PATCH: Re: _rpm odd behavior

PATCH: Re: _rpm odd behavior

[Sven Wischnowsky]

Bart Schaefer wrote:

> ...
> 
> Is it a new problem that _arguments has trouble dealing with multiple
> single-letter options packed in the same word?

It wasn't. My fault (of course, sigh). It reset the `single' flag in
parse_cadef() although it shouldn't.

IT also adds the `-c'. Any other missing options anywhere?

Bye
 Sven

Index: Completion/Linux/_rpm
===================================================================
RCS file: /cvsroot/zsh/zsh/Completion/Linux/_rpm,v
retrieving revision 1.8
diff -u -r1.8 _rpm
--- Completion/Linux/_rpm	2000/05/22 14:11:34	1.8
+++ Completion/Linux/_rpm	2000/05/24 08:29:32
@@ -96,7 +96,7 @@
   case "$lstate" in
   query)
     _arguments -s \
-      '*-v[verbose mode]' -q \
+      '*-v[verbose mode]' -q -c \
       '--rcfile:resource file:_files' \
       '--ftpproxy:FTP proxy server:_hosts' \
       '--ftpport:FTP port number:' \
Index: Src/Zle/computil.c
===================================================================
RCS file: /cvsroot/zsh/zsh/Src/Zle/computil.c,v
retrieving revision 1.22
diff -u -r1.22 computil.c
--- Src/Zle/computil.c	2000/05/22 09:28:35	1.22
+++ Src/Zle/computil.c	2000/05/24 08:29:34
@@ -628,7 +628,6 @@
 
     all = ret = alloc_cadef(oargs, single, match, flags);
     optp = &(ret->opts);
-    single = flags = 0;
     anum = 1;
 
     sargs = args;
@@ -664,7 +663,7 @@
 		set_cadef_opts(ret);
 		ret = ret->snext = alloc_cadef(NULL, single, NULL, flags);
 		optp = &(ret->opts);
-		single = flags = nopts = ndopts = nodopts = 0;
+		nopts = ndopts = nodopts = 0;
 		anum = 1;
 	    }
 	    continue;

--
Sven Wischnowsky                         wischnow@informatik.hu-berlin.de

Re: PATCH: Re: _rpm odd behavior

[Bart Schaefer]

On May 24, 10:29am, Sven Wischnowsky wrote:
} Subject: PATCH: Re: _rpm odd behavior
}
} [The patch] also adds the `-c'. Any other missing options anywhere?

Depends on the version of rpm.  `-c' may not be valid in the 2.x versions
of rpm; there are some more that are valid in 3.x.  I don't think we have
a coherent scheme for dealing with revisions of the same command, do we?
Mostly we just complete for the newest one we know about.

I may take a look at _rpm later if no one else gets there first.  But it
may be tomorrow, so if someone else has time, don't wait for me.

We could try to parse the output of "rpm --help"; but it's heirarchical,
representing the modes selected by one of -q -e -U -b -t, so we'd still
need to assist it a bit.  I probably *won't* attempt that.

-- 
Bart Schaefer                                 Brass Lantern Enterprises
http://www.well.com/user/barts              http://www.brasslantern.com