PATCH: 3.1.6-dev-20: _dpkg thinko

PATCH: 3.1.6-dev-20: _dpkg thinko

[Clint Adams]

Allegedly this is causing "-i=irs" to be exported, though I can't
reproduce it.

Index: Completion/Debian/_dpkg
===================================================================
RCS file: /cvsroot/zsh/zsh/Completion/Debian/_dpkg,v
retrieving revision 1.1.1.5
diff -u -r1.1.1.5 _dpkg
--- Completion/Debian/_dpkg	2000/01/30 00:59:22	1.1.1.5
+++ Completion/Debian/_dpkg	2000/04/03 15:20:50
@@ -23,7 +23,7 @@
 case "$words[1]" in
 dpkg)
 _arguments -s '(--install)-i[install packages]:Debian package:_files -g \*.deb' \
-           '--install(-i):Debian package:_files -g \*.deb' \
+           '(-i)--install:Debian package:_files -g \*.deb' \
            '--unpack[unpack package]:Debian package:_files -g \*.deb' \
            '(--record-avail)-A[record available]:package file:_files' \
            '(-A)--record-avail:package file:_files' \

Re: PATCH: 3.1.6-dev-20: _dpkg thinko

[Clint Adams]

> Index: Completion/Debian/_dpkg

Should I submit this to the Patch Manager?  Does someone want to give
me commit access?

Re: PATCH: 3.1.6-dev-20: _dpkg thinko

[Bart Schaefer]

On Apr 3, 12:55pm, Clint Adams wrote:
> Subject: Re: PATCH: 3.1.6-dev-20: _dpkg thinko
> > Index: Completion/Debian/_dpkg
> 
> Should I submit this to the Patch Manager?

I was actually wondering how whether and how well the "recent patches"
archive on www.zsh.org/mla/ is automated.  Could we arrange for zsh-list
patches to go straight into the sourceforge patch manager?

RE: PATCH: 3.1.6-dev-20: _dpkg thinko

[Andrej Borsenkow]

> >
> > Should I submit this to the Patch Manager?
>
> I was actually wondering how whether and how well the "recent patches"
> archive on www.zsh.org/mla/ is automated.  Could we arrange
> for zsh-list
> patches to go straight into the sourceforge patch manager?
>


The problem is, half (at least) of all patches do not have PATCH:
prefix. And I think, we should distinguish between

- CVS commits
- user submitted patches

It is hard to intermix them ... CVS commits depend on other patches and
mostly cannot be applied to clean version while with user's patches it
is exactly other way round - they probably need modification/merging to
apply to current CVS.

Somehow I'm getting the feeling that "patch archive" is not as useful at
all ...

-andrej

exporting -variables [was Re: PATCH: 3.1.6-dev-20: _dpkg thinko]

[Clint Adams]

> Allegedly this is causing "-i=irs" to be exported, though I can't
> reproduce it.

I can.  If ALL_EXPORT is on, variables beginning with
hyphens will be exported during _arguments.  Is this a problem?
bash seems to think such things are invalid.

e.g.:
sh: invalid character 45 in exportstr for -f