Bug#599652: openoffice.org-common: Please document reason for libtextcat-data-utf8 recommends, and please consider changing dependency level
On Mon, Oct 11, 2010 at 10:05:31AM +0200, Rene Engelhard wrote:
> On Sun, Oct 10, 2010 at 06:04:13PM -0700, Josh Triplett wrote:
> > > > Alternatively, if openoffice.org really can do without this package for
> > > > most purposes, please consider documenting the need for this package in
> > >
> > > It can, but that feature would fail for many languages (and we'd have
> > > dangling symlinks here) for the languages where we get these patterns
> > > from libtextcat directly.
> >
> > So, if OO.o needs this functionality to the point of having dangling
> > symlinks without it, why not just Depends on it? Hardly seems like an
>
> Because that package is a hack. And if some package (unlikely, though)
> started to depend on libtextcat0 it would need to use the
> patterns from libtextcat-data. Which conflicts with libtextcat-data-utf8
> because both share the same paths.
>
> Of course, this would be a case at any time, but when OOo depended on
> libtextcat-data-utf8 there would be no way out of it.
I see. Sounds like OO.o needs to recommend it rather than depending on
it primarily because libtextcat-data{,-utf8} can't coexist, and someone
might want the version not hacked for use with OO.o. Ouch.
(Could OO.o use a libtextcat-utf8-0 which uses -data-utf8, letting
libtextcat0 use -data?)
> > OO.o user would notice a 250k package, but they'd notice missing
> > functionality. (Or worse yet, they *wouldn't*.)
>
> Yes, in most cases they wouldn't, true... That feature is "hidden" in
> the statusbar...
Exactly the kind of information I didn't know about OO.o to begin with. :)
> > This just seems like taking Recommends to an extreme. I'd tend to think
> > that packages shouldn't use Recommends for something that only takes up
> > a bit of space (and takes up proportionally far less than the package
> > itself), as opposed to something that a user seriously might not want on
> > their system for other reasons (like an extra running daemon).
>
> Recommends is not a decisison between space or not space.
Well aware of that. I intended to make precisely that argument, in
reverse: that if a package strictly speaking *can* do without some
functionality, but space represents the only reason to omit it, and not
much space at that, then using a Recommends seems excessive. A case of
giving the user more choices than they want or need.
That said, the description you gave earlier in your mail makes it clear
why this dependency uses Recommends: because otherwise nobody could
install the non-UTF8-hacked version. Barring a solution to that
problem, I agree that it doesn't seem possible to make this a Depends.
> From the policy:
[...snip text I read before filing this bug. :) ...]
> Language guessing imho is not that important to fit into the first category,
> but as OOo normally includes an own copy of those files in it's common partm
> it can be argued that it's "found together with this one in all but unusual
> installations"
What do you mean "normally includes an own copy of those files"? Does
this represent a case where OO.o normally bundles a copy of
what libtextcat-data-utf8 contains, and you split it out for
modularity/maintainability?
In any case: arguably, OO.o has so much functionality that few
*individual* pieces can claim to provide "a significant amount of
functionality", but nevertheless when people install OO.o they expect
all of its features to work. To make an analogous argument, someone
could split the programming language mode for some obscure language out
of the vim or emacs packages, since it doesn't qualify as "a significant
amount of functionality", but nevertheless when people install vim or
emacs they expect the whole set of functionality.
> > such documentation so that the user will see it when installing packages
> > in a package manager. Unless you'd propose trying to make OO.o use
> > packagekit or similar to auto-install it on demand? :)
>
> ? People can look up what libtextcat-data-utf8 is about in a package
> manager?
But again, this doesn't explain why that applies to OO.o.
For example, you mentioned that it helps OO.o display some language
information in the status bar. I'd alternatively wondered if not having
libtextcat-data-utf8 would break spellchecking or grammar checking
(since OO.o needs to know what language to check), or some other
functionality entirely. Without that information, I don't know whether
I need libtextcat-data-utf8 or not, and I just have to blindly install
it, or not.
Does that make sense?
- Josh Triplett
Reply to: