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

Re: cvs and debian-cd

Le Wed, Jun 14, 2000 at 04:44:05PM +0100, Philip Hands écrivait:
> It was mentioned in someone else's mail to the list, and seemed like a
> good plan.  I'm of the opinion that trying it in a test cycle is the
> quickest way of finding out if people are going to get stressed about
> whatever it is that drops off the end of the CD as a result.
> I'd hope that some sort of consensus will emerge by release time, so
> that the hate mail about package XYZ not being on CD#1 doesn't all come
> to me.

As soon as you get this consensus, please commit this little change then. :)

> I do think having a limit expressed in terms of MB and calculated such
> that you could say that the limit was 650MB, and be assured that you
> were going to get something that fits on a CD would be a step forward
> though.  The alternative caused me significant frustration, which I
> wouldn't wish on anyone else.

Of course, I agree with you, however I'm no more than a yet another perl
hacker that needed potato CD at that thime when slinkcd was painful to
use for potato (that's why I wrote YACS) and I really don't know the
internals of iso9660 ... 

However I have an idea that you could try, instead of generating the
complete CD set, you could just do everything until you get to make
packages (don't launch it, it would install the packages in the temp tree). 
At that point, you check the size of the temporary tree and you use this
information to subtract it from 640MB (yes we need to have some free space
for the Packages* and md5sum.txt hat will come after).

for ARCH in i386 m68k alpha sparc powerpc arm
	export ARCH
	make distclean
	make status
	make list COMPLETE=1
	make bootable
	make upgrade
	echo "Size of binary CD 1, 2, 3 for $ARCH :"
	# Don't forget 1 -> CD1, 2 -> CD2, 3 -> CD3 if you already have
	# the non-US patch sent by Steve on this list

Raphaël Hertzog -+- http://strasbourg.linuxfr.org/~raphael/

Reply to: