Hi Martin,
On Wed, Jul 16, 2008 at 1:59 PM, Martin Krauskopf
<[EMAIL PROTECTED]> wrote:
> Thomas E Enebo wrote:
>> We plan on trying to release JRuby 1.1.3 by the end of the week
>> (hopefully on Thursday). If people have "must-haves" or
>> "it-would-be-dreadfully-nice-if" issues, then please let us know.
>> We have a three or so days to square things away. One issue is
>> adding 64 bit support for Linux POSIX functions...others?
>
> RubyKernel#proc and RubyModule#ancestors (with a bunch of other methods)
> changed their signature in rev. 7131:
>
> http://fisheye.codehaus.org/changelog/jruby/?cs=7131
>
> RubyArray ancestors()
> RubyProc proc(IRubyObject recv, Block block)
>
> becomes
>
> RubyArray ancestors(ThreadContext context)
> RubyProc proc(ThreadContext context, IRubyObject recv, Block block)
>
> Any chance they will be taken back as deprecated for JRuby 1.1.3?
Ah, I agree that we shouldn't break the API like that, and the
deprecated methods should stay.
> I've changed call:
>
> RubyProc proc = RubyKernel.proc(recv, block);
>
> to:
>
> ThreadContext ctx = ThreadContext.newContext(recv.getRuntime());
> RubyProc proc = RubyKernel.proc(ctx, recv, block);
>
> Is that right? If yes, the original method could be taken back.
How about this one:
ThreadContext ctx = recv.getRuntime().getCurrentContext();
> If they are not taken back it would mean that already released
> extensions like (j)ruby-debug-base will not work with JRuby 1.1.3. I've
> supposed that during minor version increase (1.1.x) the API is kept. Is
> that right assumption so above is a bug? I would file it.
Please file a blocker bug with 1.1.3 target release. That way, we'll
deal with it one way or another before the release.
Btw, are those two methods the only ones that you see breaking your code?
Thanks,
--Vladimir
---------------------------------------------------------------------
To unsubscribe from this list, please visit:
http://xircles.codehaus.org/manage_email
