Stage 1/3 Installation Support - 2005.0 & GCC 3.4.3

[Bob P]

I suppose a few people could call my C-flags a bit OTT, but I managed to follow the original guide fine.

if you are having problems compiling glibc, i can offer two suggestions. i hope that they will help:

1. the original set of "insane" cflags listed in the guide were used in an example for intel architecture, and i can't say anything about using them on AMD. my personal recommendation would be that if anyone is having problems compiling, you should revert to the stable/recommended/minimal cflags in the guide and assure that the problem is replicated with the conservative set of cflags before asking for help.

2. failure to compile glibc is not related in any way to a defect in the Guide. it is a problem that crops up intermittently when you're working with an experimental ebuild that has problems. this is just a fact of life in the testing branch. if you are having problems with the current version of glibc, then look at bugzilla for existing bug reports. if there is not an existing bug report, then file one. your best bet would probably be to mask the offensive ebuild and try again with one that is not presently in a state of revision.

[Sith_Happens]

sorry, i'm forgetting i'm compiling on a pentium4 3.06 gig slap myself round the face and be a bit more patient

I'm currently trying to replicate your problems on a pentium 4 3.0 ghz, so I should be able to get the job soon. In the meantime here is the output of emerge info for test install:

Code: Select all

Portage 2.0.51.19 (default-linux/x86/2005.0, gcc-3.3.5, glibc-2.3.4.20040808-r1, 2.6.11-love2 i686)
=================================================================
System uname: 2.6.11-love2 i686 Intel(R) Pentium(R) 4 CPU 3.00GHz
Gentoo Base System version 1.4.16
Python:              dev-lang/python-2.3.4-r1 [2.3.4 (#1, Mar 11 2005, 13:17:01)]
dev-lang/python:     2.3.4-r1
sys-devel/autoconf:  2.59-r6, 2.13
sys-devel/automake:  1.7.9-r1, 1.8.5-r3, 1.5, 1.4_p6, 1.6.3, 1.9.4
sys-devel/binutils:  2.15.92.0.2-r1
sys-devel/libtool:   1.5.10-r4
virtual/os-headers:  2.6.8.1-r2
ACCEPT_KEYWORDS="x86"
AUTOCLEAN="yes"
CFLAGS="-O2 -march=i686 -fomit-frame-pointer"
CHOST="i686-pc-linux-gnu"
CONFIG_PROTECT="/etc /usr/kde/2/share/config /usr/kde/3/share/config /usr/lib/X11/xkb /usr/share/config /var/qmail/control"
CONFIG_PROTECT_MASK="/etc/gconf /etc/terminfo /etc/env.d"
CXXFLAGS="-O2 -march=i686 -fomit-frame-pointer"
DISTDIR="/usr/portage/distfiles"
FEATURES="autoaddcvs autoconfig ccache distlocks sandbox sfperms"
GENTOO_MIRRORS="http://gentoo.llarian.net"
MAKEOPTS="-j2"
PKGDIR="/usr/portage/packages"
PORTAGE_TMPDIR="/var/tmp"
PORTDIR="/usr/portage"
SYNC="rsync://rsync.gentoo.org/gentoo-portage"
USE="x86 X apm arts avi berkdb bitmap-fonts crypt cups emboss encode foomaticdb fortran gdbm gif gnome gpm gtk gtk2 imlib ipv6 jpeg kde libg++ libwww mad mikmod motif mp3 mpeg ncurses nls nptl oggvorbis opengl oss pam pdflib perl png python qt quicktime readline sdl spell ssl svga tcpd truetype truetype-fonts type1-fonts xml2 xmms xv zlib"
Unset:  ASFLAGS, CBUILD, CTARGET, LANG, LC_ALL, LDFLAGS, PORTDIR_OVERLAY

And here is the make.conf (it's stripped down to only the essentials for this stage of the install):

Code: Select all

# These settings were set by the catalyst build script that automatically built this stage
# Please consult /etc/make.conf.example for a more detailed example
CFLAGS="-O2 -march=i686 -fomit-frame-pointer"
CHOST="i686-pc-linux-gnu"
CXXFLAGS="${CFLAGS}"
GENTOO_MIRRORS="http://gentoo.llarian.net"
USE="nptl"

[Bob P]

I'm currently trying to replicate your problems on a pentium 4 3.0 ghz, so I should be able to get the job soon.

young whippersnapper!

[Sith_Happens]

I'm currently trying to replicate your problems on a pentium 4 3.0 ghz, so I should be able to get the job soon.

young whippersnapper!

Ha..
An update on that I've just finished rebuilding the glibc with the 2005.0 i686 stage 3 tarball, so cheops05, your problem probably lies either in your make.conf, or in the commands you used to rebuild the toolchain. Why don't you post both and we'll see whats up.

[Bob P]

okay, tell me ... how long does it take to compile glibc on a 3.0 GHz P4? inquiring minds want to know!

[Sith_Happens]

I recompiled gcc-config (that one doesn't really count ), glibc, binutils, and gcc in under 2 hours.

[grx]

I have a couple questions--forgive me for being a n00b...

Why not do this type of install from the Universal CD with it's stage 3 tarball instead of using the minimal CD?

And is there a reason this install method uses ReiserFS, or when you're putting the filesystem on should it be more of a <place your favorite fs here> step?

[Sith_Happens]

I have a couple questions--forgive me for being a n00b...

Why not do this type of install from the Universal CD with it's stage 3 tarball instead of using the minimal CD?

And is there a reason this install method uses ReiserFS, or when you're putting the filesystem on should it be more of a <place your favorite fs here> step?

If you want to you can do this install from a Universal CD. The only reason for not using it is that it's a big download, and you really only need a snapshot, a stage 3 tarball, and a minimal cd. I can't speak for Bob, but I too would recommend ReiserFS (3.6 not 4) as a fast and stable filesystem. However, if you want to use another filesystem, go ahead (again not speaking for Bob P). You shouldn't run into any problems, and you won't have to deviate from the guide enough to obviate at least my support if you run into any problems.

[Bob P]

I can't speak for Bob, but I too would recommend ReiserFS (3.6 not 4) as a fast and stable filesystem. However, if you want to use another filesystem, go ahead (again not speaking for Bob P). You shouldn't run into any problems, and you won't have to deviate from the guide enough to obviate at least my support if you run into any problems.

EVERYONE SHALL USE REISER FS 3.6 AND THEY SHALL LIKE IT!

just kidding. it doesn't matter which fs you choose to use, as long as you know what you are doing. just don't try reducing the size of the /boot partition if you user reiser, as you could get into trouble if you fail to take into account the size of the journal files.

[kimchi_sg]

Morning all...

I've just watched this thread grow by 1 page in the course of my sleep!

A warning to all n00bs who are reading this thread:

If you do not understand what exactly each of the steps in stage 1-on-3 do, stick with the Standard Gentoo Installation Handbook. DO NOT EVER LET US CATCH YOU TRYING TO BOOTSTRAP WITH THE STAGE 1-ON-3 TUTORIAL AND / OR SETTINGS.

[Sith_Happens]

If you do not understand what exactly each of the steps in stage 1-on-3 do, stick with the Standard Gentoo Installation Handbook. DO NOT EVER LET US CATCH YOU TRYING TO BOOTSTRAP WITH THE STAGE 1-ON-3 TUTORIAL AND / OR SETTINGS.

Well they can try, just don't ask me for help.

[kimchi_sg]

If you do not understand what exactly each of the steps in stage 1-on-3 do, stick with the Standard Gentoo Installation Handbook. DO NOT EVER LET US CATCH YOU TRYING TO BOOTSTRAP WITH THE STAGE 1-ON-3 TUTORIAL AND / OR SETTINGS.

Well they can try, just don't ask me for help.

Please don't weaken the force of my warning.

2 n00bs breaking themselves on "stage 1-on-3" in 1 night is cause enough for concern.

[grx]

Just a little defense for the n00bs--

My main concern with this install is getting gcc 3.4.3, because I'm using a centrino laptop, and I'd like to have the pentium-m optimization. Doesn't mean I'm going to follow this particular guide and then complain bitterly if it breaks. Just trying to learn more, and this thread is related to what I'm trying to do.

[kimchi_sg]

My main concern with this install is getting gcc 3.4.3, because I'm using a centrino laptop, and I'd like to have the pentium-m optimization.

The optimisation in gcc for the pentium-m is minuscule, compared to optimising for pentium 3 using gcc 3.3.5. My opinion is that it is far too risky just for this change.

Remember that gcc 3.3.5 was cutting edge not too long ago. In other words, do not look down on gcc 3.3.x.

Far more optimisation can be obtained by configuring system correctly in other areas. For example, choosing kernel options wisely, and using as few USE flags as necessary.

[Bob P]

My main concern with this install is getting gcc 3.4.3, because I'm using a centrino laptop, and I'd like to have the pentium-m optimization. Doesn't mean I'm going to follow this particular guide and then complain bitterly if it breaks. Just trying to learn more, and this thread is related to what I'm trying to do.

this pentium-m issue with gcc 3.3.4 is widely documented. everyone with a pentium-m should search and read the many, many threads that will come up. in fact, duplicate posts went up on this topic today because a n00b didn't bother to search. in reading them, it seems evident that the person who is making the most noise has absolutely no clue what he is doing.

[grx]

it seems evident that the person who is making the most noise has absolutely no clue what he is doing.

That's possible--but not everyone new to this distro is clueless. I've been doing my research and reading those threads too, as well as ones about using 3.3.x with pentium-m. I'm just researching and comparing this method to the installation manual, which seems nearly the same if you use a stage-3 install, but with some specific cflag/USE variable changes. I certainly wouldn't think of trying to bootstrap using this as a guide since it's not intended for that at all. But I do want to learn more about what I'm doing before I sit down and let the machine crank out my os. Long story short (and hopefully ended ) I'm glad for this thread so I can learn more about what I'm going to be doing to my computer.

[Bob P]

well, this guide is pretty clear. use pentium for the first compile, and then when you have gcc 343 on your computer, edit your make.conf settings and recompile your toolkit in the final form. its a chipshot for you to put pentium-m in at that stage of the game. have fun!

[cheops05]

I have used the 2004.3 guide so many times that I can practically do it in my sleep just because my post count is not touching the moon does not mean that I am clueless and dont know what I'm doing I have just never needed help, I thought the idea of this topic is to get the rough draft 100% fullproof for the new release of 2005.0, I am not moaning I am just telling you that glibc breaks, trying to get gcc 3.4.3 compiler onto my machine just because it doesn't break on your machines does not mean it doesn't break on someone elses.

Anyway I just to let you know I am trying again for the 4th time from scratch, changing my CFLAGS etc until I pinpoint the problem

[sn0wrider]

Hi!

Excellent piece of work Bob!

Just one thing though..when I compile gcc I get a couple of errors...problem is that I can't seem to find out how I'm supposed to copy-aspte the error messages in the terminal!

Could someone enlighten me in this small matter?

///Sn0wie

[sn0wrider]

Morning all...

If you do not understand what exactly each of the steps in stage 1-on-3 do, stick with the Standard Gentoo Installation Handbook. DO NOT EVER LET US CATCH YOU TRYING TO BOOTSTRAP WITH THE STAGE 1-ON-3 TUTORIAL AND / OR SETTINGS.

Oh and one more thing:

My problem is that I do not understand each step in Bobs guide.. BUT it seems to me, however, that the old manual doesn't work...when I've followed it step by step and then tries to emerge gentoo-sources I get a message sayin: "portage can not find any ebuilds to satisfy your request"
And nothing happens..... exactly the same thing happened when I tried (out of curiosity) to emerge development-sources as well.

Then I tried the Bobs 2005.0 "manual" and got errors when compiling gcc ...(and the error messages for that I will have to get back to you about).

So me as a n00b is quite stuck here...what are my greates chances of success? Must I DL the entire LibveCD and make a networkless install? And then put my computer up for a few days of catching everything up? (upgrading)

Or is there other possibilities? I mean for me it seems "madness" to do a networkless install when I have a 10mbit connection to get everything up to date immediatteky...


///Sn0wie

[Gentoonie]

now how's this for an idea -- instead of forcing every user that wants to install GCC 3.4.3 to rebuild the entire toolkit, why don't we just make an unauthorized live CD and stageballs that feature an ~x86 toolkit?

it would save an awful lot of time to do the heavy lifting once, and burn the results to an alternative Live CD. who's in?

This is a real excellent idea. Building toolkits and especially this F**ing glibc takes ages even on my Pentium 4 2.8 GHz.

An already prepared live CD would save so many time to the users

[LucaSpiller]

Just one thing though..when I compile gcc I get a couple of errors...problem is that I can't seem to find out how I'm supposed to copy-aspte the error messages in the terminal!

Could someone enlighten me in this small matter?

No problem. If you are in a chrooted enviroment, hopefully you ran screen. If so just detatch it, and then open it in a terminal under X. If not then just type:

Code: Select all

gpm -m /dev/mouse -t ExplorerPS/2

-m is where you mouse is, and -t is the protocol; both of which are in the X config file (well X.org at least).

I did a bit of reasearch and found that enabling sandbox mode can sometimes cause errors when compiling glibc, so I disabled it (removing sandbox and usersandbox from /etc/make.conf FEATURES="...") and I didn't get the above error any more. Unfortunately I got another one, but sort of removed the screen instead of detaching it - so I will post the results later.

[sn0wrider]

Just one thing though..when I compile gcc I get a couple of errors...problem is that I can't seem to find out how I'm supposed to copy-aspte the error messages in the terminal!

Could someone enlighten me in this small matter?

No problem. If you are in a chrooted enviroment, hopefully you ran screen. If so just detatch it, and then open it in a terminal under X. If not then just type:

Code: Select all

gpm -m /dev/mouse -t ExplorerPS/2

-m is where you mouse is, and -t is the protocol; both of which are in the X config file (well X.org at least).

I did a bit of reasearch and found that enabling sandbox mode can sometimes cause errors when compiling glibc, so I disabled it (removing sandbox and usersandbox from /etc/make.conf FEATURES="...") and I didn't get the above error any more. Unfortunately I got another one, but sort of removed the screen instead of detaching it - so I will post the results later.

Problem is that I have not come so far as to emerging xorg yet ... I'm trying to do a stage 3 install using genkernel (as of now..I tried maunally configuring the .conf file at first..it turned out to be some things I couldn't get right though) ...and following Bobs guide I must recompile gcc to get it working... and thats where I get my errors.

[vipernicus]

glibc fails maketest everytime for me, and I'm using your default settings, when i remove maketest, it works.

[kimchi_sg]

To anyone who adds to / changes settings when following this tutorial:

You are free to add to or amend the settings we have listed in the tutorial proper. However, do keep in mind that by doing this, we are also free not to help you.