Creation of fedora CD install.

[Joel Andres Granados]

Hi list:

I wanted to use the configuration values that express the amount of cds 
that pungi should generate and their related size.  When I executed 
pungi with "discs =2" and "cdsize=700", pungi threw a traceback that had 
something to do with not finding a directory.  On further investigation 
I found out that it was anaconda-runtime's splittree.py script that was 
causing the traceback.  It created the full directory structure for the 
first cd but did not do the same for the cds after that.
I noticed that the addition of one line "fixed" it :

--- /tmp/splittree.py   2007-04-23 12:50:23.000000000 -0400
+++ splittree.py        2007-04-23 12:51:35.000000000 -0400
@@ -228,6 +227,7 @@
                        pass

else:
+                os.makedirs("%s-disc%d/%s" % (self.dist_dir, i, 
self.product_path))
                self.linkFiles(self.dist_dir, "%s-disc%d" 
%(self.dist_dir, i), self.common_files)
            self.createDiscInfo(i)

I also noticed that this line was recently removed from the splittree.py 
file and really don't understand why, since the idea is for pungi to 
create cd and dvd.  Will the process to create CD isos change??  In any 
case, the addition of this line to splittree.py fixes the traceback I 
was getting.

Regards.

[Jesse Keating]

On Tuesday 24 April 2007 11:17:48 Joel Andres Granados wrote:
> --- /tmp/splittree.py   2007-04-23 12:50:23.000000000 -0400
> +++ splittree.py        2007-04-23 12:51:35.000000000 -0400
> @@ -228,6 +227,7 @@
>                         pass
>
> else:
> os.makedirs("%s-disc%d/%s" % (self.dist_dir, i,
>
> self.product_path))
>                 self.linkFiles(self.dist_dir, "%s-disc%d"
> %(self.dist_dir, i), self.common_files)
>             self.createDiscInfo(i)
>          
> I also noticed that this line was recently removed from the splittree.py
> file and really don't understand why, since the idea is for pungi to
> create cd and dvd.  Will the process to create CD isos change??  In any
> case, the addition of this line to splittree.py fixes the traceback I
> was getting.

Hrm, good to know, but that actually needs to be sent to 
[e-mail] and fixed there.

[Joel Andres Granados]

Jesse Keating wrote:
> On Tuesday 24 April 2007 11:17:48 Joel Andres Granados wrote:
> > --- /tmp/splittree.py   2007-04-23 12:50:23.000000000 -0400
> > +++ splittree.py        2007-04-23 12:51:35.000000000 -0400
> > @@ -228,6 +227,7 @@
> >                         pass
> >         else:
> >            os.makedirs("%s-disc%d/%s" % (self.dist_dir, i,
> >
> > self.product_path))
> >                 self.linkFiles(self.dist_dir, "%s-disc%d"
> > %(self.dist_dir, i), self.common_files)
> >             self.createDiscInfo(i)
> >
> > I also noticed that this line was recently removed from the splittree.py
> > file and really don't understand why, since the idea is for pungi to
> > create cd and dvd.  Will the process to create CD isos change??  In any
> > case, the addition of this line to splittree.py fixes the traceback I
> > was getting.
>
> Hrm, good to know, but that actually needs to be sent to 
> [e-mail] and fixed there.
>
> --
> Fedora-buildsys-list mailing list
> [e-mail]
> https://www.redhat.com/mailman/listinfo/fedora-buildsys-list

Sorry, I didn't make myself clear enough.  I checked the anaconda 
changelog and saw the following lines:

830     * textw/upgrade_text.py (UpgradeSwapWindow.__call__): Likewise.
  831
  832     * scripts/splittree.py: Don't generate RPMS/ dir (Jesse Keating)
  833
  834     * iw/progress_gui.py (InstallProgressWindow.setShowPercentage):

I thought that the missing line was part of some new feature that was 
not yet complete.  This is the reason why I posted it here, but I guess 
anaconda guys took away the whole line when they should have just erased 
the RPMS part of the command.  I'll post in on the anaconda list and see 
what they have to say.
Regards.

[Jesse Keating]

On Wednesday 25 April 2007 05:24:58 Joel Andres Granados wrote:
> I thought that the missing line was part of some new feature that was
> not yet complete.  This is the reason why I posted it here, but I guess
> anaconda guys took away the whole line when they should have just erased
> the RPMS part of the command.  I'll post in on the anaconda list and see
> what they have to say.
> Regards.

Oh dear, that may have been _me_ that broke it with an improper patch.  
Whoops.  I'll follow up with anaconda folks and fix that up hopefully 
tomorrow.

[Joel Andres Granados]

Jesse Keating wrote:
> On Wednesday 25 April 2007 05:24:58 Joel Andres Granados wrote:
> > I thought that the missing line was part of some new feature that was
> > not yet complete.  This is the reason why I posted it here, but I guess
> > anaconda guys took away the whole line when they should have just erased
> > the RPMS part of the command.  I'll post in on the anaconda list and see
> > what they have to say.
> > Regards.
>
> Oh dear, that may have been _me_ that broke it with an improper patch.  
> Whoops.  I'll follow up with anaconda folks and fix that up hopefully 
> tomorrow.
>
> --
> Fedora-buildsys-list mailing list
> [e-mail]
> https://www.redhat.com/mailman/listinfo/fedora-buildsys-list

For Reference:  Bug #233384

[Joel Andres Granados]

Jesse Keating wrote:
> On Wednesday 25 April 2007 05:24:58 Joel Andres Granados wrote:
> > I thought that the missing line was part of some new feature that was
> > not yet complete.  This is the reason why I posted it here, but I guess
> > anaconda guys took away the whole line when they should have just erased
> > the RPMS part of the command.  I'll post in on the anaconda list and see
> > what they have to say.
> > Regards.
>
> Oh dear, that may have been _me_ that broke it with an improper patch.  
> Whoops.  I'll follow up with anaconda folks and fix that up hopefully 
> tomorrow.
>
> --
> Fedora-buildsys-list mailing list
> [e-mail]
> https://www.redhat.com/mailman/listinfo/fedora-buildsys-list

The bug has been patched by Jeremy.
Regards