> However it causes a segfault on load with the current unstable head. I > *think* this is bidi bug. At least, gdb tells me the crash is caused > by an invalid delete of fribidi chars in the field code in fp_Run.cpp.
I think I fixed it now, there was an out-of-bounds write in the call to fribidi (the behaviour of the function has changed since since the previous version we were using, it now 0-terminates all the output, it did not use to). However, I am still getting a strange assert in the fp_FieldRun::_defaultDraw() function. I have fixed a similar bug in ap_Dialog_Lists.cpp. Tomas
