GNOME Bugzilla – Bug 543377
No multisession
Last modified: 2009-06-23 10:44:56 UTC
Please describe the problem: "Import session" is always inactive, doesn't matter if disc was recorded with brasero or other application. Version 0.8 compiled in Debian Lenny Steps to reproduce: 1. 2. 3. Actual results: Expected results: Does this happen every time? Other information:
Created attachment 114701 [details] output from shell
It concerns only DVDs, i could write session on CD but only two, after third i got message "volume is unknown"
Thanks for the report. Could you please: - attach a log for the CD you can't add/import any more sessions - try SVN trunk where I committed a change to workaround your problem Thanks
Created attachment 114902 [details] session1
Created attachment 114903 [details] session2
Created attachment 114904 [details] session3
I can't try svn version because i have access to net only using windows. Could you please tell me which files should i download and replace in 8.0 version
SVN didn't remove the problem, Brasero still can't see previous sessions
Here is the current state of my investigations: the problem seems to come from your using libburn. libburn uses an address to start writing the start of a session which differs by + or - 150 sectors from the one used by brasero. So if you want, you can try using wodim or cdrecord and turning off libburn for the moment. To do that put -1 in the gconf key /app/brasero/config/priority/libburn-burn. That said, how hard as I tried, I couldn't reproduce the bug with my version of libburn (home compiled 0.5.0) so my question is what version of libburn do you have please?
I'm afraid it's not libburn(0.4.8), using wodim it's the same . It's strange that you can't reproduce this problem, i can do it with my two different drives pioneer and old liteon
I committed a fix to trunk. Could you test it please? NOTE that you must erase your medium (you seem to use a CDRW) before starting again for testing. The error/discrepancy about the next writable address I mentionned above messed your medium.
Created attachment 116026 [details] log-libburn
Created attachment 116027 [details] log-wodim
It's still the same
This time I'm pretty sure I got it. The problem came from the fact that a drive must be in TAO mode before retrieving NWA. That isn't true for all drives since one of mines didn't but I have another that does and it seems the change I made to trunk worked for it. I'm closing this bug. Of course, should this not work, please let me know and reopen the bug. Thanks for your help.
Now you will think i'm spiteful :) but did you try DVD+RW? For me it's still not working and i encountered some problem with cd-rw when verifying is on, i will make some more tests. While talking about verifying, why don't you let the user choose whether he wants to do it or not? It would be more convenient to have it as an option to check/uncheck under "windows comatibility" and "multisession"
Before i was testing multisession by starting and quiting Brasero with disc present in drive. Now i left Brasero started and ejected/inserted cd. 1.insert blank cdrw 2.record something 3.empty project 4.eject cd 5.insert cd 6.import session 7.record something 8.message: disc could not be mounted (max attemps reached)- brasero-session.log 9.empty project 10.import session 11.crash - brasero-bugreport.txt
Created attachment 116476 [details] brasero-bugreport
Created attachment 116477 [details] brasero-session
Created attachment 116478 [details] logcdrw
I think I committed a change to trunk that may fix your problem (the crash) as well as some other problems. Please test =).
I also increased the number of attempts at mounting that should help.
btw, last but not least could retry with DVD+RW. Retry to first burn a first session with brasero on one and then see if the import button works when it is re-inserted. Also please attach logs (as usual) for the whole process. There is something strange in your first log (#1) since brasero sees your DVD+RW as a blank one since it can't find any file system on it.
Import button is inactive
Created attachment 117119 [details] logdvdgrowiso
Created attachment 117120 [details] logdvdlibburn
Thanks for the logs. I committed some changes to trunk. Please test and report. In case of failure, you just need to attach a log with the begining. The problem is in brasero which doesn't seem to get along with your drive while detecting it.
I added another fix that should improve things. I used another SCSI function to access the medium since your drive didn't like the one that was originally used.
Now after recording is finished brasero disappears and i get some message about memory protection violation. I think there's nothing interesting in log but i provide it anyway. Please tell me what is the way to give you debug information about it
Created attachment 117501 [details] logdvd
Thanks for your patience. The best would be to get a stack trace by running brasero inside gdb: - gdb brasero - in the new gdb console type "run" to start brasero - CRASH - in gdb console type "thread apply all bt" - cut and paste the backtrace to a file and attach it. NOTE: it could also be helpful to have the debug packages installed for gtk2 and glib2. According to the distros they should be named like "glib2-debug" and "gtk2-debug". Thanks in advance.
Created attachment 117564 [details] logdvd I removed messages about missing icons and powermanager to make it smaller. Brasero started from gdb doesn't crash but hangs.
Thanks a lot. Your crash should be fixed by now in the SVN.
Btw, I added some more debugging to improve your next log.
Strange thing, it worked for a while, i could write some sessions until i placed some files in an existing folder on dvd. Then some error occurred and i could'n even start brasero (i provide log from gdb). Now i don't know whether it is problem of brasero or of my system. Another thing is that brasero don't warn clearly about too deep structure of catalogues but says that unknown error occurred and information about it can be found in message that appears when you click view or save.
Created attachment 117824 [details] logdvdcrash
So basically import session for DVD+RW is working again right? If so, that's a progress. You seem to have uncovered another deeper bug but completely unrelated. The problem you have is with project loading that gets stupidly stuck at one point when automatically loading your project. Please could you attach ~/.config/brasero/brasero-tmp-project to the bug after brasero crashed please? To unlock brasero do "rm ~/.config/brasero/brasero-tmp-project -f"
Yes, import is working but there is a problem i tried to put some folder and some files into existing one on dvd
Created attachment 117900 [details] logdvd
Created attachment 117901 [details] brasero-tmp-project
Created attachment 117902 [details] brasero-session
Thanks for the help and log, I fixed the crash and imported multisessioned projects for good in SVN trunk. NOTE: brasero doesn't import the session from the disc when you load a project you still have to do it with the button but it should safely import it. Please test it and tell me if all works well.
Crash is fixed now, import works. The only thing is that i have written above, when i added files into existing folder got warning that some parent directory could not be found, log is 117902.
Thanks again for all this new bug. The bug was triggered by the use of libisofs (though libisofs was not itself the culprit). It's fixed in SVN trunk. Not closing the bug, in case you find another one ;).
And i foud it. First: if you insert multisession cd and will not import data but add some new, brasero will not erase it before writing. Second: it does not warn when directories' structure is too deep, you have to find it in log which is called unknowned error. I was afraid that my mission has finished but luckily no :)
I fixed the multisession bug and the lack of warning in SVN trunk. But I guess you're probably going to run into a problem with deep nested directory again. Indeed, even if I had no warning about a directory being too deep as well, I didn't have the failure while burning. So it seems I fixed half the problem here. Could you send me a log please?
Now i got a warning but still couldn't start burning
Created attachment 118725 [details] brasero-session.log
Created attachment 118726 [details] logcd
Sorry for the delay. I think it's now fixed in SVN trunk. Thanks for your patience. I'm waiting for the next bug report ;) before closing the bug.
This time something about wrong address
Created attachment 120017 [details] brasero-session.log
Thanks a lot for your help. I committed a fix to trunk. The libburn plugin was using a function only designed for DVD+RW and such.
I couldn't do any tests because Brasero doesn't detect dicsc
Created attachment 121630 [details] log
Thanks. That's strange, HAL doesn't return your drive in the list of available drives. So brasero doesn't even try to probe it. Have you done any system upgrade recently? Does your drive appear in applications using nautilus-cd-burner like say sound-juicer? What's your version of HAL?
Hal is 0.5.11, i did upgrade but sound juicer works and what is more strange Brasero 0.8.0 (Debian package)works too
Thanks. Same here for hal version and since sound-juicer sees your drive then it lets me think that the problem is in brasero. I added some more debugging to catch the problem. Could you attach a new log with latest SVN version please. Last note, don't close brasero immediatly as I recently changed its behavior regarding drive/medium initialization. Now brasero probes them asynchronously which allows for a faster start and also avoids locks up when a medium is damaged or when the drive is in a bad state. So even if brasero is loaded it may not mean that your drives and media are fully probed (it can take a few seconds sometimes).
Hal works now. So we can go back to deeply nested directories problem. I couldn't find exact way to reproduce this problem, but i thik it happens when i added some directories to compilation and then some other ones into them
Created attachment 121993 [details] brasero-session
Created attachment 121994 [details] log
I found the problem; you had a deep directory nesting in the session you imported and you added a file/a directory beyond the depth of 6 and brasero failed to detect that. I fixed the problem in trunk. Thanks a lot.
There are still some errors sometimes. I didn't get any message about it but it didn't finish verification, the window with progress bar just disappeared. DVDs are not burned at all, only sums are created
Created attachment 122847 [details] logcd
Created attachment 122848 [details] logdvd
Again I don't know how to thank you. I fixed the problem with DVD+RW and libburn. I also fixed the problem with the error message not appearing (recent change that had a mistake in it on error). I'll check the rest. Thanks again.
I have forgotten to say that there is a roblem with multisession when folders are nested and there are many sessions. There is another strange thing at the eng of writing, look at the picture. It is changing to 100% and proper size after the burning is finished.
Created attachment 122927 [details] screenshot
Tomasz: thanks. As for the size thing, it should have been fixed some time ago (I know it's been a long time since I last worked on this bug; sorry but I had other things to do first). I can't see it anymore with trunk at least. Many changes have taken place in the tree and I was wondering if you could test if everything works fine (if you happen to have some multisession discs to burn). Btw if you use a version > 0.9.1 you now need to run brasero like that: "brasero -g --brasero-media-debug > log 2>&1". Thanks.
Hey Tomasz, is there any chance you can test this with a more recent brasero? If not, I think closing this bug report as INCOMPLETE would be the correct solution then. Philippe: Isn't there another bug tracking multisession? And isn't there a GSoC Project going on? :)