> On 05/03/2012 12:55 PM, Timothy Pearson wrote: >>> On 05/03/2012 07:38 AM, David C. Rankin wrote: >>>> On 05/03/2012 01:53 AM, Mag. Dr. Nikolaus Klepp wrote: >>>>> This might be a stupid question, but did you try to run kwirte with >>>>> valgrind? >>>>> >>>>> Nik >>>> >>>> There are no stupid questions when it comes to "did I do something?" >>>> The >>>> answer >>>> is probably no... I'm a great monkey on this side of the keyboard, but >>>> I'm not a >>>> career developer, so you guys that have ideas of what needs to be >>>> done, >>>> just ask >>>> and I can carry it out on this end. I'll give valgrind a go and let >>>> you >>>> know as >>>> well as the g++ command suggested by /dev/ammo42. >>>> >>> >>> OK, >>> >>> What tool in valgrind is the one we want? I tried cachegrind and that >>> didn't >>> produce anything: >> <snip> >> >> I would use the default memchecker tool, as my initial guess would be >> memory corruption based on the widely varied backtraces. >> >> Tim >> >> > > Tim, Doc, will do, > > I should have more time this evening to work with valgrind. I have > little > familiarity with it, so I'll just run it successively with the various > tools and > post any output I get. > > Any other tests you can think of, just let me know. > Just make sure you have the debugging symbols installed so that we can get line numbers, and that you have compiled from either the latest GIT head or a recent revision, preferably one with the gcc 4.7 fixes in place. It would be a good idea to post the valgrind output to the bug report. Tim