[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: debian-cd 2.2.13 and woody.



On Mar 16, Philip Charles wrote:
> I am in the process of configuring debian-cd to produce woody images. 
> Some interesting results.  non-US and non-free (initially) were included. 
> cvs was updated earlier today and no change found.  The MULTIBOOT option
> was enabled. I am running a woody system.
> 
> New base-woody and task-woody files were generated.  There were only minor
> differences between the new files and the old that came with debian-cd.  A
> seven CD set was then built.  Packages with "Task: foo" fields were
> scattered throughout the seven CDs. 
> 
> A second set was built with COMPLETE=0 and "popularity-contest-woody"
> deactivated along with the other non-essential bits and the result was
> about 1.3 CDs.
> 
> A reduced task-woody list was then generated using
> 
> cat task.list | while read LINE; do \
> apt-cache dumpavail | grep-dctrl -F Task $LINE -n -s Package;done > \
> task-woody
> 
> task.list being a partial list of the available tasks.
> 
> A third CD set was built along the lines of the second.  The result was
> 1.5 CDs, bigger than the second set.  On investigation it was found that
> the latest task-woody was bigger than the original.
> 
> A full task-list file was generated using my script.  The seven CD set
> built.  This time the packages with "Task: foo" were confined to the first
> two CDs with only one of these packages being on the third.
> 
> Conclusion 1.  The mini script for generating task-list in debian-cd is
> NBG.  I cannot see why this is the case.

The script in debian-cd is erroneous, and only picks up the names of
tasks with hyphens in them; I fixed it in CVS a few days back.  The
list should now be correct, too.

> For the final build COMPLETE=0 was set and popularity-contest deactivated,
> but the other bits in ../tasks/Debian_woody left active.  non-free was
> excluded and non-US included for this build.
> 
> Result, the second CD was 500 MB.
> 
> Conclusion 2.  We will have to do some hard thinking about what should be
> on the first and second CDs.

I can get all of the non-i18n tasks on CD 1 and have space left over
for lots of interesting packages.  My gut feeling is that the i18n
tasks can either be on CD #2, or a separate "woody-i18n" set with some
of the other tasks dropped from CD #1 may be appropriate.

As an alternative, I think if you dropped the non-Western European
language tasks (which pull in a lot of fonts and other big packages)
you could still have a decent #1 CD; the ISO-8851-1/15 language tasks
don't bring in a lot of stuff (dictionaries, mainly).  But I haven't
tried this.


Chris
-- 
Chris Lawrence <cnlawren@olemiss.edu> - http://www.lordsutch.com/chris/

Instructor and Ph.D. Candidate, Political Science, Univ. of Mississippi
208 Deupree Hall - 662-915-5765



Reply to: