[Surfraw-devel] Bug#209155 and namespace issues

[Ian Beckwith]

The list lives!

Kevin said:
> Hello,
> 
> The list has been pretty quiet lately, so I wanted to see what we are 
> planning on doing about #209155 and the associated namespace issues.  
> What is the rough consensus on what we should do, and what is still 
> left to be done?

Hi. As I said in a previous message (Oct 1st), I've had a stab at
fixing the problems, and checked it into a NAMESPACE branch under cvs.

Has anyone had a chance to look at it yet? Any comments?

to check it out, do:

cvs -d:pserver:yourname@cvs.alioth.debian.org:/cvsroot/surfraw checkout -r NAMESPACE surfraw

Sarge is fast approaching, and it would be nice to get everything
resolved in time for that. If we can't, someone should at least do an
interim upload fixing #209155, the conflict with translate. On the
main branch I've uncommented the line in debian/rules that renames
translate to stranslate.

A possible problem with the namespace stuff and surfraw-update-path: I
was reading debian-policy, and it mentioned that packages should not
modify /etc/profile, as it belongs to base-files. Is that specifically
because it is in base-files, or because it belongs to a package other
than surfraw? Is the fact that we ask in debconf first enough to make
it ok?  If not, how about if we changed the question to a note and
just told them about surfraw-update-path? would that satisfy policy?

I've checked in (NAMESPACE branch too) two of the elvi Moritz
Muehlenhoff <jmm@informatik.uni-bremen.de> posted to bug #217799

leo: for dict.leo.org a german<->english dictionary. 
I've added option handling to this, and renamed it to leodict, as it
looks like leo.org provides various other searchable services.

happypenguin: for www.happypenguin.org, a linux games site.
I didn't bother adding options for this, as it looks like the search
interface may be in flux.

I haven't checked in grgoogle, a groups.google.com elvi, but it brings
up the question of should we rename the deja elvi to something more
obviously googley.

I've also upped the Standards-Version in both branches to 3.6.1.0
As far as I can tell, no changes are needed for this.

One perplexing thing, which could be my setup. While building the
NAMESPACE branch, it kept screwing up (something to do with prefix)
and tried to write/delete surfraw files from the actual system. I
fixed this by changing prefix= to DESTDIR= in debian/rules (which
incidently solves the problem with ELVIDIR in configure.in). However,
the main branch still builds fine using prefix= in debian/rules. I've
run fakeroot debian/rules clean, aclocal, automake, autoconf on both
branches, but the discrepancy remains. Odd. I've also run into similar
prefix-related weirdness messing with some other packages (perl modules 
that used MakeMaker). any ideas?

Ian.