This is now tracked in ticket #109 (http://trac.common-lisp.net/armedbear/ticket/109). Bye, Erik. On Fri, Oct 8, 2010 at 10:23 AM, Erik Huelsmann <ehuels@gmail.com> wrote:
Hi Ole,
On Thu, Oct 7, 2010 at 8:53 PM, Ole Myren Rohne <ole.rohne@gmail.com> wrote:
Hi ABCL-developers,
Thank you for releasing abcl-0.22!
I'm playing with AbclScriptEngine (with Electric VLSI) and found a tiny problem with config.lisp: EXT:MAKE-THREAD should be THREADS:MAKE-THREAD. Unfortunately the problem was unnecessary hard to debug because in the absence of TPL::REPL the debugger ends up quitting the entire application.
Thanks for your report! The EXT:MAKE-THREAD vs THREADS:MAKE-THREAD will be fixed in the next release.
I'm wondering if in that case it would be more useful to throw a Java error, eg like SYS::%DEBUGGER-HOOK-FUNCTION?
Interesting idea indeed. Especially since uncaught errors themselves cause Java applications to be terminated too. It's definitely something to be studied more. I'll file an enhancement ticket for it - sorry can't address the issue right now.
Regards,
Erik.