compctl-related core dump

compctl-related core dump

[John Harres]

[-- Attachment #1: Type: text/plain, Size: 1356 bytes --]

I got no response to this, so I guess email to zsh-users isn't sent to
zsh-workers...

John
------- Forwarded Message

Date: Tue, 25 Feb 1997 16:57:29 -0700
From: John Harres <Harres@uwyo.edu>
Subject: compctl-related core dump
To: zsh-users@math.gatech.edu

While working on another odd command's compctl, I got a core dump out of 
3.0.2.  First, the compctl:

   compctl -x 's[-]' -k '(logical physical server database contents object as 
type dependencies from stripe incremental stripe replace copyover merge 
no_replace no_index indexfile sql_insert charset audits no_audits no_data 
data_only merge no_segments on_segment cache no_cache on_index_segment 
on_text_segment interfaces user password sybasedump masterinfo bcp_batch_size 
db_refs dryrun code verbose no_prompt no_status quiet debug)' - \
'c[-1,from][-1,stripe][-1.indexfile][-1,interfaces][-1,sybasedump][-1,masterinf
o]' -f -- dtsload

I get a core dump when I do

dtsload <tab>

Note that this compctl has nothing defined for this case.  I'm not sure what
it should do, but dumping core ain't it.


------- End of Forwarded Message
-- 
John Harres <harres@uwyo.edu>        http://horseman.uwyo.edu/~harres
<A HREF="http://www.uwyo.edu">University of Wyoming</A> 
Finger harres@horseman.uwyo.edu for PGP public key.
Key fingerprint =  FC B4 72 D6 37 A7 A1 73  BE 8B F3 E2 99 68 BF 5A 



[-- Attachment #2: Type: application/pgp-signature, Size: 239 bytes --]

Re: compctl-related core dump

[Xris Laas]

On 27 Feb, John Harres wrote:
> I got no response to this, so I guess email to zsh-users isn't sent to
> zsh-workers...

I'm not sure if it is (I'm subscribed to both), but I've CC'd both to
make sure it gets to both.  Sorry if this produces duplicates.

> Date: Tue, 25 Feb 1997 16:57:29 -0700
> From: John Harres <Harres@uwyo.edu>
> Subject: compctl-related core dump
> To: zsh-users@math.gatech.edu
> 
> While working on another odd command's compctl, I got a core dump out of 
> 3.0.2.  First, the compctl:
> 
>    compctl -x 's[-]' -k '(logical physical server database contents object as 
> type dependencies from stripe incremental stripe replace copyover merge 
> no_replace no_index indexfile sql_insert charset audits no_audits no_data 
> data_only merge no_segments on_segment cache no_cache on_index_segment 
> on_text_segment interfaces user password sybasedump masterinfo bcp_batch_size 
> db_refs dryrun code verbose no_prompt no_status quiet debug)' - \
> 'c[-1,from][-1,stripe][-1.indexfile][-1,interfaces][-1,sybasedump][-1,masterinf
> o]' -f -- dtsload
> 
> I get a core dump when I do
> 
> dtsload <tab>
> 
> Note that this compctl has nothing defined for this case.  I'm not sure what
> it should do, but dumping core ain't it.

The problem isn't that there's nothing defined for that case.  The
problem is merely compctl syntax in the line:

> 'c[-1,from][-1,stripe][-1.indexfile][-1,interfaces][-1,sybasedump][-1,masterinf
                        ^^^^^^^^^^^^^^
Change that to "[-1,indexfile]" .

BTW, dumping core still isn't a good solution for bad compctl syntax. 
Perhaps compctl should do a syntax check when it is invoked, instead of
causing problems like this.

--Chris

-- 
Prudence, indeed, will dictate that governments long established
should not be changed for light and transient causes; and accordingly
all experience hath shown that mankind are more disposed to suffer,
while evils are sufferable, than to right themselves by abolishing the
forms to which they are accustomed.
    -- Thomas Jefferson, Prologue, "Declaration of Independence"

{ Chris "Xris" Laas }-----{ chrisl@cybercom.net }-----{ xris@qfl.com }
<*>--------------{ http://www.cybercom.net/~chrisl/ }--------------<*>

Re: compctl-related core dump

[Peter Stephenson]

Xris Laas wrote:
> On 27 Feb, John Harres wrote:
> > While working on another odd command's compctl, I got a core dump out of 
> > 3.0.2.  First, the compctl:
> > 
> >    compctl -x 's[-]' -k '(logical physical server database contents object 
> as 
> > type dependencies from stripe incremental stripe replace copyover merge 
> > no_replace no_index indexfile sql_insert charset audits no_audits no_data 
> > data_only merge no_segments on_segment cache no_cache on_index_segment 
> > on_text_segment interfaces user password sybasedump masterinfo bcp_batch_si
> ze 
> > db_refs dryrun code verbose no_prompt no_status quiet debug)' - \
> > 'c[-1,from][-1,stripe][-1.indexfile][-1,interfaces][-1,sybasedump][-1,maste
> rinf
> > o]' -f -- dtsload
> 
> The problem isn't that there's nothing defined for that case.  The
> problem is merely compctl syntax in the line:
> 
> > 'c[-1,from][-1,stripe][-1.indexfile][-1,interfaces][-1,sybasedump][-1,maste
> rinf
>                         ^^^^^^^^^^^^^^
> Change that to "[-1,indexfile]" .

The test in zsh is insufficient when the incorrect condition has
another bracketed expression after it:  it finds the next comma after
and doesn't realise it's gone too far.  This fixes it (for 3.1.1, it
will apply to 3.0.x if you can find the file with get_xcompctl() in
it).

*** Src/Zle/compctl.c.bug	Sun Jan  5 22:07:31 1997
--- Src/Zle/compctl.c	Fri Feb 28 11:11:14 1997
***************
*** 598,605 ****
  		    c->u.l.b[l] = ztrdup(tt);
  		} else {
  		    /* remaining patterns are number followed by string */
! 		    for (; *t && *t != '\201'; t++);
! 		    if (!*t) {
  			zwarnnam(name, "error in condition", NULL, 0);
  			freecompcond(m);
  			return 1;
--- 598,605 ----
  		    c->u.l.b[l] = ztrdup(tt);
  		} else {
  		    /* remaining patterns are number followed by string */
! 		    for (; *t && *t != '\200' && *t != '\201'; t++);
! 		    if (!*t || *t == '\200') {
  			zwarnnam(name, "error in condition", NULL, 0);
  			freecompcond(m);
  			return 1;

-- 
Peter Stephenson <pws@ifh.de>       Tel: +49 33762 77366
WWW:  http://www.ifh.de/~pws/       Fax: +49 33762 77413
Deutsches Elektronen-Synchrotron --- Institut fuer Hochenergiephysik Zeuthen
DESY-IfH, 15735 Zeuthen, Germany.