A Dissabte, 23 d'octubre de 2010, [email protected] va escriure: > On Fri, 22 Oct 2010 19:50:52 +0200 > > <[email protected]> wrote: > >http://alt.swiecki.net/j/poppler_2010.10.22.tbz > > Now I'm checking PDFs in this tarball. SIGABRT for > > > SIGABRT.PC.0x7ffff70a9a75.CODE.-6.ADDR.0x1102000000156.INSTR.cmp_rax,_0xff > fff000.pdf > > is designed behaviour. In poppler/Object.h, there is > a macro to check the object type and unwanted type > makes the program aborted, aslike: > > 41 #define OBJECT_TYPE_CHECK(wanted_type) \ > 42 if (unlikely(type != wanted_type)) { \ > 43 error(0, (char *) "Call to Object where the object was type > %d, " \ 44 "not the expected type %d", type, > wanted_type); \ 45 abort(); \ > 46 } > 47 > 48 #define OBJECT_2TYPES_CHECK(wanted_type1, wanted_type2) \ > 49 if (unlikely(type != wanted_type1) && unlikely(type != > wanted_type2)) { \ 50 error(0, (char *) "Call to Object where the > object was type %d, " \ 51 "not the expected type %d or > %d", type, wanted_type1, wanted_type2); \ 52 abort(); \ > 53 } > > This is consistent with the stracktrace attached in the > end of this message. I'm not sure if abort() is the best > way, but this is designed behaviour, not bug.
Yes this is a bug, it should check for isNum before calling getNum. Albert _______________________________________________ poppler mailing list [email protected] http://lists.freedesktop.org/mailman/listinfo/poppler
