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

Re: RFS: viewglob -- A graphical display of directories referenced at the shell prompt

On Sat, Sep 04, 2004 at 02:50:35PM +0100, James Troup wrote:
> Colin Watson <cjwatson@debian.org> writes:
> > On Fri, Sep 03, 2004 at 11:02:26PM +0200, Michael Schiansky wrote:
> >> Why do you call dpatch 'obfuscated' ? 
> >> Before I used it for one of my packages i quickly checked the code and it
> >> simply works. No big fancy stuff etc.
> >> 
> >> Did I miss a part?
> >
> > Compared to simply making the source changes directly, it's obfuscated.
> Personally I find a bunch of unrelated changes lumped together in a
> big diff (like, say, in openssh to pick a purely random example) is a
> much worse obfuscation than being forced to learn './debian/rules
> patch'.

Or 'debian/rules patched-source' or 'debian/rules source.make' or 'make
-f debian/scripts/somethingorother blah' or any number of other things;
every time I unpack a separated-patches source package and want to look
at it I always have to either read lots of twisty included makefiles
(and my make is pretty fluent, I dread to think what it would be like if
it weren't) or else run random debian/rules targets until I find one
that works. Sometimes I just give up, run 'debian/rules build', and
Ctrl-C it part-way through, which is just crap. If they all had the
common decency to use the same simple target name I probably wouldn't
mind, but every patch system in Debian is so freaking NIH that they're
never going to cooperate enough for that to happen.

openssh is not my most shining example of source packaging; it's
maintained in a strange way (CVS, God help me) for historical reasons.
Some day I'll fix it.

> > I recommend using a good revision control system instead, which
> > offers similar benefits to developers while leaving things clear for
> > users.
> A good revision control system is of little use to other developers if
> it's not available to them ...

Like I say, I'm happy with the patches being exposed in debian/patches/,
I just want them pre-applied if they're there. perl does this, and it's
*so* much easier to deal with.

Lots of people arguing for patch systems are arguing for their own
convenience, not for other people's. Revision control should be able to
solve that part of the puzzle.


Colin Watson                                       [cjwatson@debian.org]

Reply to: