PATCH: Re: Approximate completion

PATCH: Re: Approximate completion

[Sven Wischnowsky]

Peter Stephenson wrote:

> Approximate completion seems to be be broken in some non-obvious way.  If I
> try to complete 50dc.z to 50cd.ztst, the right completion appears as an
> argument to compadd, but nothing is inserted.  Correction with ^Xc seems to
> work.  Indeed, completing 50dc.zts works, too: since I have max errors set
> to 2, this suggest that for some reason it isn't matching with a * at the
> end of the test string at some point.

No. I had to make globbing flags be treated specially lately for
functions like _path_files that use hidden prefixes. The code now
takes the flags (if they are at the beginning of $PREFIX) and moves
them after the matched prefix. But I forgot to update the cached
prefix-length after extracting the flags.

Bye
 Sven

diff -u -r oldsrc/Zle/compcore.c Src/Zle/compcore.c
--- oldsrc/Zle/compcore.c	Wed Dec 29 22:26:15 1999
+++ Src/Zle/compcore.c	Mon Jan  3 19:58:57 2000
@@ -1680,6 +1680,7 @@
 			p[1] = sav;
 
 			lpre = p + 1;
+			llpl -= gfl;
 		    }
 		}
 		s = dat->ppre ? dat->ppre : "";

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

Re: PATCH: Re: Approximate completion

[Sven Wischnowsky]

I wrote:

> But I forgot to update the cached
> prefix-length after extracting the flags.

Uh oh. Two more of these...

Bye
 Sven

diff -ru ../z.old/Src/Zle/compcore.c Src/Zle/compcore.c
--- ../z.old/Src/Zle/compcore.c	Tue Jan  4 10:15:43 2000
+++ Src/Zle/compcore.c	Tue Jan  4 11:35:20 2000
@@ -1696,6 +1696,7 @@
 		    }
 		    pline = matchparts;
 		    lpre += ml;
+		    llpl -= ml;
 		    bcp = ml;
 		    bpadd = strlen(s) - ml;
 		} else {
@@ -1721,6 +1722,7 @@
 		    }
 		    sline = revert_cline(matchparts);
 		    lsuf[llsl - ml] = '\0';
+		    llsl -= ml;
 		    bcs = ml;
 		    bsadd = strlen(s) - ml;
 		} else {

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