Mailing-List: contact cygwin-help AT cygwin DOT com; run by ezmlm
Sender: cygwin-owner AT cygwin DOT com
Mail-Followup-To: cygwin AT cygwin DOT com
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain; 	charset="iso-8859-1"
Subject: RE: Bash login breaks if too many environment variables are set
Date: Fri, 16 Sep 2005 15:55:14 +0200
Message-ID: <67B13185D126644BA1AC929BE1F942AD7A5ADD@dekomplm001.net.plm.eds.com>
From: "Hommersom, Fred" <fhommers AT ugs DOT com>
To: <cygwin AT cygwin DOT com>
Content-Transfer-Encoding: 8bit

Yes sure. You can see this in the header of the dump its says:
DLL version:  1005.19, api: 0.138
DLL build:    20050916 00:00:39SNP

In order to be sure that we are talking about the same things:
I have all these variables in DOS and start bash from a CMD window with command
c:\cygwin\bin\strace -o fhbashtrace.txt c:\cygwin\bin\bash 

As indicated in previous mails the stackdump occurs only with an exact number of characters
for all environment names and values together. My names and values have sizes of 1-1000 characters. No extremes.
But this dll behaves differently from the "production".
Now if I have one character less then the 'dumping' number some commands (e.g. ls) works, others don't.
I did not find a pattern in it. I'll keep trying.

But if I add 10 characters to an enviroment variable ( so 10 more then the number that gives the dump)
then bash exits immediately without dump. Strace gives:
   27     324 [main] bash 3024 set_myself: myself->dwProcessId 3024
   23     347 [main] bash 3024 time: 1126876171 = time (0)
  422     769 [main] bash 3024 environ_init: 0x10010238: !C:=C:\Dat
No more output here, no stackdump nothing. just exit


It looks to me as if a buffer or stack is reused if some maximum is exceeded with effect that the system sometimes works.
Fred

-----Original Message-----

Sure you've used the latest snapshot DLL?  I tried to reproduce this with the latest snapshot, as well as a self-build DLL from CVS and I can't reproduce this behaviour.  My test environment consisted of 1400 variables with a size of 98K, one of the variables taking roughly 31K alone.
However, what's strange in your strace output is this:

>   442     785 [main] bash 3504 environ_init: 0x10010238: X$
>    37     822 [main] bash 3504 environ_init: 0x10010248: 甔$
>    97     919 [main] bash 3504 handle_exceptions: In cygwin_except_handler exc 0xC0000005 at 0x610D6971 sp 0x22EE64

The first two lines should contain some valid environment entries, but both of them look broken.  There's no hint why this is, of course.

I observerd that tcsh doesn't like variables with a length of 31K, though.  ash, bash, zsh and pdksh could handle that long environment varibale just fine, tcsh on the other hand printed this:

  $ echo $VERY_LONG_ENV_VAR
  Word too long.


--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/