Re: Bug#252632: ITP: wmtemp -- WindowMaker dock applet displaying lm_sensors temperature values
Lars Steinke <firstname.lastname@example.org> writes:
> On Sat, Jun 05, 2004 at 04:30:32PM +0200, Eduard Bloch wrote:
>> Lars Steinke schrieb am Samstag, den 05. Juni 2004:
>> > you are speaking my mind exactly. Anyway, I am not a programmer, so
>> > merging the code of wmtemp and wmgtemp is not feasible for me and I
>> > don't think a 14k .deb is such a waste of project resources that it is
>> > worth the effort (of a code merge or even a lengthy discussion on
>> > debian-devel...) - or is it ?
>> Wrong, it is a waste. You do not only add ONE 14k .deb, you add a 17k
>> source plus dsc plus orig plus diff plus 10*14kB (for each arch, and
>> maybe more than that), and, last but not least, you make apt and dpkg
>> slower for EVERYONE even if many people do not care about some cute toy
>> for some weird WM.
> Fair enough. Sorry, I am in the process of being taught to simplify
> problems to the point of irrelevance ;).
> Well, I can follow your line of reasoning. But if you spin that thought
> further we would ultimately strive to create a single unified
> package "wm-dockapps" containing the roughly 90 dock apps for
> WindowMaker packaged individually in Debian, even though probably
> no one wants all of them on their system anyway...
If the package of all 90 dockapps is just 90*20K = ~2Mb do you realy
think anyone would care? Thats about the same size as an backup
kernel-iamge (without modules). The time saved searching through 90
dockapp packages to find one the user likes is probably more anoying
than that little space waste. A new user would probably just install
all of them anyway and then try out which to use.
> And sadly enough, I don't think it's the only class of apps affected by
> the problem of explosive package number growth - the numerous small
> KDE or Gnome apps come to my mind spontaneously, or xmms plugins.
Just wait for multiarch. That will roughly tripple the number of
packages for biarch systems and quadruple it for triarch systems like
powerpc (with qemu for i386) or mips.
> As you can't realistically expect that all package developers affected
> will sort this out between them (they haven't by now), I would
> think this problem needs to be addressed in a more general way, perhaps
> via Debian Policy (which I find so detailed already that it's unfeasible
> to work with without lintian) or by a technical solution (e.g. an
> automatic merging of packages with common dependencies based on a
> certain ruleset. When thinking about responsibility for bug reports,
> a merge would give rise to a lot of complexity that would have to be
> handled here...)
When some people find they have such overlaps in their intrest they
sometimes join up and start a common mailinglist and revision control
system somewhere, like the Gnome team.
You don't neccessarily have to merge all 90 wm dockapps into a single
package, you could keep them all seperate and still make it a joined
effort. Having a mailinglist and some rcs means that you can inform
each other about upcoming changes or subsitute easily during holidays.
>> And since the applications are similar and have exactly the same
>> Dependencies and Build-Dependencies, they should be merged into one
>> package. You don't have to merge the code, just combine the source
> Is there any documented process for merging source packages, a search on
> d.o does not seem to yield anything ? How am I supposed to handle the
> different copyrights, ChangeLogs and so on ? Also, I am not the packager
> of wmgtemp so it's not only my effort asked for here - two people have
> to sacrifice their time for this venture... While I do find the cause of
> saving time for all Debian users just and noble, developers usually
> don't have it in abundance either.
> Just my 2 (Euro)cents,