On Friday 07 March 2003 11:03 am, Lars Gullik Bjønnes wrote: > Angus Leeming <[EMAIL PROTECTED]> writes: > | On Friday 07 March 2003 5:04 am, [EMAIL PROTECTED] wrote: > | > Hello, Angus: > | > > | > On Thu, 6 Mar 2003, Angus Leeming wrote: > | > > Hmmm. This rings a bell. > | > > CG did I see that you added a static boost::signal to the code? I > | > > believe taht older gcc compilers had problems with these static > | > > signals (generated incorrect code) but can't remember the details > | > > anymore. Anyway, if you did add a static boost::signal in your patch, > | > > then we devised a work-around. Could you let me know as I don't have > | > > your source to hand. > | > > | > Yes, I did. It is in src/frontends/WorkArea.h, > | > > | > "boost::signal2<void, int,char const *> workAreaCJK_IMprocess;" > | > | No, this is fine. Only static signals can be problematic. The problem > | lies elsewhere then. > > Are you sure that only static signals can be prolematic?
Well you were the one who dug deep in all this. I'm merely reporting what I remember. Anyway, the crash occurs in Toolbar::Pimpl::~Pimpl, called from boost::checked_delete; complain about a missing pure virtual function. Seen anything like this before? Angus
