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

Re: [SCM] Debian package checker branch, master, updated. 38002a14d5e0b494449410efb1d8555ead1cef30



On Sun, 2008-07-06 at 12:29 -0700, Russ Allbery wrote:
> "Adam D. Barratt" <adam@adam-barratt.org.uk> writes:

> > This could be easily fixed by renaming the postinst variant to
> > maintainer-script-loads-obsolete-confmodule - or just
> > loads-obsolete-confmodule. Does renaming tags cause any major issues?
> > (other than invalidating any overrides for the tag).
> 
> Nope.  And that's what I'd do there as well.

Thanks. I've just done so, with the additional step of merging the
config tag as well. (Unlike some of the others, the
config-loads-obsolete-confmodule tag was identical to the postinst tag
other than s/postinst/config/, so merging them seemed harmless in this
case).

> > It also has the potential to require
> > yet-another-does-not-load-confmodule tag, unless we renamed the postinst
> > tag to be more generic; there is a fair amount of postinst-specific
> > information in its description currently.
> 
> I think that the check for other scripts is fundamentally different than
> the one for config and postinst.  The last two have special requirements
> that are part of how debconf works, whereas the other scripts just need to
> load it if they use it.

Yeah. config and postinst are easy - if templates is present, both of
them must be using debconf. For the others we'd need a reliable check
for "uses debconf", which is fairly simple - if long-winded - for shell
scripts but somewhat more prone to errors when dealing with perl
maintainer scripts.

Adam


Reply to: