On Fri, 1 Aug 2008 17:16:26 +0200, "Matthias Andreas Benkard" <[EMAIL PROTECTED]> said:
> The program below, which does little more than raising an exception > from a method invoked by means of an NSInvocation object, crashes with > a SIGABRT by calling abort() from objc_exception_throw() as indicated > by the attached gdb output. > With either libgnustep-base1.16-libffi or a manually compiled stable > GNUstep Base (1.16.3), as well as with a current Subversion checkout > of GNUstep Base (both compiled using ffcall, not libffi), the program > instead handles the exception as intended: That's odd. Are there any differences between the way you compile GNUstep Base, and the way the package is built? e.g. what flags do you pass to configure? Do you have libffi-dev installed? With GNUstep Base 1.16.3, it will use libffi if it can find it, even if you say --enable-ffcall, unless you specifically say --disable-libffi. Anyways, this may all be moot, since you say it doesn't happen with libgnustep-base1.16-libffi. After lenny is released, we will be defaulting to using libffi, since upstream prefers it now. The reason we haven't switched yet is because we wouldn't be able to sufficiently test such a big change before the release of lenny. -- Hubert Chathi <[EMAIL PROTECTED]> -- Jabber: [EMAIL PROTECTED] PGP/GnuPG key: 1024D/124B61FA http://www.uhoreg.ca/ Fingerprint: 96C5 012F 5F74 A5F7 1FF7 5291 AF29 C719 124B 61FA -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
