Re: Many bugs in latest version

[Maxim Levitsky <maximlevitsky gmail com>]

On Mon, 2009-10-05 at 01:59 +0200, Maxim Levitsky wrote: 
> On Mon, 2009-10-05 at 01:27 +0200, Maxim Levitsky wrote: 
> > On Sat, 2009-10-03 at 14:26 +0200, Philippe Rouquier wrote:
> > > Le lundi 28 septembre 2009 à 23:32 +0200, Maxim Levitsky a écrit : 
> > > > Hi, I did another round of testing.
> > > > Burn settings are shown correctly now when adding 2nd session.
> > > > 
> > > > This time I have used libburn for all the work.
> > > > To be sure, I removed wodim, cddao, dvd+rw-tools, and genisoimage.
> > > > 
> > > > Here are problems I found, mostly libburn/libiso bugs:
> > > > 
> > > > 1 - disk label (normal) is written in all caps. Joiliet label is written
> > > > normally but can be truncated due to size limit. currently, latest -git
> > > > of util-linux-ng will use normal label only if its first 16 chars are
> > > > same as in joiliet label.
> > > 
> > > Yes, they stick closely to the standard specifications. Perhaps there
> > > is an options to relax that ..
> > 
> > Yes, and the attached patch fixed that issue.
> > 
> > 
> > > 
> > > > 2 - next sessions use always the label from 1st session (normal label)
> > > > thus now both joilet and normal label are same, but uppercase
> > > > 
> > > This is annoying, I'll have a look at it.
> > 
> > Fixed, thanks!
> > 
> > > > 
> > > > 3 - writing CD-RW with simulation fails (drive doesn't want to close
> > > > track or so, log (attached) brasero-session-1.log
> > > > 
> > > > 
> > > Thanks I'll take a look.
> > Now is disabled, not a big deal.
> > 
> > 
> > > > 4 - audio recording fails (tried on two disks)
> > > >    first track is written normally, then brasero begins to write 2nd
> > > > one, and then fails (log brasero-session-2.log)
> > > > This isn't devkit bug, I stopped it, (and removed its dbus startup
> > > > file), to be sure.
> > > > 
> > > This one is really annoying. I can reproduce but not fix and I'm
> > > starting to wonder if it is not in libbrasero burn itself. It only
> > > happens with TAO not DAO and seems to be linked to the fact that after
> > > writing the first track libburn asks the next writable sectors which
> > > seems to return 1 (the beginning of the session) which has been
> > > already written by the first track. But I may be wrong here.
> > 
> > I am sure I didn't use the simulation mode. However now works just fine.
> > 
> > Few minor nitpicks:
> > * I am allowed to set disk label, but it has no place on audio disk
> > (maybe cd-text?)
> > 
> > * I am allowed to 'leave disk open', however brasero (at least with
> > libburnia backend), doesn't support importing audio sessions.
> > 
> > 
> > > 
> > > > 
> > > > 5 - not sure if this is bug, but adding new session on DVD+RW, shows for
> > > > some time that it is 'blanking' the disk, although no information is
> > > > erased. Does it actually blank new area it will write, can you explain?
> > > > 
> > > > 
> > > It's sort of a bug. It turns out that DVD+RW  need a small formatting
> > > (which should concern only the sectors that are going to be written)
> > > before writing. Since in libburn plugin we listen to the drive
> > > actions, brasero reported blanking (which is an oversimplified way to
> > > say it's formatting). Anyway, I've added some context and now brasero
> > > should say 'start recording' which is far less confusing.
> > Fixed as well, thanks!
> > 
> > 
> > 
> > 
> > Well, here you see the results of my checklist:
> > 
> > 
> > * CD-RW:blank disk 
> >      - success
> > 
> > * CD-RW:write first session
> >      - success
> > 
> > * CD-RW:write second session
> >      - tested with different label
> >      - success
> >      
> > * CD-RW:write third session (close the disk + simulation)
> >      - for addition testing, unchecked all options in burn setting
> > dialog
> >      - success
> > 
> > * CD-RW:write 5 sound tracks
> >      - did see a warning for disk blank (as expected) + warning about
> > CDRW (nice)
> >      - success
> > 
> > 
> > 
> > * DVD+RW:blank
> >      - success (didn't see progress bar though)
> > 
> > * DVD+RW:write 1 session
> >      - success
> > 
> > * DVD+RW:write 2 session
> >      - tested with different label
> >      - success
> > 
> > 
> > Just couldn't find any problem.
> > And this is using only the libburn and libisofs!
> 
> I addition to that, I have just added a session to DVD-R, and everything
> went smooth.

And futhermore, I have reverted the patch that disabled cd simulation,
and it does work now (tested 2 sessions)

I did test simulation on DVD-R too.


In fact everything seems to be prefect.

Of course except that brasero still is very often confused after an
operation (once it even insisted on blanking the disk, while I added new
session...), gnome/devkit/gdu continues the try to steal that poor disk
from brasero (how many 'error' messages I have seen about 'unable to
mount blank disk...' it is especially funny to see that error message,
past a successful burn ....) there are similar messages of that kind,
like unable to mount disk, its busy, etc... in other words lets them fix
that mess.  Maybe best solution is to add an api to tell devkit and
friends to leave the disk alone, and use that api in brasero, don't
know.

Best regards,
Thanks,
Maxim Levitsky