Alban Bedel <albeu@free.fr> writes:
The bug seems to be pretty apparent from the backtrace, I'm just not sure how to track down which module the string came from. could you add the following to your backtrace : up print *opt
so i could know wich option is causing this bug. It's probaly an option related to dvdnav, smb or live.com.
Indeed, "dvdnav". (gdb) p *opt $1 = {name = 0x8376297 "dvdnav", p = 0x83762a0, type = 0x842e3e0, flags = 0, min = 0, max = 1, priv = 0x0}
That doesn't make sense at all as i have a different cpu, a different compiler and don't have dvdnav, smb and livecom. We ask that to know the exact compile environment that you used. ... Sure that i probably need to reproduce it locally to fix it. But for the same reasons as above getting your binary make no sense.
Well, yes, it wouldn't necessarily be easy. I could compile statically to remove the library dependencies, and could enable run-time cpu detection. But then it would get even bigger. A lot bigger I expect. But the original point of my statement was exactly to explain that there wasn't much more that I could practically provide to help. Actually I'm a bit embarrassed, I already had printed *opt in gdb but cut it off when I sent the e-mail. I don't know what I was thinking there. -- greg