Bug w/matching control

Bug w/matching control

[Felix Rosencrantz]

[-- Attachment #1: Type: text/plain, Size: 684 bytes --]

I'm seeing a problem with matching control and characters being deleted. I'm
using the data from workers/13320

I've attached a patch to 54compmatch.  I think the output should be the same as
the previous testcase.  Though this time with input of "g-2", we get a
"glibc-", with the 2 deleted.  This isn't checked in, since it won't pass.

The actual output seen from the test is:
! line: {tst glibc}{-2.1.94-3.i386.rpm}
  COMPADD:{}
! INSERT_POSITIONS:{9:27}
--- 1,3 ----
! line: {tst glibc-}{}
  COMPADD:{}
! INSERT_POSITIONS:{10}

-FR


__________________________________________________
Do You Yahoo!?
Get email at your own domain with Yahoo! Mail. 
http://personal.mail.yahoo.com/

[-- Attachment #2: test54.txt --]
[-- Type: text/plain, Size: 400 bytes --]

Index: 54compmatch.ztst
===================================================================
RCS file: /cvsroot/zsh/zsh/Test/54compmatch.ztst,v
retrieving revision 1.8
diff -r1.8 54compmatch.ztst
623a624,630
>  test_code $workers_13320_matcher workers_13320_list
>  comptest $'tst g-2\t'
> 0:Test from workers 13320
> >line: {tst glibc}{-2.1.94-3.i386.rpm}
> >COMPADD:{}
> >INSERT_POSITIONS:{9:27}
> 

Re: Bug w/matching control

[Sven Wischnowsky]

Felix Rosencrantz wrote:

> Thanks, Sven, that seems to fix the problem with the characters being deleted,
> but the insert positions are different than the other case, they should be
> 9:27, there seems to be an extra 11 in there.
> 
> When I run with the patch I sent before I'm getting a diff for
>   line: {tst glibc}{-2.1.94-3.i386.rpm}
>   COMPADD:{}
> ! INSERT_POSITIONS:{9:11:27}

Ah, right. Better reset the CLF_MISS flag if it's in the wrong cline
struct.


Bye
 Sven

Index: Src/Zle/compmatch.c
===================================================================
RCS file: /cvsroot/zsh/zsh/Src/Zle/compmatch.c,v
retrieving revision 1.32
diff -u -r1.32 compmatch.c
--- Src/Zle/compmatch.c	2001/03/05 10:48:11	1.32
+++ Src/Zle/compmatch.c	2001/03/06 11:29:06
@@ -1923,7 +1923,10 @@
 			 ((tn->flags & CLF_NEW) || !cmp_anchors(o, tn, 0));
 		     t = tn);
 		if (tn) {
+		    int of = o->flags & CLF_MISS;
+
 		    diff = sub_join(o, n, tn, 0);
+		    o->flags = (o->flags & ~CLF_MISS) | of;
 
 		    if (po && po->prefix && cmp_anchors(n, pn, 0)) {
 			po->flags |= CLF_MISS;
@@ -2069,7 +2072,10 @@
 				    cmp_anchors(tn, o, 1)) break;
 
 			    if (tn) {
+				int of = o->flags & CLF_MISS;
+
 				if ((diff = sub_join(o, n, tn, 0))) {
+				    o->flags = (o->flags & ~CLF_MISS) | of;
 				    if (po) {
 					po->flags |= CLF_MISS;
 					po->max += diff;

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

Re: Bug w/matching control

[Felix Rosencrantz]

Thanks, Sven, that seems to fix the problem with the characters being deleted,
but the insert positions are different than the other case, they should be
9:27, there seems to be an extra 11 in there.

When I run with the patch I sent before I'm getting a diff for
  line: {tst glibc}{-2.1.94-3.i386.rpm}
  COMPADD:{}
! INSERT_POSITIONS:{9:11:27}

-FR


__________________________________________________
Do You Yahoo!?
Get email at your own domain with Yahoo! Mail. 
http://personal.mail.yahoo.com/