Re: BUG: listmatches called with bogus list

Re: BUG: listmatches called with bogus list

[Sven Wischnowsky]

Tanaka Akira wrote:

> `echo =zsh<TAB>-<M-x>expand-word<CR>' shows `BUG: listmatches called
> with bogus list'.

Time for a bit of defensive programming, I think.

Avoiding the error message was simple but finding out how to avoid the 
display bug that showed up after that...

Bye
 Sven

diff -ru ../z.old/Src/Zle/compresult.c Src/Zle/compresult.c
--- ../z.old/Src/Zle/compresult.c	Wed Jan  5 16:41:42 2000
+++ Src/Zle/compresult.c	Thu Jan  6 10:24:36 2000
@@ -1827,9 +1827,9 @@
 mod_export int
 invalidate_list(void)
 {
-    if (showinglist == -2)
-	listmatches();
     if (validlist) {
+	if (showinglist == -2)
+	    zrefresh();
 	freematches(lastmatches);
 	lastmatches = NULL;
 	hasoldlist = 0;
diff -ru ../z.old/Src/Zle/zle_misc.c Src/Zle/zle_misc.c
--- ../z.old/Src/Zle/zle_misc.c	Wed Jan  5 16:41:43 2000
+++ Src/Zle/zle_misc.c	Thu Jan  6 09:51:28 2000
@@ -640,7 +640,8 @@
 executenamedcommand(char *prmt)
 {
     Thingy cmd;
-    int len, l = strlen(prmt), ols = listshown, feep = 0, listed = 0, curlist = 0;
+    int len, l = strlen(prmt), feep = 0, listed = 0, curlist = 0;
+    int ols = (listshown && validlist);
     char *ptr;
     char *okeymap = curkeymapname;
 

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

BUG: listmatches called with bogus list

[Tanaka Akira]

`echo =zsh<TAB>-<M-x>expand-word<CR>' shows `BUG: listmatches called
with bogus list'.

Z(2):akr@is27e1u11% Src/zsh -f
is27e1u11% bindkey -e; autoload -U compinit; compinit -D
is27e1u11% zmodload zsh/complist
is27e1u11% echo =zsh<TAB>

->
is27e1u11% echo =zsh-<M-x>expand-word<CR>
zsh                zsh-3.1.5-pws-12   zsh-3.1.5-pws-3    zsh-3.1.6-pws-11 
zsh-3.0.5          zsh-3.1.5-pws-14   zsh-3.1.5-pws-4    zsh-3.1.6-pws-12 
zsh-3.0.6          zsh-3.1.5-pws-15   zsh-3.1.5-pws-5    zsh-3.1.6-pws-13 
zsh-3.0.7          zsh-3.1.5-pws-16   zsh-3.1.5-pws-6    zsh-3.1.6-pws-2  
zsh-3.1.4+         zsh-3.1.5-pws-17   zsh-3.1.5-pws-7    zsh-3.1.6-pws-3  
zsh-3.1.4++        zsh-3.1.5-pws-18   zsh-3.1.5-pws-8    zsh-3.1.6-pws-4  
zsh-3.1.4+3        zsh-3.1.5-pws-19   zsh-3.1.5-pws-9    zsh-3.1.6-pws-5  
zsh-3.1.4+4        zsh-3.1.5-pws-20   zsh-3.1.6          zsh-3.1.6-pws-6  
zsh-3.1.4+5        zsh-3.1.5-pws-21   zsh-3.1.6-bart-7   zsh-3.1.6-pws-9  
zsh-3.1.5          zsh-3.1.5-pws-22   zsh-3.1.6-bart-8   zsh-3.1.6-test-1 
zsh-3.1.5-4533     zsh-3.1.5-pws-23   zsh-3.1.6-dev-14   zsh-3.1.6-test-2 
zsh-3.1.5-4564     zsh-3.1.5-pws-24   zsh-3.1.6-pws-1    zsh-3.1.6-test-3 
zsh-3.1.5-pws-11   zsh-3.1.5-pws-25   zsh-3.1.6-pws-10   zsh.old          

->
is27e1u11% echo =zsh-
BUG: listmatches called with bogus list
-- 
Tanaka Akira

BUG: listmatches called with bogus list

[Zoltan Hidvegi]

Here is the bug:

% zsh -f
bolyai% setopt autolist                                             
bolyai% bindkey '^[[19~' where-is 
bolyai% 
Where is: kill-_
BUG: listmatches called with bogus list

As I remember, it is somewhere in Zefram's code which causes this bug (i.e. it
is not present in beta10).

Zoltan

Re: BUG: listmatches called with bogus list

[Zoltan Hidvegi]

I wrote:
> Here is the bug:
> 
> % zsh -f
> bolyai% setopt autolist                                             
> bolyai% bindkey '^[[19~' where-is 
> bolyai% 
> Where is: kill-_
> BUG: listmatches called with bogus list
> 
> As I remember, it is somewhere in Zefram's code which causes this bug (i.e. it
> is not present in beta10).

I forgot to mention that the the actual message cames when I press TAB or
newline.

Zoltan

Re: BUG: listmatches called with bogus list

[Zefram]

-----BEGIN PGP SIGNED MESSAGE-----

Zoltan wrote:
>Here is the bug:
>
>% zsh -f
>bolyai% setopt autolist                                             
>bolyai% bindkey '^[[19~' where-is 
>bolyai% 
>Where is: kill-_
>BUG: listmatches called with bogus list
>
>As I remember, it is somewhere in Zefram's code which causes this bug (i.e. it
>is not present in beta10).

Yup.  I put in a sanity check, because of the previous bug or two that
it would catch (and because of the possibility of there being more such
bugs, of course).

In this case, the sanity check got it wrong.  The list is absolutely
fine.  It's just that validlist wasn't set.  You can observe this: type
"ls <tab>", so that a list is formed, then do a where-is, and it will
happily list the editor function names.  Here's the patch:

 *** Src/zle_tricky.c.1.20	1995/08/09 05:33:48
 --- Src/zle_tricky.c	1995/09/07 10:30:55
 ***************
 *** 3704,3715 ****
   {
       int hw = haswhat, ip = ispattern;
       char *lp = lpre, *ls = lsuf;
 !     int nm = nmatches;
       char **am = amatches;
       char *ex = expl;
   
       haswhat = HAS_MISC;
       ispattern = 0;
       lpre = lsuf = "";
       expl = NULL;
   
 --- 3704,3716 ----
   {
       int hw = haswhat, ip = ispattern;
       char *lp = lpre, *ls = lsuf;
 !     int nm = nmatches, vl = validlist;
       char **am = amatches;
       char *ex = expl;
   
       haswhat = HAS_MISC;
       ispattern = 0;
 +     validlist = 1;
       lpre = lsuf = "";
       expl = NULL;
   
 ***************
 *** 3720,3725 ****
 --- 3721,3727 ----
       expl = ex;
       amatches = am;
       nmatches = nm;
 +     validlist = vl;
       lpre = lp;
       lsuf = ls;
       ispattern = ip;

 -zefram

-----BEGIN PGP SIGNATURE-----
Version: 2.6.i

iQBVAgUBME7MZmWJ8JfKi+e9AQG8DQH/VJyf9tfMEAHNfdthxCNEBaHjfIhkAjE/
6fNvRO2CtGi6DtjbWUrS2yy3hVvuovi8yB6EL1R10Q9/O+LULRpiIA==
=I40w
-----END PGP SIGNATURE-----