patching upstream -- why not?

[G. Milde]

On 31.08.05, Jörg Sommer wrote:
> Hello G.,
> 
> G. Milde schrieb am Wed 31. Aug, 10:19 (+0200):
> > On 23.08.05, Jari Aalto wrote:

...
> > > Please move 
> > > 
> > >     /etc/jed.conf
> > >     /etc/jed-init.d/
> > > 
> > > Under common directory => /etc/<package> like
> > > 
> > >     /etc/jed/config 
> > >     /etc/jed/jed-init.d/
> > 
> > Unfortunately, the path /etc/jed.conf is "hardcoded" into the upstream
> > release, changing it would introduce one more Debian-specific patch (and
> > incompatibility with original documentation ...).
> 
> Why do you think we should not patch upstream files? It's common
> practice. 

I gave one reason, (incompatibility with original documentation ...).
Patches always mean additional work for the packagers and can introduce
consistency problems, so IMHO they should only be a last resort.

Besides patching site.sl (a file that is likely to change with every
release) we had to patch all documentation that says "looks for
/etc/jed.conf" or "looks for jed.conf in ..." (and keep this patches up
to date too).

IMO it is not worth the effort in order to fix a minor problem
(actually, IMHO the bug should be labeled "enhancement request" instead
of "minor").

BTW, `locate *.d` showed all *.d/ directories right under /etc/, so
keeping /etc/jed-init.d/ is also a case of consistency.

> John is IMO a person who would see that it works before he do any
> changes. It's hard to get any changes into JED or SLang, more than ever
> if they a distribution specific.

This is true, I doubt that John would like to expand the "config file
search path" easily (nor would I).

Still, I think contacting the upstream author should always have
priority. Maybe I am allergic, as I was stunned to see a bunch of
patches for my Jedmodes modes in jed-extra fixing bugs I never received a
report about. 

Guenter

-- 
G.Milde web.de