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

Re: Upgrade procedure for tetex

'Craig Sanders wrote:'
>On Sat, 1 Mar 1997, Chris Fearnley wrote:
>> I think tetex should be removed from bo and the old tex reinstated.
>> The new packaging scheme is incompatible with a smooth upgrade process
>> and I haven't seen (nor been able to conceive) of any way to fix
>> tetex before the code freeze. Any thoughts I have that /might/ work
>> (renaming all the tetex packages to names compatible with the old
>> tex in such a way that upgrading is possible) are extensive enough
>> in nature that risking keeping tetex in the distribution seems very
>> unwise to me.
>I think this is a bit hasty.
>A simple solution would be to get rid of the "Replaces:" lines in the
>control file and just have "Conflicts". Then put instructions on how to
>switch from the old tex packages to tetex in the Description field.

That "solution" would throw the user into the conflict resolution
screen in dselect with all of the old TeX packages and the one tetex
package which the user had selected.  Because so many things depend on
TeX, the user will be unable to resolve the dependencies without lots
of trial-and-error and a thorough understanding of dpkg and Debian
packaging.  Therefore, I don't consider it a sufficient solution.  I
think it would be worth testing, I could be wrong after all.  But it
seems dubious to test these types of radical solutions so close to the
release of our next "stable".  The old TeX though buggy, at least
installs properly with dselect.

If it can't be easily upgraded via dselect, it's broken and shouldn't
be in the next stable.  Period.

The only solution that will work (as far as I can see) is if the tetex
packaging were to be totally reworked so that each package replaced
only one of the original TeX packages (i.e., each were functionally
equivalent to the one it is replacing).  Even this may have problems.
[It may be that the names of the original packages need to be
maintained as well.]  Hence my insistence that a longer period of
testing be provided.

Christopher J. Fearnley            |    Linux/Internet Consulting
cjf@netaxs.com, cjf@onit.net       |    UNIX SIG Leader at PACS
http://www.netaxs.com/~cjf         |    (Philadelphia Area Computer Society)
ftp://ftp.netaxs.com/people/cjf    |    Design Science Revolutionary
"Dare to be Naive" -- Bucky Fuller |    Explorer in Universe

Reply to: