Hello,
I plan to take the current slime from cvs (taken at the moment I post this) and call it a release.
Given the many features and fixes added since 1.3.1, I propose calling/numbering it
slime 2.0
If there aren't any objections, I will publish it in about 24 hours.
Regards, Mario.
"Good news, everyone! I've taught the toaster to feel love!" [1]
Can we get a last few patches in for some compatibility bugs? Maybe we then want to have it as FAIRLY-STABLE for a few days to make sure they don't introduce a regression?
I only followed the 1.2.1 conversation cursorily but wasn't it launched by some complaint over unstable named release? Let's try to make a good one here; it's the big Two Point Zero!
Then we can bump the version number with big fanfare and hooplah on Planet Lisp, Slashdot, and Reddit. Get some good press for our favorite development environment.
Anyone know who maintains (is it dead) the Lisp In a Box? Maybe try to get a simultaneous release?
Problems and patches:
Slime CVS has been broken on XEmacs for several weeks now (due to use of assoc-default). Originally reported with bugs: http://common-lisp.net/pipermail/slime-devel/2006-April/004737.html Better fix after discussion: http://common-lisp.net/pipermail/slime-devel/2006-April/004757.html This alone is fixed with the attached patch slime-filename-translations.diff
Slime CVS won't work on some windows setups (not accounting for :device). http://common-lisp.net/pipermail/slime-devel/2006-April/004736.html This alone is fixed with the attached patch swank-source-files.diff
Slime-loader doesn't work nicely with CLC. I don't know too much about it; I came across this patch tracking down another problem. http://common-lisp.net/pipermail/slime-devel/2006-March/004664.html The patch swank-loader-ALL.diff wraps up the patches given in that thread and my separate patch for windows compat. I have only tested the #-common-lisp-controller branch of this;
All of those patches together are given in patch nb-4-12-06.diff
I think these are fairly stable. I have tried to test them with X and Gnu Emacs on WinXP + ACL7-trail, and Debian2.6 + SBCL 0.9.10 without hitting any problems.
Regards, Nathan Bird
[1] I don't know, it just seemed appropriate somehow. http://en.wikiquote.org/wiki/Futurama
-----Original Message----- From: slime-devel-bounces@common-lisp.net [mailto:slime-devel- bounces@common-lisp.net] On Behalf Of Mario S.Mommer Sent: Wednesday, April 12, 2006 5:36 PM To: slime-devel@common-lisp.net Subject: [slime-devel] slime release
Hello,
I plan to take the current slime from cvs (taken at the moment I post this) and call it a release.
Given the many features and fixes added since 1.3.1, I propose calling/numbering it
slime 2.0If there aren't any objections, I will publish it in about 24 hours.
Regards, Mario.
slime-devel site list slime-devel@common-lisp.net http://common-lisp.net/mailman/listinfo/slime-devel
"Nathan Bird" nathan@acceleration.net writes:
"Good news, everyone! I've taught the toaster to feel love!" [1]
Can we get a last few patches in for some compatibility bugs? Maybe we then want to have it as FAIRLY-STABLE for a few days to make sure they don't introduce a regression?
Ok, fair enough.
Let's wait until after easter then...
Regards, Mario.
"Nathan Bird" nathan@acceleration.net writes:
Anyone know who maintains (is it dead) the Lisp In a Box? Maybe try to get a simultaneous release?
The Lisp In a Box on cl-net is, as far as I know, dead. As in, the maintainer, Matthew Danish, stopped responding to emails quite some time ago. (Or maybe he just hates me.) My Lispbox (http://www.gigamonkeys.com/lispbox/) is not dead though I don't update it as aften as I might like. I mostly just include the CVS head of SLIME from whenever I update since the SLIME releases are few and far-between. So if you release 2.0 and I get around to updating the Lispboxen, it'll get a 2.0 plus whatever version of SLIME.
-Peter
Mario S.Mommer wrote:
I plan to take the current slime from cvs (taken at the moment I post this) and call it a release.
Given the many features and fixes added since 1.3.1, I propose calling/numbering it
slime 2.0
[...]
I'm pretty new to Slime development, but could you tell me which CVS branch you intend to bless as 2.0? Do you mean what is now in HEAD?
If it is, I'd like to report a problem with Slime + SBCL 0.9.11 + OS X/PPC regarding autodoc. I tried looking into a bit yesterday without much success, but if there really is an impending release, I'll try a bit harder (I also "maintain" the DarwinPorts version of Slime, which would be nice to update from slime-1.2.1).
Mark Evenson
* Mark Evenson [2006-04-14 16:31+0200] writes:
I'm pretty new to Slime development, but could you tell me which CVS branch you intend to bless as 2.0? Do you mean what is now in HEAD?
Yes, the main trunk.
If it is, I'd like to report a problem with Slime + SBCL 0.9.11 + OS X/PPC regarding autodoc. I tried looking into a bit yesterday without much success, but if there really is an impending release, I'll try a bit harder (I also "maintain" the DarwinPorts version of Slime, which would be nice to update from slime-1.2.1).
If that's the "Error: (void-variable name)" bug then it should fixed with Greg's patch.
Helmut.
Helmut Eller wrote:
- Mark Evenson [2006-04-14 16:31+0200] writes:
I'm pretty new to Slime development, but could you tell me which CVS branch you intend to bless as 2.0? Do you mean what is now in HEAD?
Yes, the main trunk.
If it is, I'd like to report a problem with Slime + SBCL 0.9.11 + OS X/PPC regarding autodoc.
[...]
If that's the "Error: (void-variable name)" bug then it should fixed with Greg's patch.
Indeed it was, and the patch committed to HEAD fixed it. Things are now fine from what limited testing I am able to do towards blessing HEAD as slime-2.0.
Thanks for the help.
Mark Evenson
Dear Slimers,
Recently, (well, OK, not that recently, but I only now finally got tired enough to report it) tracing broke in some situations under LispWorks. In something like this:
(defmethod foo :after ((bar blah)) (bang bar))
If point is on `foo' and you hit C-c C-t Slime offers to trace (:defmethod foo :after (blah))
This may work in other lisps, but it doesn't in LW.
In LW, the following _would_ work, however: (trace (method foo :after (blah)))
Tracing of FLETs and LABELs just doesn't work, AFAIK, so I never get bitten trying to trace those.
I wonder if there's some way of changing what gets sent to the underlying lisp based on some choice of implementation variable. The work seems to be done in SLIME-EXTRACT-CONTEXT, and there are no current hooks.
I have a local patch, but it basically clobbers the whole functionality and is obviously not suitable for slime as a whole.
Cheers,
Alain Picard
Alain Picard Alain.Picard@memetrics.com writes:
Dear Slimers,
Recently, (well, OK, not that recently, but I only now finally got tired enough to report it) tracing broke in some situations under LispWorks. In something like this:
i've been using (but for some reason haven't commited) this for openmcl:
(defimplementation toggle-trace (spec) "We currently ignore just about everything." (ecase (car spec) (setf (ccl::%trace spec)) (:defmethod (ccl::%trace (second spec))) (:defgeneric (ccl::%trace (second spec))) (:call (toggle-trace (third spec))) ;; mb: FIXME: shouldn't we warn that we're not doing anything for ;; these two? (:labels nil) (:flet nil)) t)
I wonder if there's some way of changing what gets sent to the underlying lisp based on some choice of implementation variable. The work seems to be done in SLIME-EXTRACT-CONTEXT, and there are no current hooks.
personally i think it's cleaner to just implementation toggle-trace admititng that we don't actually do much.
-
Alain Picard -
Helmut Eller -
Marco Baringer -
Mario S.Mommer -
Mark Evenson -
Nathan Bird -
Peter Seibel