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

Re: Perl files in base.



On Wed, 17 Sep 1997, Fabrizio Polacco wrote:

> Christian Schwarz wrote:
> > 
> > I still don't understand the reason (is this just me? how do the
> > others think about this?). If a package installs /usr/bin/foo, why
> > can't this program be used in the "postinst" script?
> > 
> 
> Well, probably I've skipped some part of my reasoning so, I'll try to be
> more deep :-)
> 
> Probably there is not a problem in using "installed" modules inside a
> postinst script, but ... there are also preinst, postrm and prerm
> scripts. Some of these scripts aren't only called during "normal"
> installation, but are also called during wind up, the reversal path that
> dpkg uses to rebuild a stable state when the installation fails. The
> intere process is described in the programming manual (it was there last
> time I sought, if it hasn't been moved to a new manual :-) and is quite
> complicated, although very clever.
> 
> What I meant was that we cannot make a policy that says 
>  "a maintainer can use a perl module installed by the package inside
>   the script postinst but not in prerm, and in postrm only if it's a
>   rainy day"

Why I can't we make such a policy? This sounds very reasonable, IMHO, and
I agree 100%. 

I'm against disallowing the use of the modules in the postinst scripts
just because it's dangerous in the other maintainer scripts. 

For example, we are still looking for a solution with the tetex-* postinst
script. This script could be made much faster by using a simple Perl
script that uses the dpkg-perl modules. Unless these modules are included
in in perl-base we have to include them in tetex-* (since "Depends:" are
really dangerous here). The modules would only be used in the postinst
script.

As I understand dpkg, the package is garantueed to be fully unpackaged by
the time the postinst script is called. (That's why it's called
"postinst", BTW :-)

I CC this mail to Klee--maybe he could verify if we've missed something.

[snip]
> I hope I've been plain and not too rude :-)

Everything's fine here!


Thanks,

Chris

--          _,,     Christian Schwarz
           / o \__   schwarz@monet.m.isar.de, schwarz@schwarz-online.com,
           !   ___;   schwarz@debian.org, schwarz@mathematik.tu-muenchen.de
           \  /        
  \\\______/  !        PGP-fp: 8F 61 EB 6D CF 23 CA D7  34 05 14 5C C8 DC 22 BA
   \          /         http://fatman.mathematik.tu-muenchen.de/~schwarz/
-.-.,---,-,-..---,-,-.,----.-.-
  "DIE ENTE BLEIBT DRAUSSEN!"


Reply to: