On Fri, May 27, 2011 at 10:22 AM, anuj agarwal <[email protected]> wrote: > Hi Chris, > > Thanks for replying. > > I have the source code. I have moved my source code from Unix to Linux, > compiled the code and then facing this problem. > > I tried gdb, but it is showing me the line number where the program is > dumping core. I don't see any problem with that line because its a simple > integer variable declaration. Now i tried printing the variables passed to > the function which is failing and they are also fine. > Main problem is that our code runs through tuxedo so i can't create the > binaries using gdb and can't put debugs on line but only attaching the gdb > with binary process id. > > The program runs fine in Unix and also not showing any warnings while > compiling in linux.
It might have happened on a different thread. IIRC gdb is fantastically stupid about showing you the stack trace for the right thread when a non-main thread is the segfault. I don't think it's really probable that gdb would be unable to find the problem. It's a very powerful tool, capable of grabbing stack traces not only from threads but from libdispatch containers owned by the OS as well. So I'm fairly certain that it's still detectable in gdb, though I'm not at all familiar with Tuxedo, so it could be something well beyond my understanding. :) Good luck! (I'm about out of advice for the situation, so I'll be quiet now). -- Registered Linux Addict #431495 For Faith and Family! | John 3:16! fsdev.net | 0x5f3759df.org | chrismiller.at -- You received this message because you are subscribed to the Linux Users Group. To post a message, send email to [email protected] To unsubscribe, send email to [email protected] For more options, visit our group at http://groups.google.com/group/linuxusersgroup Please remember to abide by our list rules (http://tinyurl.com/LUG-Rules or http://cdn.fsdev.net/List-Rules.pdf)
