I propose to freeze the CVS on sunday 28th or monday 29th, and cut the
next release on 6th of July. (Hopefully making less of a mess of
things this time around!)
I also plan announce the freeze on sbcl-announce in the hopes of
getting more testers during the freeze-week.
I built cvs head (220.127.116.11) using sbcl 18.104.22.168 on a first generation
MacBook (Intel Core duo) running MacOS 10.5.7 after running clean.sh
Running run-tests.sh got an unexpected error in compiler.pure.lisp
I have pasted the output from
sh run-tests.sh compiler.pure.lisp
-- Sidney Markowitz
From: Paul Khuong <pkhuong@gm...> - 2009-06-26 13:05:37
On 26-Jun-09, at 3:02 AM, Sidney Markowitz wrote:
> I built cvs head (22.214.171.124) using sbcl 126.96.36.199 on a first
> MacBook (Intel Core duo) running MacOS 10.5.7 after running clean.sh
> Running run-tests.sh got an unexpected error in compiler.pure.lisp
> I have pasted the output from
> sh run-tests.sh compiler.pure.lisp
Fix is on its way.
2009/6/25 Nikodemus Siivola <nikodemus@...>:
> I also plan announce the freeze on sbcl-announce in the hopes of
> getting more testers during the freeze-week.
I've been thinking more about this. If there are no objections:
* When freeze starts, stick .rc1 at the end of version.lisp-expr. Make
a source tarball out of that and put it up on the release area.
* In the freeze announcement, ask for "no regressions here" messages
in addition to problem reports. Promise the massive boon of being
listed as a "release tester" in NEWS for the brave souls who do this.
Also include a download link to the release candidate in the freeze
* If/when commits are made during the freeze, increment the rcN
counter instead of the tertiary version number. Depending on the
testing and nature of fixes reports, possibly announce a new .rcN
version along with a new source tarball on the lists (no need to
announce all of them, I think.)
How does this sound?