On Sun, Oct 24, 2010 at 12:56:39PM +0100, Caolán McNamara wrote: > This sucks, looks like catch throw and catch catch don't work on NetBSD. > Or maybe its too old a gdb or something. Maybe add a "break main" and > reorganize the gdb script in the hope that it needs libstdc++ to be > loaded before it'll work. i.e. something like the (untested) attached.
"br main" let it find the symbols. > The registering step should be effectively dlopening the modules that it > wants to register, running the component_writeInfo function and friends > in order to record into a registry file which .sos implement/support > what named services. So manual debugging like adding some printf around > the dlopen/dlsym calls in sal can help find what is the last .so to be > dlopened, and adding/debugging the component_writeInfo etc calls in > that .so can see if they're getting called at all, and sort of narrow > down where things are going wrong. I added some debugging printfs around dlopen/dlsym in sal/, as suggested, and found that it was just another instance of rpath problems (finding pyuno) which I currently just work around -- I added a workaround for that and could continue my build :) Thanks for your explanations and suggestions! Thomas _______________________________________________ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
