Yeah, my plan was to first update the compilers, then the rest.

Thnx

I'll report back later (I expect several days )

Hi all,

I have been a bad gentoo user, and not done updates regularly for like 2 years (I made updates, but not all).

I ran "emerge -uvD world -p", I don't think anyone will look through it, so I will not attach it. I could think of some possible problematic packages, namely gcc, glibc, phython ...

thnx for your answer.

I'll try it again later. My system is quite outdated, I think I'll be working on that in the next days.

Hi,

bought that all-in-one printer yesterday, finally printed the first page some minutes ago (most time was consumed by system upgrades (X server, pam))

So I installed the newest hplip (2.7.9, ~x86 masked), worked through all the accompanying upgrades and problems.
Started hp-setup, printer was ...

I just looked at my "glxinfo" output, and it turned out that direct rendering is not working for me too.
Well, at least glxgears works on both displays with the second setup (which was not the case with the first config), and there was a notable speedup...

perhaps someone else got this working ...

whoever may be interested in this (and I assume there some people are, since there are several threads related to this one on the first side)

As mentioned above, xinerama does a god job, but there is no graphics acceleration. This turned out to be a problem, since even opera started jumping when I ...

well, I got it working, with Xinerama

The trick was to specify the same BusID for both devices

so here is my new xorg.conf:
Section "dri"
Mode 0666
EndSection


Section "Module"
Load "dbe" # Double buffer extension
SubSection "extmod"
Option "omit xfree86-dga" # don't initialise the DGA ...

EDIT:
Found some solutions, see 2nd and 3rd post. I just noticed that this Thread already got ~150 views. wow. Would be glad to know if it worked for some of you too

So for two days now I am trying to set up a dual head system.

Radeon 9600 XT with one DVI and one VGA out
1280x1024 lcd connected ...

thnx, had similar problem

Solution:

Upgrade to ~x86 portage (2.1 something) solves the problem


It seems to solve most (if not all) insecure RUNPATH problems, so perhaps there should be a sticky until the new portage becomes stable.

same problem here, updating from 5.8.5

I can remember having (had) that problem quite some time ago.
It had been a problem with gtk, gtk2 USE Flags set/not set, I think I had to recompile mplayer without gtk2 support.

Well, set up my system newly, and I didn't use mplayer for a while until I emerged it today. It is blue. The gtk2 flag ...

I am emerging kile currently, and I just saw passing by this piece of information (emerge 3/10):
You must rebuild other packages depending on t1lib.
You may use revdep-rebuild (from app-portage/gentoolkit)
to do all necessary tricks.
Ok, no problem for me, as I am setting up a new System.
But now ...

It looks simple enough.

sm3600.h:279: __SM3600EXPORT__ TState GetScannerModel(unsigned short idVendor, unsigned short idProduct);

sm3600-scanmtek.c:68: __SM3600EXPORT__
TModel GetScannerModel(unsigned short idVendor,
unsigned short idProduct)

You need to patch sm3600.h to declare ...

I am still failing to compile sane-backends. Errors:
gcc -c -march=pentium4 -O2 -pipe -fomit-frame-pointer -W -Wall -DHAVE_CONFIG_H -I. -I. -I../include -I../include -DPATH_SANE_CONFIG_DIR=/etc/sane.d -DPATH_SANE_DATA_DIR=/usr/share -DV_MAJOR=1 -DV_MINOR=0 -DSTUBS -DBACKEND_NAME=snapscan snapscan-s ...

According to gcc-3.4 manual, there is no such cpu-ype supported.

http://gcc.gnu.org/onlinedocs/gcc-3.4.0 ... %20Options

Only pentium3 and pentium4 are valid options for -march=cpu-type (of course there are others, but nothing for Pentium M).

You need to recompile all the kde packages with gcc 3.4 for things to get normal.
I had that problem with arts too, recompile kdemultimedia and it should go away.
You should recompile kdemultimedia with DO_NOT_COMPILE="noatun noatun-plugins" and before that make sure you recompile id3tag and ...

I am running happily with the following flags:

"-march=pentium4 -O3 -fomit-frame-pointer -pipe"

nothing else. P4 flag didn't make any problems so far. Of course you should choose Pentium M in your Kernel, the Centrino CPU is simply an Pentium M.

So, a little summary of my experiences:

qt-3.3.2 compiled without problems, so did arts, kdelibs, kdebase
arts keeps crashing, so does kcminit (part of kdebase), this could be because of new qt, new gcc, glibc or whatever.
some video codecs seem to be broken, don't know exactly which ones ...

qt 3.3.2 is went out today. It should compile with gcc 3.4.0:

Qt/Embedded: Compiles with gcc 3.4.0 (prerelease).

No ebuild yet.

Thanx, that's it.
So, the easyest workaround is to put FEATURES="-sandbox" in your make.conf. It is set as default in make.globals, but this is not the place to change it.

Well, even quicker is this one:

Doesn't work for me. I never had FEATURES set. I tried setting them, removing some of them, it is all the same. Perhaps you could post the output of "emerge info" to see what you have set.

Thanx guys.
Didn't have time neither to test this nore to reply here. I will check it out, now I too remember some issue with changing back to xfree-opengl and such things. Also, stopping X first makes sense.
I'll try.

EDIT: Ebuild still not working. Don't know what to do. The Access Violation ...

Hi.
My Problem:
gentoo root # emerge ati-drivers
Calculating dependencies ...done!
>>> emerge (1 of 1) media-video/ati-drivers-3.7.6-r1 to /
>>> md5 src_uri ;-) fglrx-4.3.0-3.7.6.i386.rpm
* X11 implementation is xfree.
>>> Unpacking source...
>>> Unpacking fglrx-4.3.0-3.7.6.i386.rpm
* Applying ...

Thanx Bob and Evil, I followed your explanations and it is working now.

First I had to change my kernel, cause mm based love-sources had problems with hotplug, but now it is working. great.