zsh 4.3.0 dev 1 with cygwin1.dll 1.5.15

zsh 4.3.0 dev 1 with cygwin1.dll 1.5.15

[S. Cowles]

I don't have a clear perception which mailing list is the more appropriate for 
this question, so I've sent a version of this message to the cygwin users' 
mailing list and to the zsh users' list.  Any help would be appreciated.

The problem is that I have had no success running zsh 4.3.0 dev 1 on the last 
two releases of cygwin1.dll, 

CYGWIN_NT-5.1 host 1.5.14(0.126/4/2) 2005-04-01 13:40 i686 
CYGWIN_NT-5.1 host 1.5.15(0.127/4/2) 2005-04-18 12:20 i686

The builds of zsh proceed with no undue problems, but test A05 fails.  The 
following is the message from running the cvs version of 4.3.0 dev 1 (pulled 
about 10h00 PDT, GMT +0700, 19 apr 2005) against cygwin1.dll 
1.5.15(0.127/4/2):

./A05execution.ztst: starting.
Test ./A05execution.ztst failed: bad status 1, expected 0 from:
  path=($ZTST_testdir/command.tmp/dir1
        $ZTST_testdir/command.tmp/dir2
        .)
  tstcmd
  path=($storepath)
Error output:
      6 [main] zsh 2052 fork_parent: child 3084 died waiting for longjmp 
before initialization
(eval):4: fork failed: interrupt
Was testing: path (1)
./A05execution.ztst: test failed.
 249108 [main] zsh 2052 fork_parent: child 3640 died waiting for longjmp 
before initialization
ZTST_cleanup:2: fork failed: no such file or directory


When I try to run this build of zsh, every call to a non-builtin resource 
results in a fork failure. I have asked the cygwin mailing list if there 
might be some user configurable resource that would fix this problem.  On the 
other hand, the cygwin release of zsh 4.2.4 does run under this latest 
release of cygwin1.dll. Before I jump in, I wondered if anyone had already 
run into this problem and had any pointers to causes.

Re: zsh 4.3.0 dev 1 with cygwin1.dll 1.5.15

[Peter A. Castro]

On Tue, 19 Apr 2005, S. Cowles wrote:

> I don't have a clear perception which mailing list is the more appropriate for
> this question, so I've sent a version of this message to the cygwin users'
> mailing list and to the zsh users' list.  Any help would be appreciated.

This is really a Cygwin issue.  If you can wait, I'm building this version
right now and maybe I'll be able to give you a better answer as to what's
going on.

> The problem is that I have had no success running zsh 4.3.0 dev 1 on the last
> two releases of cygwin1.dll,
>
> CYGWIN_NT-5.1 host 1.5.14(0.126/4/2) 2005-04-01 13:40 i686
> CYGWIN_NT-5.1 host 1.5.15(0.127/4/2) 2005-04-18 12:20 i686
>
> The builds of zsh proceed with no undue problems, but test A05 fails.  The
> following is the message from running the cvs version of 4.3.0 dev 1 (pulled
> about 10h00 PDT, GMT +0700, 19 apr 2005) against cygwin1.dll
> 1.5.15(0.127/4/2):
>
> ./A05execution.ztst: starting.
> Test ./A05execution.ztst failed: bad status 1, expected 0 from:
>  path=($ZTST_testdir/command.tmp/dir1
>        $ZTST_testdir/command.tmp/dir2
>        .)
>  tstcmd
>  path=($storepath)
> Error output:
>      6 [main] zsh 2052 fork_parent: child 3084 died waiting for longjmp
> before initialization
> (eval):4: fork failed: interrupt
> Was testing: path (1)
> ./A05execution.ztst: test failed.
> 249108 [main] zsh 2052 fork_parent: child 3640 died waiting for longjmp
> before initialization
> ZTST_cleanup:2: fork failed: no such file or directory
>
>
> When I try to run this build of zsh, every call to a non-builtin resource
> results in a fork failure. I have asked the cygwin mailing list if there
> might be some user configurable resource that would fix this problem.  On the
> other hand, the cygwin release of zsh 4.2.4 does run under this latest
> release of cygwin1.dll. Before I jump in, I wondered if anyone had already
> run into this problem and had any pointers to causes.
>
>

-- 
Peter A. Castro <doctor@fruitbat.org> or <Peter.Castro@oracle.com>
 	"Cats are just autistic Dogs" -- Dr. Tony Attwood