1.5.21-1 DLL Loading Problem

[Rob Hatcherson]

All:

After installing 1.5.21-1 (cygcheck output attached) I've starting 
experiencing a DLL loading problem whose presentation is similar to that 
of a problem described by Gary Zablackis in a 12-Apr-2006 post with 
subject "Patch for silent crash with Cygwin1.dll v 1.5.19-4".  
Examination of dll_init.cc shows that some variation of the suggested 
patch is part of 1.5.21-1.


The general situation is that I have a program that depends on "B.dll", 
which in turn depends on "A.dll".  Both DLLs provide C++ classes.  The 
symptom is that an attempt to launch the program comes back to the shell 
prompt without doing anything, i.e. no error messages, no output, no 
anything else.  If I statically link the program then it always works.


Attempts to run in gdb in the dynamically linked case result in output 
similar to this:

Starting program: /cygdrive/c/WorkAreas/...

Program received signal SIGSEGV, Segmentation fault.
Error: dll starting at 0x4d1000 not found.

Program received signal SIGSEGV, Segmentation fault.

Program received signal SIGSEGV, Segmentation fault.

Program exited with code 030000000005.
You can't do that without a process to debug.


The line "dll starting at 0x4d1000 not found" seemed to suggest that I 
have a path issue, but by inspection I don't believe this is the case, 
plus if I move the DLLs in question out of the way I get a different 
response from gdb.


Google'ing hasn't yielded much yet.  The post I referenced above is the 
closest thing I've found so far to the symptom I'm seeing.

I've been trying to track this down with gdb, but haven't made much 
progress.  I put the cygwin1.dbg from the src distro into the /bin dir, 
which at least allowed some level of breakpoint setting, but so far none 
of the ones I've tried (dlopen, dll::init(), and a few others in that 
vicnity) have been hit.  I'm not too familiar with the cygwin source 
base, so I may be snooping around in the wrong spot.

I've also been working on reducing the scenario down to some very simple 
libraries that reproduce the symptom reliably, but this hasn't yielded 
anything worth reporting yet.


In the meantime, does anyone have any suggestions regarding what's going 
on, and/or some other technique to catch this problem?

Rob

-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: cygcheck.out
URL: <http://cygwin.com/pipermail/cygwin/attachments/20060726/cb1ccaec/attachment.ksh>
-------------- next part --------------
--
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/