Re: [PATCH] New Dpkg::Deps module to replace parsedep() and showdep()
On Wed, Oct 17, 2007 at 11:01:04PM +0200, Raphael Hertzog wrote:
> On Wed, 17 Oct 2007, Frank Lichtenheld wrote:
> > Some notes:
> > - Should the parsing algorithm enforce the negated arch syntax, i.e.
> > that either none or all architectures need to be negated?
> I have no opinion here. There are many other stuff that could be more
> thoroughly checked/enforced as well. For example using arch restriction
> list on Arch: all packages (or even on dependencies in general, it's
> officially restricted to Build-Dependencies currently but I believe it
> works well on Arch: any package, i.e. they are expanded as required
> depending on the architecture).
> I don't see the point to add it immediately in any case. The current parser
> didn't enforce it either.
Yeah, it is certainly no condition for inclusion. But it is a good time
to ask the question :)
> > - for future patch sendings to the list I would prefer using a method
> > like on LKML, etc. I.e. sending each patch as a single mail. This
> > makes reviews piece-by-piece waay more pleasant. The git-send-email
> > tool can help with doing such mass-sendings.
> Fine, I hesitated and since I didn't want to annoy people with too many
> mails I choosed this approach.
debian-dpkg isn't really a high-traffic mailinglist...
As long as all the mails are sent as one thread it doesn't really add
that much work for people uninterested in the topic and it makes reviews
way more pleasant.
> > And generally I would rather err on the side of too much strings marked
> > for translation and let the translators decide whether they are worth
> > translating.
> IMO, most translators do not even consider the possibility to not
> translate a string... and they always target 100% of translation.
Probably true. For error messages I still stand by my opinion, though.
(e.g. debug messages are a whole other topic).
Frank Lichtenheld <firstname.lastname@example.org>