| View previous topic :: View next topic |
| Author |
Message |
franzf
Advocate
Joined: 29 Mar 2005
Posts: 4565
|
 Posted: Mon Nov 21, 2011 4:05 pm Post subject: paludis can't read resume-file/stuck in cave resolve -e |
 |
|
Hi,
I did a GCC-Update yesterday (4.4.6->4.6.2), running cave resolve -e installed-packages after that, to rebuild all packages with the new gcc.
As it should be, some packages failed - no problem, unimportant ones. I mounted /var/tmp/paludis into RAM, to speed up compilation. Unforunately, /var/tmp/paludis went full, resulting in failures.
So I stopped paludis (ctrl+C), unmounted /var/tmp/paludis, did cave resume -Rr - and now I get this:
| Quote: | | * Eek! Internal error at paludis::Deserialisation::Deserialisation(const string&, paludis::Deserialiser&) at serialise.cc:207: can't parse string (paludis::InternalError) |
YEAH! I managed to build 829 packages out of 1129 successfully, and now I can't resume! I did a backup, but that has it's last finished build @247...
Actually, now all the interesting packages would have started. thunderbird, firefox, kde, ...
One full day compiling for nothing?
Or is there a way to get a working resume-file?
The resume-file has ~24MB (The broken one has 10...), editing manually is a mess, I would do it, if it was understandable at least 
Is there someone out, who could help me? I won't start from scratch, today, nor tomorrow - completely lost interest in rebuilding ATM
Thx in advance
Franz |
|
| Back to top |
|
 |
yoshi314
l33t
Joined: 30 Dec 2004
Posts: 850
Location: PL
|
|
| Back to top |
|
|
franzf
Advocate
Joined: 29 Mar 2005
Posts: 4565
|
| Posted: Tue Nov 22, 2011 3:11 pm Post subject: |
|
|
| yoshi314 wrote: | | maybe you should ask on #paludis on freenode, or on its mailing list. |
Thx for the tip (I haven't been on IRC since years, so this idea did not pop in...)
I think it is not possible to recover from the broken resume-file. Several packages I know could not be found by grep, If the list is not complete, the recovery is quite useless.
In the meantime, have managed to start from the point of failure, though.
I had the back up. I resumed from that, stopped after the initial state paludis prints (succesfuil built, failed, skipped, ...) and copy-pasted all lines from the last succesful package on downwards, saved it to a file and did the following command:
| Code: | | sed 's/skipped: install /=/g' cave_list_raw | sed 's/ to ::installed replacing [0-9a-z_.-]*//' > cave_packages |
I'm not that familiar with that kind of file-editing, so maybe there is a better way, but it works 
I just needed to do
| Code: | | cave resolve $(cat cave_packages) |
and now the rest of the packages get merged
No real solution, if you did not back up your resume-file... |
|
| Back to top |
|
|
yoshi314
l33t
Joined: 30 Dec 2004
Posts: 850
Location: PL
|
| Posted: Wed Nov 23, 2011 1:28 pm Post subject: |
|
|
i suppose it's not possible this way.
you might want to create a resume command, and simply cut it at an appropriate point based on package build logs in /var/log/paludis when it fails again.
does in ram building speed up compilation that much?
_________________
~amd64
shrink your /usr/portage with squashfs+aufs |
|
| Back to top |
|
|
|
|
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
|
|
Unsupported Software
