View previous topic :: View next topic |
Author |
Message |
hielvc Advocate
Joined: 19 Apr 2002 Posts: 2805 Location: Oceanside, Ca
|
Posted: Wed May 31, 2006 11:36 pm Post subject: |
|
|
Actually I think it was dirtyEpic who first mentioned it. Basicly this my approche after a couple of years of building the TC, stage1 stage2, stage3 install, bootstraping stage1s 2s and 3s, command line TC builds, wathcing and doing the gcc-beta threads and most importantly what has worked for me. In most cases your starting from a stable TC. When I started this thread I didnt have the experince I have now and I knew I couldnt supoort the way Bob_P does so to play it safe and from my experince of then thats what I said. Back to the topic, if I have a stable TC it will build a stable gcc-update. If its not capable of doing this then linux, heck all OS's have an insermountable problem.
Originlly bootstrap meant that you built your TC and a few other progs 3 times. That meant gcc was built 9 times. Somebody somewhere said to heck with this and only built the TC progs once and had no problems. That was in late 2003 I think maybe early 2004 . When gcc is built it bootstraps itself, in other words builds its self 3 times. 1st gcc-old builds gcc-new, 2nd time gcc-new[old] builds gcc-new[new-old] and then 3rd time gcc-new[new-old] builds gcc-new[new-new-old] at each stage gcc runs checks to see if its working correctly and in the stage2 and stage3 that compiliers are the same. This happens every time you emerge gcc.
So if your only upgrading gcc then I ve found that building it first. then the rest of the TC works and is quicker. If its glibc major upgrade then linux-hheaders are thrown in to be safe. Then glibc binutils gcc. There are times though when this, for gcc-testing, that you have to build binutils first then gcc, but those guys know this. This script is for regular users arent realy interested in the fine points of the TC or emergeing, they just want a save easy way to do it and or break things into smaller chunks. They actually use their computers for more than or only for web browsing email school and work. They think Im nuts for playing "break my gentoo" _________________ An A-Z Index of the Linux BASH command line |
|
Back to top |
|
|
hielvc Advocate
Joined: 19 Apr 2002 Posts: 2805 Location: Oceanside, Ca
|
Posted: Wed May 31, 2006 11:52 pm Post subject: |
|
|
zigver The system build time dependencies are fairly simple and a single go through will usually 99% of the time build them correctly, though with all x11 stuff starting to show up it WOULD BE wiser to build the system-TC twice.
Yes -s and -w just synonyms for system and world and yes its system-TC and world-TC. The "-W" is world-system-TC. Ill look at the documentation. _________________ An A-Z Index of the Linux BASH command line |
|
Back to top |
|
|
zigver Tux's lil' helper
Joined: 09 May 2003 Posts: 87
|
Posted: Thu Jun 01, 2006 2:57 am Post subject: |
|
|
Running through the first "emwrap.sh -W", it hung on the "these failed to build" list. I CTRL-D'd and then it output this:
Code: |
You can now enter " f " to build the faild file list.If you dont enter anything
then the faild file will be deleted and emwrap.sh system -e will continue.
You only need to run "emwrap.sh faild" after doing a emptytree system or world.
If you want to stop and scratch your head do a CRTL-C
Input was empty I will continue with emwrap.sh system -e
|
This was a world (-W) build and not system. I decided to stop and scratch my head so what would have happened next if I allowed it to continue? All of the world packages were rebuilt except the failed list. |
|
Back to top |
|
|
hielvc Advocate
Joined: 19 Apr 2002 Posts: 2805 Location: Oceanside, Ca
|
Posted: Thu Jun 01, 2006 3:20 am Post subject: |
|
|
Just re-run emwrap.sh with the same flags when the faild message pops up do "f" ENTER. It will then move the failed list to the build file and emerge those packages. As you know emerges can fail for bunchs of reasons. If they fail to build again then you can " cat /tmp/emwrap/failed and emerge them one at a time. This will give a better idea why they are failing. _________________ An A-Z Index of the Linux BASH command line |
|
Back to top |
|
|
zigver Tux's lil' helper
Joined: 09 May 2003 Posts: 87
|
Posted: Thu Jun 01, 2006 3:44 am Post subject: |
|
|
hielvc wrote: | Just re-run emwrap.sh with the same flags when the faild message pops up do "f" ENTER. It will then move the failed list to the build file and emerge those packages. As you know emerges can fail for bunchs of reasons. If they fail to build again then you can " cat /tmp/emwrap/failed and emerge them one at a time. This will give a better idea why they are failing. |
Yep, already ran the failed singly. But what is left to do when the script says "If you dont enter anything then the faild file will be deleted and emwrap.sh system -e will continue."? I believe it should say "emwrap.sh world -e will continue" but at this point shouldn't it just exit? |
|
Back to top |
|
|
kiboko n00b
Joined: 29 May 2003 Posts: 59
|
Posted: Tue Jun 13, 2006 4:08 am Post subject: problems with latest eselect/gcc-config |
|
|
I was trying to use emwrap.sh 4.0.5 for the first time on my AMD64 box after upgrading to GCC 4.1.1 and ran into a problem at the end of of a "emwrap.sh -Sb" where the changes to gcc-config and eselect confuse the script and cause it to abort early.
gcc-config -l now reports Quote: | :~# gcc-config -l
Available compilers for CTARGET i686-pc-linux-gnu
[1] x86_64-pc-linux-gnu-3.4.4/x86-hardened
[2] x86_64-pc-linux-gnu-3.4.4/x86-hardenednopie
[3] x86_64-pc-linux-gnu-3.4.4/x86-hardenednopiessp
[4] x86_64-pc-linux-gnu-3.4.4/x86-hardenednossp
[5] x86_64-pc-linux-gnu-3.4.4/x86-vanilla
[6] x86_64-pc-linux-gnu-3.4.6/x86-hardened
[7] x86_64-pc-linux-gnu-3.4.6/x86-hardenednopie
[8] x86_64-pc-linux-gnu-3.4.6/x86-hardenednopiessp
[9] x86_64-pc-linux-gnu-3.4.6/x86-hardenednossp
[10] x86_64-pc-linux-gnu-3.4.6/x86-vanilla
[11] x86_64-pc-linux-gnu-4.1.1/x86-vanilla
Available compilers for CTARGET x86_64-pc-linux-gnu
[12] x86_64-pc-linux-gnu-3.4.4/amd64-hardened
[13] x86_64-pc-linux-gnu-3.4.4/amd64-hardenednopie
[14] x86_64-pc-linux-gnu-3.4.4/amd64-hardenednopiessp
[15] x86_64-pc-linux-gnu-3.4.4/amd64-hardenednossp
[16] x86_64-pc-linux-gnu-3.4.4/amd64-vanilla
[17] x86_64-pc-linux-gnu-3.4.6/amd64-hardened
[18] x86_64-pc-linux-gnu-3.4.6/amd64-hardenednopie
[19] x86_64-pc-linux-gnu-3.4.6/amd64-hardenednopiessp
[20] x86_64-pc-linux-gnu-3.4.6/amd64-hardenednossp
[21] x86_64-pc-linux-gnu-3.4.6/amd64-vanilla
[22] x86_64-pc-linux-gnu-4.1.1/amd64-vanilla
Activated profiles:
i686-pc-linux-gnu x86_64-pc-linux-gnu-4.1.1/x86-vanilla
x86_64-pc-linux-gnu * x86_64-pc-linux-gnu-4.1.1/amd64-vanilla
|
When using gcc-config to set the profile I now have to use either the profile number, (e.g. 22) or the profile full name, (e.g. x86_64-pc-linux-gnu-4.1.1/amd64-vanilla). Using the CHOST name x86_64-pc-linux-gnu causes an error. emwrap.sh uses the CHOST derived profile name which now breaks gcc-config.
Also of note is the fact that gcc-config is now deprecated in favour of eselect complier. The error message: Code: | BEFORE selecting new GCC:
Available compilers for CTARGET i686-pc-linux-gnu
[1] x86_64-pc-linux-gnu-3.4.6/x86-hardened
[2] x86_64-pc-linux-gnu-3.4.6/x86-hardenednopie
[3] x86_64-pc-linux-gnu-3.4.6/x86-hardenednopiessp
[4] x86_64-pc-linux-gnu-3.4.6/x86-hardenednossp
[5] x86_64-pc-linux-gnu-3.4.6/x86-vanilla
[6] x86_64-pc-linux-gnu-4.1.1/x86-vanilla
Available compilers for CTARGET x86_64-pc-linux-gnu
[7] x86_64-pc-linux-gnu-3.4.6/amd64-hardened
[8] x86_64-pc-linux-gnu-3.4.6/amd64-hardenednopie
[9] x86_64-pc-linux-gnu-3.4.6/amd64-hardenednopiessp
[10] x86_64-pc-linux-gnu-3.4.6/amd64-hardenednossp
[11] x86_64-pc-linux-gnu-3.4.6/amd64-vanilla
[12] x86_64-pc-linux-gnu-4.1.1/amd64-vanilla
Activated profiles:
i686-pc-linux-gnu x86_64-pc-linux-gnu-4.1.1/x86-vanilla
x86_64-pc-linux-gnu * x86_64-pc-linux-gnu-4.1.1/amd64-vanilla
**********************************************************************
I'm going to execute: # gcc-config x86_64-pc-linux-gnu-4.1.1
**********************************************************************
Kill me NOW, or forever hold your peace!
(sleeping for 15 seconds...)
15 14 13 12 11 10 9 8 7 6 5 4 3 2 1
(continuing...)
* gcc-config is deprecated and is just a frontend to the compiler
* eselect module. In the future, gcc-config will be removed
* from portage. Please see 'eselect compiler help'
!!! Error: Invalid profile: x86_64-pc-linux-gnu-4.1.1
exiting.
ERROR: : gcc-config x86_64-pc-linux-gnu-4.1.1 failed!
|
Any possibilities on help with this?
Also on a seperate note, it would be nice if the emwrap.sh scripts uses of emerge included the --ignore-default-opts option. I could not get the emwrap.sh script to work correctly without removing the values for EMERGE_DEFAULT_OPTS from my make.conf file. Also if one has emerge aliased, the output from that emwrap.sh depends on could be different and cause the parse to break easily. I modifed my copy to fix this. The patch is as follows: Code: | *** dist/emwrap.sh 2006-05-31 14:47:55.000000000 -0600
--- /usr/local/bin/emwrap.sh 2006-06-12 16:30:17.000000000 -0600
***************
*** 456,462 ****
if [ ! -e build ] || [[ -e build && ! -s build ]];then
# added emerge to get the version specific info. It also sorts the tc packages.
:>tc.lst
! emerge -qp $TC_build |grep "/"|cut -c17-65|cut -f1 -d " " >tc.lst
for t in $(echo $tc_list); do
grep $t tc.lst>>build
done
--- 456,462 ----
if [ ! -e build ] || [[ -e build && ! -s build ]];then
# added emerge to get the version specific info. It also sorts the tc packages.
:>tc.lst
! \emerge --ignore-default-opts -qp $TC_build |grep "/"|cut -c17-65|cut -f1 -d " " >tc.lst
for t in $(echo $tc_list); do
grep $t tc.lst>>build
done
***************
*** 580,586 ****
# If for blocking regeneration of list if resume or do_tc is used
# Added exclude_filter to filters here. That way it works "W" 5-8-06
if [ "$resume" != "yes" ];then
! emerge $bclass -${eargs}pq >wrld.lst
# checking for blocked or masked packages
if grep "blocks" wrld.lst ; then
echo
--- 580,586 ----
# If for blocking regeneration of list if resume or do_tc is used
# Added exclude_filter to filters here. That way it works "W" 5-8-06
if [ "$resume" != "yes" ];then
! \emerge $bclass --ignore-default-opts -${eargs}pq >wrld.lst
# checking for blocked or masked packages
if grep "blocks" wrld.lst ; then
echo
|
|
|
Back to top |
|
|
hielvc Advocate
Joined: 19 Apr 2002 Posts: 2805 Location: Oceanside, Ca
|
Posted: Tue Jun 13, 2006 2:27 pm Post subject: |
|
|
kiboko I posted a "em_No_gccselect.sh" which removes the call to gcc-config . This is a nother reason why I held off from doing this I new the devs would change it. As to "--ignore-default-opts " I'll fix it in a day or 2, Im swamped todoy. _________________ An A-Z Index of the Linux BASH command line |
|
Back to top |
|
|
kiboko n00b
Joined: 29 May 2003 Posts: 59
|
Posted: Thu Jun 15, 2006 6:13 am Post subject: |
|
|
hielvc wrote: | kiboko I posted a "em_No_gccselect.sh" which removes the call to gcc-config . This is a nother reason why I held off from doing this I new the devs would change it. As to "--ignore-default-opts " I'll fix it in a day or 2, Im swamped todoy. |
em_No_gccselect.sh works better. Was able to complete a couple of -Sb runs. Thanks |
|
Back to top |
|
|
curmudgeon Veteran
Joined: 08 Aug 2003 Posts: 1741
|
Posted: Thu Jun 15, 2006 6:18 am Post subject: |
|
|
I have a couple of problems now.
First, I can't reach the site http://68.8.82.201:9002/ (and I have tried for days). It worked fine before.
So, of course, I tried copying the script on the first page of the thread. That didn't work either:
Code: | $ ./emnew.sh
./emnew.sh: line 685: syntax error: unexpected end of file |
It looks like some kind of quote mismatch (the latest version has a LOT more quotes in it).
I found a couple of small (not the source of the problem) typos.
Line 105 (part of the help) has an improper posessive ("switch's" instead of "switches")
Line 183 (basically a comment) mixes quote types:
Code: | ${GR} New if emwrap.sh is run and it detects a ${CY}"failed' or "build"${GR} file |
I couldn't find (or figure out) why I have that problem (I did try copying and pasting the file more than once).
Thank you for your help. |
|
Back to top |
|
|
StifflerStealth Retired Dev
Joined: 03 Jul 2002 Posts: 968
|
Posted: Thu Jun 15, 2006 10:48 am Post subject: |
|
|
I too have been having issues connecting to your server. Can you not just get a SourceForge account and put all the Downloads up there? You can even put stuff in CVS and have revision control. _________________ Nothing to read in this sig. Move along. |
|
Back to top |
|
|
hielvc Advocate
Joined: 19 Apr 2002 Posts: 2805 Location: Oceanside, Ca
|
Posted: Thu Jun 15, 2006 3:20 pm Post subject: |
|
|
Apparently my cable connection was upgraded. I gat a 754.4k/sec donwload speed this morning The down. Its now fixedside is that my address changed. Its now fixed.This has been really busy and remains so. I havent had a chance to really fix anything. Tonight Ill try to paste the no gcc-config to the front page. _________________ An A-Z Index of the Linux BASH command line |
|
Back to top |
|
|
Bob P Advocate
Joined: 20 Oct 2004 Posts: 3355 Location: Jackass! Development Labs
|
Posted: Thu Jun 15, 2006 5:08 pm Post subject: |
|
|
StifflerStealth wrote: | I too have been having issues connecting to your server. Can you not just get a SourceForge account and put all the Downloads up there? You can even put stuff in CVS and have revision control. |
Having rolled my own supported projects and having dealt with server problems and users that are eager to download, I'll throw my two-cent answer to the SourceForge question.
Its alot more convenient for users if a project is hosted on SourceForge. The bandwidth is great, the server redundancy makes downloads alot easier and the servers are almost never unavailable. One problem with this, though, is that when things become so easy from the user's perpective, they tend to lose all of the appreciation of how hard it is for an independent developer to keep his project up and running.
SourceForge isn't a universal solution for everyone. Getting a project hosted on SourceForge is not a trivial undertaking. A developer has to qualify for hosting on SourceForge according to their criteria. If the project meets the criteria for elegibility then the developer has to jump through more flaming hoops to get the project hosted on SourceForge. Then he has to make a long term commitment to guarantee that the project will continue under developement and continue to be supported on SourceForge. Asking a developer to move their project to Sourceforge, then, amounts to asking them to do an awful lot of work that might not even be necessary. For a project that involves a small download and only needs a thread in an existing forum for support, serving the files from your own private webspace can be an ideal solution.
If a developer of a small, private open source project doesn't already have their project hosted on SourceForge, there's probably a reason for it. Some projects are one-man-bands and don't have the resources to qualify for SourceForge hosting. Some projects don't have the development staff or the extra time to support SourceForge hosting. Some projects are hosted on a home-based server because its cool to own your own server and to run stats on the server logs. Asking developers to change the way that they do things just to make downloads easier ends up making their job harder. If a project's server problems are getting in your way as a downloader, instead of asking someone to change the way that they do things to make your life easier, maybe it would be good to consider doing something to make their life easier. If you really like their project, you could volunteer to help them or you could just send them a donation.
I don't mean to seem like I'm chastising anyone. That's not my intent. but when someone says "Could you not just get a sourceforge account," that makes it sound as if you consider the task of getting a sourceforge account to be as trivial as turning on a light switch. In reality, its much harder than that and it involves a long term commitment that may not be suitable for everyone.
Just my $0.02. _________________ .
Stage 1/3 | Jackass! | Rockhopper! | Thanks | Google Sucks |
|
Back to top |
|
|
dol-sen Retired Dev
Joined: 30 Jun 2002 Posts: 2805 Location: Richmond, BC, Canada
|
Posted: Fri Jun 16, 2006 4:34 am Post subject: |
|
|
Perhaps if you find a good, but small project, and have a little space you could mirror it on your webspace to help spread the load a little.
I did for McQuires earlier re-write of emwrap. It has not been kept up with advancements that hielvc has been making. If I can get more time again. I'll clean up a few things for some space and offer to mirror your current releases. It is on my ISP's personal webspace so should have reasonable bandwidth. _________________ Brian
Porthole, the Portage GUI frontend irc@freenode: #gentoo-guis, #porthole, Blog
layman, gentoolkit, CoreBuilder, esearch... |
|
Back to top |
|
|
skutnar n00b
Joined: 13 Jun 2006 Posts: 65
|
Posted: Sat Jun 17, 2006 3:40 pm Post subject: |
|
|
hielvc wrote: | Apparently my cable connection was upgraded. I gat a 754.4k/sec donwload speed this morning The down. Its now fixedside is that my address changed. Its now fixed.This has been really busy and remains so. I havent had a chance to really fix anything. Tonight Ill try to paste the no gcc-config to the front page. |
How about a dynamic DNS service? |
|
Back to top |
|
|
MoonWalker Guru
Joined: 04 Jul 2002 Posts: 510
|
Posted: Sun Jun 18, 2006 7:52 pm Post subject: |
|
|
hielvc,
I think you need to fix this in your emwrap.sh
After first emerging gcc-4.1.1 once, I executed Code: | # eselect compiler list
#eselect compiler set 6 (number of my gcc-4.1.1 profile)
# env-update && source /etc/profile
#emwrap.sh -seb1 |
and after completion of gcc-4.1.1 compilation I got this
Code: | >>> /usr/i586-pc-linux-gnu/gcc-bin/4.1.1/c++ -> i586-pc-linux-gnu-c++
>>> /usr/i586-pc-linux-gnu/gcc-bin/4.1.1/g++ -> i586-pc-linux-gnu-g++
>>> /usr/i586-pc-linux-gnu/gcc-bin/4.1.1/cpp -> i586-pc-linux-gnu-cpp
>>> /usr/i586-pc-linux-gnu/gcc-bin/4.1.1/gcc -> i586-pc-linux-gnu-gcc
>>> /usr/i586-pc-linux-gnu/gcc-bin/4.1.1/i586-pc-linux-gnu-gcc-4.1.1 -> i586-pc-linux-gnu-gcc
>>> /usr/i586-pc-linux-gnu/gcc-bin/4.1.1/gfortran -> i586-pc-linux-gnu-gfortran
* The following compiler profile has been activated based on your previous profile:
* i586-pc-linux-gnu-4.1.1/vanilla
* If you have issues with packages unable to locate libstdc++.la,
* then try running 'fix_libtool_files.sh' on the old gcc versions.
>>> No packages selected for removal by clean.
>>> Auto-cleaning packages...
>>> No outdated packages were found on your system.
* Regenerating GNU info directory index...
* Processed 6 info files.
BEFORE selecting new GCC:
Available compilers for CTARGET i586-pc-linux-gnu
[1] i586-pc-linux-gnu-3.4.6/hardened
[2] i586-pc-linux-gnu-3.4.6/hardenednopie
[3] i586-pc-linux-gnu-3.4.6/hardenednopiessp
[4] i586-pc-linux-gnu-3.4.6/hardenednossp
[5] i586-pc-linux-gnu-3.4.6/vanilla
[6] i586-pc-linux-gnu-4.1.1/vanilla
Activated profiles:
i586-pc-linux-gnu * i586-pc-linux-gnu-4.1.1/vanilla
**********************************************************************
I'm going to execute: # gcc-config i586-pc-linux-gnu-4.1.1
**********************************************************************
Kill me NOW, or forever hold your peace!
(sleeping for 15 seconds...)
15 14 13 12 11 10 9 8 7 6 5 4 3 2 1
(continuing...)
* gcc-config is deprecated and is just a frontend to the compiler
* eselect module. In the future, gcc-config will be removed
* from portage. Please see 'eselect compiler help'
!!! Error: Invalid profile: i586-pc-linux-gnu-4.1.1
exiting.
ERROR: : gcc-config i586-pc-linux-gnu-4.1.1 failed!
|
Actually, I tried once first to use gcc-config this way and it didn't like when feeding it the profile string but like to have the diggit. But anyhow your script need to be updated to use eselect.
I haven't done it yet but assume setting the gcc-config again properly with eselect and then run same emwrap command will resume the operation? Well I will see otherwise I guess I can just kill it and run just the -e system excluding TC.
I think it's a very nice script otherwise
EDIT: Just another note, running the same emwrap.sh -seb1 again gave me the option to resume, nice, but as you can see from my emerge.log below it starts to emerge gcc ahgain although it has completed succesfully. The error comes with the gcc-config usage after gcc is compiled.
Code: | 1150642299: Started emerge on: Jun 18, 2006 14:51:39
1150642299: *** emerge --oneshot --nodeps --quiet =sys-devel/gcc-4.1.1
1150642301: >>> emerge (1 of 1) sys-devel/gcc-4.1.1 to /
1150642301: === (1 of 1) Cleaning (sys-devel/gcc-4.1.1::/usr/portage/sys-devel/gcc/gcc-4.1.1.ebuild)
1150642303: === (1 of 1) Compiling/Merging (sys-devel/gcc-4.1.1::/usr/portage/sys-devel/gcc/gcc-4.1.1.ebuild)
1150658146: >>> AUTOCLEAN: sys-devel/gcc
1150658146: --- AUTOCLEAN: Nothing unmerged.
1150658146: === (1 of 1) Post-Build Cleaning (sys-devel/gcc-4.1.1::/usr/portage/sys-devel/gcc/gcc-4.1.1.ebuild)
1150658146: ::: completed emerge (1 of 1) sys-devel/gcc-4.1.1 to /
1150658146: *** Finished. Cleaning up...
1150658153: *** exiting successfully.
1150658161: *** terminating.
1150659953: Started emerge on: Jun 18, 2006 19:45:53
1150659953: *** emerge --oneshot --nodeps --quiet =sys-devel/gcc-4.1.1
1150659959: >>> emerge (1 of 1) sys-devel/gcc-4.1.1 to /
|
have to wait a few hours to see if this is a "mary goes around in circles" poping up the same error again...
EDIT2: It complete the resume w/o borking on gcc-config but then it never do the system -e files, I understand it as they are included in -seb1?
Anyhow I ran it from start again as building the TC a second time can't hurt, except for the time it takes but now again it bork on the gcc-config stuff. I wont resume it this time as no point in building gcc a 4th time, just do the libstdc++-v3-3.3.6 manually and then see if I can get it started on just doing the system w/o TC.
Just a final remark on this, the script do seem to go around in an endless loop borking on gcc-config. If that has to do with me running i586 (2.4 kernel) I don't know but the gcc-config is depreciated anyhow and oght to be fixed. _________________ /Joakim
Living on earth is expensive, but it includes a free trip around the sun
every year.
Last edited by MoonWalker on Mon Jun 19, 2006 12:14 pm; edited 1 time in total |
|
Back to top |
|
|
zajelo3 n00b
Joined: 05 Sep 2005 Posts: 60
|
Posted: Mon Jun 19, 2006 2:54 am Post subject: |
|
|
When doing an "emwrap.sh -set" am I supposed to do that once or twice? The output of -setp shows it building some of the TC twice but not all of it. Just figured I'd ask to be sure. |
|
Back to top |
|
|
dol-sen Retired Dev
Joined: 30 Jun 2002 Posts: 2805 Location: Richmond, BC, Canada
|
Posted: Tue Jun 20, 2006 3:24 am Post subject: |
|
|
Once should be enough. As long as the packages are emerged in a specific order (emwrap.sh does already) and you are not having any real, nasty problems with your system. _________________ Brian
Porthole, the Portage GUI frontend irc@freenode: #gentoo-guis, #porthole, Blog
layman, gentoolkit, CoreBuilder, esearch... |
|
Back to top |
|
|
hielvc Advocate
Joined: 19 Apr 2002 Posts: 2805 Location: Oceanside, Ca
|
Posted: Tue Jun 20, 2006 3:08 pm Post subject: |
|
|
Sorry for my lack of responding but I have been hammered at work. After 14hour days my brain just goes on strike. HOoopefully in the next 2 or 3 days things will slow down. _________________ An A-Z Index of the Linux BASH command line |
|
Back to top |
|
|
mauricev Apprentice
Joined: 22 Mar 2004 Posts: 198
|
Posted: Tue Jun 20, 2006 9:37 pm Post subject: |
|
|
Quote: | Portage doesn't know about all the optional compile-time links. So it's not good enough for ensuring that you end up with all the app links that you originally had. To form all the links, another emerge -e world would be necessary. |
Can someone please explain what is meant by the "optional compile-time" links?
Also, why does the upgrade guide at http://www.gentoo.org/doc/en/gcc-upgrading.xml#upgrade-3.3-to-3.4 ask only for just one rebuild of the system and world? It's not even asking for the toolchain! |
|
Back to top |
|
|
kovert n00b
Joined: 23 Jun 2006 Posts: 2
|
Posted: Fri Jun 23, 2006 9:27 am Post subject: |
|
|
Uh...
So if the kernel headers are a different version from the actual kernel you are using what does that mean?
uname -a = 2.6.16-r9
linux headers = 2.6.11-r2
Forgive me...I'm a n00b! HA HAHA...HA |
|
Back to top |
|
|
StifflerStealth Retired Dev
Joined: 03 Jul 2002 Posts: 968
|
Posted: Fri Jun 23, 2006 10:09 am Post subject: |
|
|
kovert wrote: | Uh...
So if the kernel headers are a different version from the actual kernel you are using what does that mean?
uname -a = 2.6.16-r9
linux headers = 2.6.11-r2
Forgive me...I'm a n00b! HA HAHA...HA | Yes. The kernel headers can be a different version. It has been tested that everything compiles fine against those version of the headers. Also, you need to do a lot of recompiling when you update your headers, so it's just better to update them every once in a while.
Cheers. _________________ Nothing to read in this sig. Move along. |
|
Back to top |
|
|
kovert n00b
Joined: 23 Jun 2006 Posts: 2
|
Posted: Fri Jun 23, 2006 10:32 am Post subject: |
|
|
StifflerStealth wrote: |
Yes. The kernel headers can be a different version. It has been tested that everything compiles fine against those version of the headers. Also, you need to do a lot of recompiling when you update your headers, so it's just better to update them every once in a while.
Cheers. |
Thanks for the quick response...I can at least start rebuilding before I go to sleep. I thought at first this might have been an error and I couldn't find an answer if I was required to have the same kernel headers as the one I was running. |
|
Back to top |
|
|
mauricev Apprentice
Joined: 22 Mar 2004 Posts: 198
|
Posted: Fri Jun 23, 2006 3:18 pm Post subject: |
|
|
Quote: | Yes. The kernel headers can be a different version. |
This has never made any sense to me. Why is there even a package for the kernel headers? The kernel configured in /usr/src/linux has the headers in /usr/src/linux/include/linux, so why have two copies of them and to top it off it is never the version I am using as the running kernel.
I wonder what happen if one just removed the kernel header package, feigned its presence in portage and just symlinked /usr/src/include/linux to the "real" headers? |
|
Back to top |
|
|
ck42 l33t
Joined: 31 Jul 2003 Posts: 789
|
Posted: Sat Jun 24, 2006 5:11 pm Post subject: |
|
|
Does the ordering of the options for the emwrap command not have an effect?
example: if I choose -Sb1, it starts off by emerging the TC files. Since I'm having the same issue with the gcc-config/eselect deprecation thing, after the error, should I:
manually select the new gcc version via eselect compile set
manually sourece /etc/profile
manually emerge libstdv3++
run emwrap.sh -S
to get the same effect as if the error never occured? Is there another gcc build in that sequence somewhere that should occur? |
|
Back to top |
|
|
hielvc Advocate
Joined: 19 Apr 2002 Posts: 2805 Location: Oceanside, Ca
|
Posted: Sun Jun 25, 2006 12:29 am Post subject: |
|
|
I moved em_noselect_gcc.sh to emwrap.sh. When you run it with no args it will annouce its self as "emwrap.sh no gcc select v. 4.0.5 "The fuzzy tailed rats are winning, Elp"". Youll not get the gcc-config error.
TimeStamp Jun 24 16:35 emwrap.sh
ck42 no I desinged emwrap.sh to not be finicky about the flags and thier order with one exception "exclude" or -X. They have to be the last flags, switch, with its args following. _________________ An A-Z Index of the Linux BASH command line |
|
Back to top |
|
|
|