From: William H. N. <wil...@ai...> - 2005-04-14 16:13:28
|
On Thu, Apr 14, 2005 at 04:15:32PM +0100, Christophe Rhodes wrote: > William Harold Newman <wil...@ai...> writes: > > >> Is this alright with you? Is there any major work that people are > >> aware of that won't be done a little early this month? It's been a > >> busy month already; going a few days early won't make it a small > >> release by any means :-) > > > > If no surprises, I'm happy with freezing on the 20th and releasing on > > the 24th. > > This is fine with me. The 24th happens to be the day that I will be > talking in Amsterdam at the European Lisp Meeting (provisional title: > "Recent Developments in SBCL"). I haven't started writing my talk > yet, so there is still time to decide to call the version which will > release on the 24th sbcl-0.9.0, if that is agreeable to people; we > haven't exactly solved all the things either in Bill's original "for > 0.9", the cliki "Road To Zero Point Nine" page, or even the stuff in > "for early 0.8.x", *sigh*, but I'm inclined to think that there are > reasonable marketing reasons for calling this release 0.9.0, as well > as the technical ones of having done 22 months of solid technical > work, and of being unable to count above a certain number of minor > revisions :-) Indeed, making it 0.9 sounds OK too. No obviously-more-natural cutoff date has suggested itself so far, and I agree that it's past time by several reasonable standards. > Here's my list of 0.8 highlights; if I've missed off your favourite > feature, shout! > > * new ports: > ** PPC/Darwin > ** X86-64/Linux > ** (host port) builds under clisp (sometimes ;-) > * compiler efficiency enhancements: > ** {un,}signed modular arithmetic (0.8.3, 0.8.?) > ** efficient 64-bit arithmetic (David Lichteblau: 0.9.0?) > ** dynamic-extent {&rest,closures} (x86-only: 0.8.10,0.8.19) > ** many many many individual improvements including > *** float/fixnum compares; > *** bignum operations; > * new features: > ** external-format > ** Unicode > ** linkage-tables > ** sampling profiler > ** package locks > ** compiler conditions > ** other contributed modules: asdf-install, sb-simple-streams, sb-md5 > * bug fixes: > ** too many to number. (Can someone run 0.8.0 on today's PFDsuite?) Ayup. I'd say that just a 64-bit port and Unicode are probably enough to justify 0.9; and even if those two features didn't exist, then just adding up all the other new features would also be enough. And if there weren't any new features to speak of, then just the improved standards compliance and stability, and the elapsed calendar time, would probably *also* be enough... Besides, April 19th is my 40th birthday, and if one can't participate in an 0.9 release before reaching 40, then better slightly-late than later.:-| -- William Harold Newman <wil...@ai...> PGP key fingerprint 85 CE 1C BA 79 8D 51 8C B9 25 FB EE E0 C3 E5 7C "Secondly, to celebrate my birthday. Or, I should say: OUR birthday. For it is, of course, also almost the release date of the freeware system for which I am release manager..." -- Bilbo v.2005 |