle

PostPosted: Mon Aug 19, 2002 10:11 pm    Post subject: Reply with quote

You have an athlon-tbird
not a palomino, which IS the same thing as an athlon4 and XP
Back to top
View user's profile Send private message
_puck_
Tux's lil' helper
Tux's lil' helper


Joined: 12 May 2002
Posts: 111
Location: Kaiserslautern / Germany

PostPosted: Mon Aug 19, 2002 10:26 pm    Post subject: Reply with quote

With the following CFLAGS I get this error, can anybody tell me why? It is, as you can see, due the first gcc-compile-run while bootstrapping:

Code:

CFLAGS="-march=athlon -fforce-addr -frerun-loop-opt -fomit-frame-pointer -funroll-loops -O3 -pipe -mmmx -m3dnow"


Code:

multilib.h is unchanged
echo timestamp > s-mlib
stage1/xgcc -Bstage1/ -B/usr/i686-pc-linux-gnu/bin/ -DIN_GCC    -g -O2 -W -Wall -Wwrite-strings -Wstrict-prototypes -Wmissing-prototypes -Wtraditional -pedantic -Wno-long-long  -DHAVE_CONFIG_H -DGENERATOR_FILE  -o gencheck \
 gencheck.o ../libiberty/libiberty.a
stage1/xgcc -Bstage1/ -B/usr/i686-pc-linux-gnu/bin/ -DIN_GCC    -g -O2 -W -Wall -Wwrite-strings -Wstrict-prototypes -Wmissing-prototypes -Wtraditional -pedantic -Wno-long-long  -DHAVE_CONFIG_H -DGENERATOR_FILE  -o genpreds \
 genpreds.o ../libiberty/libiberty.a
./gencheck > tmp-check.h
make[2]: *** [s-check] Error 139
make[2]: *** Waiting for unfinished jobs....
make[2]: Leaving directory `/var/tmp/portage/gcc-3.2/work/build/gcc'
make[1]: *** [stage2_build] Error 2
make[1]: Leaving directory `/var/tmp/portage/gcc-3.2/work/build/gcc'
make: *** [bootstrap-lean] Error 2

!!! ERROR: The ebuild did not complete successfully.
!!! Function src_compile, Line 14, Exitcode 2
!!! (no error message)

!!! emerge aborting on  /usr/portage/sys-devel/gcc/gcc-3.2.ebuild .

_________________
regards
puck
___________________________
#There is more than one way to do it.
Back to top
View user's profile Send private message
aardvark
Guru
Guru


Joined: 30 Jun 2002
Posts: 576

PostPosted: Mon Aug 19, 2002 11:01 pm    Post subject: Reply with quote

_puck_ wrote:
With the following CFLAGS I get this error, can anybody tell me why? It is, as you can see, due the first gcc-compile-run while bootstrapping:

Code:

CFLAGS="-march=athlon -fforce-addr -frerun-loop-opt -fomit-frame-pointer -funroll-loops -O3 -pipe -mmmx -m3dnow"


Code:

multilib.h is unchanged
echo timestamp > s-mlib
stage1/xgcc -Bstage1/ -B/usr/i686-pc-linux-gnu/bin/ -DIN_GCC    -g -O2 -W -Wall -Wwrite-strings -Wstrict-prototypes -Wmissing-prototypes -Wtraditional -pedantic -Wno-long-long  -DHAVE_CONFIG_H -DGENERATOR_FILE  -o gencheck \
 gencheck.o ../libiberty/libiberty.a
stage1/xgcc -Bstage1/ -B/usr/i686-pc-linux-gnu/bin/ -DIN_GCC    -g -O2 -W -Wall -Wwrite-strings -Wstrict-prototypes -Wmissing-prototypes -Wtraditional -pedantic -Wno-long-long  -DHAVE_CONFIG_H -DGENERATOR_FILE  -o genpreds \
 genpreds.o ../libiberty/libiberty.a
./gencheck > tmp-check.h
make[2]: *** [s-check] Error 139
make[2]: *** Waiting for unfinished jobs....
make[2]: Leaving directory `/var/tmp/portage/gcc-3.2/work/build/gcc'
make[1]: *** [stage2_build] Error 2
make[1]: Leaving directory `/var/tmp/portage/gcc-3.2/work/build/gcc'
make: *** [bootstrap-lean] Error 2

!!! ERROR: The ebuild did not complete successfully.
!!! Function src_compile, Line 14, Exitcode 2
!!! (no error message)

!!! emerge aborting on  /usr/portage/sys-devel/gcc/gcc-3.2.ebuild .


Same problem here, whichever flags I use. Looks like emerging a 1.4 is not an option at the moment.
Back to top
View user's profile Send private message
_puck_
Tux's lil' helper
Tux's lil' helper


Joined: 12 May 2002
Posts: 111
Location: Kaiserslautern / Germany

PostPosted: Tue Aug 20, 2002 12:17 am    Post subject: Reply with quote

aardvark: With the standard CFLAGS all went fine. I did an emerge system after the bootstrap and it run pretty fast. I have to say, I have not the 1.4 tarball from the ibiblio.org server ... it's named stage1-x86-1.4-20020818-2037.tbz2
Code:

CFLAGS="-march=i686 -O3 -pipe"

But with those cflags, which SHOULD work, I have no success.

Nice to have someone in company, also it is not nice for both of us. ;) I thought by the time, I go insane and making an obvious failure.
_________________
regards
puck
___________________________
#There is more than one way to do it.
Back to top
View user's profile Send private message
phong
Bodhisattva
Bodhisattva


Joined: 16 Jul 2002
Posts: 778
Location: Michigan - 15 & Ryan

PostPosted: Tue Aug 20, 2002 5:28 am    Post subject: Reply with quote

Arg, I'm also having the same error (genpreds) emerging gcc - actually it's a bootstrap on a secondary system I have at home, using the 1.3b tarball, gcc-3.1.1-r1. I've done lots of installations, and haven't run into this before... Has something changed somewhere?
_________________
"An empty head is not really empty; it is stuffed with rubbish. Hence the difficulty of forcing anything into an empty head."
-- Eric Hoffer
Back to top
View user's profile Send private message
dol-sen
Developer
Developer


Joined: 30 Jun 2002
Posts: 2503
Location: Richmond, BC, Canada

PostPosted: Tue Aug 20, 2002 6:40 am    Post subject: gcc 3.2 speed Reply with quote

It sounds like gcc3.2 code runs faster :lol: . Is this only for the new cpu's or will it improve my old slug P1 cpu. Remember, slightly faster code for the new fast cpu's usually means much faster for the slugs, only if the speedup is not just for the new cpu optimizations.


Brian

Eagerly awaiting the day I can upgrade my hardware, :idea: planning how to comandeer my daughters 1gHZ. Athalon system to build a gcc 3.2 system for my slug :wink: That should shave 3-4 days off an emerge -e world.
Back to top
View user's profile Send private message
Flake
n00b
n00b


Joined: 19 Apr 2002
Posts: 44
Location: The Netherlands

PostPosted: Tue Aug 20, 2002 11:02 am    Post subject: Reply with quote

probably lacking the brains and compensating this in balls i decided to try the upgrate to gcc 3.2 too.

the choice was easy in a way, since my HD space is limited to a little less than 2gig!

more and more updates required a rebuild of X, and i can tell now, 750 MB is not enough to build X! (same for gcc3.2) So, had to clear Gnome2 as well, along with all the libraries that come with it.

it can tell you that /var/db/pkg is starting to get an empty space !

i have managed to free up to 900 MB.

i hope that this is enough to build gcc 3.2. if not, there is little left i can delete!! a base gentoo system looks like a 1 gig minimum !

already set the DISTDIR=/mnt/windows/linux/ (dont ask)
emptied /var/tmp/portage/ etc etc.

I haven't deleted any of the tmp and .o files. I couldn't find any good links on which ones can be deleted etc.

so my questions /remarks:

1. is 900 MB enough to build gcc 3.2 (i'll know in 6 hours anyhow)
2. can i remove gcc 2.9 safely thereafter and have enough space left to build glib ? and then built gcc again ??? (i'll know in a day)
3. what is the absolute minimum to build X ???
4. what are the last resort methods to free up HD space ?
Back to top
View user's profile Send private message
_puck_
Tux's lil' helper
Tux's lil' helper


Joined: 12 May 2002
Posts: 111
Location: Kaiserslautern / Germany

PostPosted: Tue Aug 20, 2002 2:46 pm    Post subject: Reply with quote

Hey folks. I did an bootstrap and emerge system with a 1.4beta with the following CFLAGS ... now I will add more of teh CFLAGS seeing if and when it fails.

Code:

CHOST="i686-pc-linux-gnu"
CFLAGS="-march=athlon -O3 -pipe -mmmx -m3dnow"
CXXFLAGS="-march=athlon -O3 -pipe -mmmx -m3dnow"

_________________
regards
puck
___________________________
#There is more than one way to do it.
Back to top
View user's profile Send private message
warrax
n00b
n00b


Joined: 20 Aug 2002
Posts: 13

PostPosted: Tue Aug 20, 2002 10:00 pm    Post subject: Reply with quote

Just thought I'd mention that I've written a Python version (with some improvements over the original) of Dolio's rmerge script.

See http://bugs.gentoo.org/show_bug.cgi?id=5667 for details. (The script is attached to the bug as "rmerge2-0.9.tar.gz").

As always: I've only tested on my own machine so: no guarantees, it may cause your computer to blow up, or your head to spontaneously combust, etc.etc. That said, it should work, and I'd certainly appreciate some testing & comments.
Back to top
View user's profile Send private message
Dolio
l33t
l33t


Joined: 17 Jun 2002
Posts: 650

PostPosted: Tue Aug 20, 2002 10:02 pm    Post subject: Reply with quote

If you take a look at the rmerge bug report:

http://bugs.gentoo.org/show_bug.cgi?id=5667

You'll see that rmerge has been re-implemented by someone else (Bardur Arantsson) in Python. He added a couple new features to the script, and made it shorter/easier to read. I'd recommend trying out his script, since he's said he's tested it pretty thoroughly (no guarantees of course), and Python is probably the way to go (easier to add to the Portage package directly, if someone so desires). Plus, I'll be gone for a couple days, so if you encounter any bugs in his script, he'll be able to fix them, while bugs in my script won't be fixed (by me at least), for a few days.

There's no ebuild for the new script yet, but it's easy enough to just untar some place and run.

Good luck with your rebuilding.
_________________
They don't have a good bathroom to do coke in.
Back to top
View user's profile Send private message
puddpunk
l33t
l33t


Joined: 20 Jul 2002
Posts: 681
Location: New Zealand

PostPosted: Wed Aug 21, 2002 6:11 am    Post subject: Reply with quote

Quote:

[gentoo-dev] Gentoo 1.2 to 1.4 update, HowTo do it and not break everything

Spider spider@gentoo.org
Mon, 12 Aug 2002 19:48:14 +0000

Previous message: [gentoo-dev] GLSA: netscape-flash
Next message: [gentoo-dev] Gentoo 1.2 to 1.4 update, HowTo do it and not break everything
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
--0enmDZ=.SLgUq(rp
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit

Wheehey.
time for the preliminary upgrade instructions for all those adventurous
souls that dont want to break everything but only something :)


Overview:
how to update from gcc 2.95.3 to 3.2 without breaking too much

1) build gcc / glibc /binutils / gettext packages to backup with
emerge --buildpkg gcc glibc gettext binutils (all in gcc 2.95.3)

Move theese aside in a place tagged as gcc 2.95.3 .


2) build an replacement gcc systems
change the profile to default-x86-2.0

export USE="-* build"
emerge --buildpkg gcc
emerge --buildpkg binutils gettext texinfo glibc
unset USE

This will now replace your gcc with one without c++ and other such
things.. "oops" gcc is now gcc 3.2.. thats good.

Now, We remove all traces of gcc, *shudder*
emerge -C gcc
and install the binary again:
emerge --usepkg gcc

now, next step... we want a fully working and competent gcc, glibc and
other things:

emerge --buildpkg gcc
emerge --buildpkg binutils gettext texinfo glibc

Lovely, a complete gcc 3.2 system installed! and not a trace of gcc
2.95.3's libstdc++ ... This is the time when you go "oops" and whine a
lot.

or.. is it?

move aside your gcc 3.2 complete binaries, copy in gcc 2.95.3 ones
emerge --usepkg =gcc-2.95.3-r7 (whatever version it was you had
installed)
emerge --usepkg gcc

Now, you've restored gcc 2.95.3's paths and data, and overwritten the
gcc, g++ and others with those of gcc 3.2

check with /usr/bin/gcc -v
it reports gcc 3.2 for me


Now, edit /etc/env.d/05gcc
add :
LDPATH=/usr/lib/gcc-lib/i686-pc-linux-gnu/2.95.3

then run "env-update"


now your old c++ code will work again... everything is normal, until ...
soon..


This is as far as I've come today, my system works, c++ code works,
galeon works, blackbox runs.

Beware though... when you start to mix c++ codes it will probably break
in strange ways. Solution: recompile :)

I hope this helps.....


//Spider


All heil spider!!!!! thank you very much muh man. I hope this can help other people see the light of GCC with out killing something...
Back to top
View user's profile Send private message
Flake
n00b
n00b


Joined: 19 Apr 2002
Posts: 44
Location: The Netherlands

PostPosted: Wed Aug 21, 2002 11:24 am    Post subject: Reply with quote

Well i can tell you now you need approx. 800+ MB to compile gcc-3.2 :D

(not making a binary at the same time, though i think that would require little extra space)

Unfortunately i see spiders how-to too late.


i already compiled gcc3.2, chanced the default compiler, piled glibc and gcc again. no binaries to replace the stuff with.

well, i'll see where the ship hits the beach... (Dutch saying)

here goes nothing.....


i still don't see anything about compiling a new kernel.. is it not nececairy ???
Back to top
View user's profile Send private message
dizzy
n00b
n00b


Joined: 04 Aug 2002
Posts: 50
Location: Melb.OZ.Terra.Sol

PostPosted: Wed Aug 21, 2002 4:17 pm    Post subject: Reply with quote

There's a chance you might lose the timezone stuff from glibc r7. Just something to look out for.
Back to top
View user's profile Send private message
Display posts from previous:   
Reply to topic    Gentoo Forums Forum Index Portage & Programming All times are GMT
Goto page Previous  1, 2, 3, 4, 5, 6, 7, 8  Next
Page 3 of 8

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum