freenode/#clim - IRC Chatlog
Search
8:31:54
loke
beach: GPL specifically establishes that “linking” the GPL-licensened code with other cause, requires the other code to also be covered by the GPL.
8:32:15
loke
What “linking” even means in terms of Lisp isn't very clear, and as far as I know has not ever been established.
8:32:52
red-dot
As well, there is a lot of commercial entities backing open source, but nearly always with an Apache license.
8:33:03
loke
With that in mind, is obvious why many companies legal departments simply decide to avoid any risk by simply not allowing GPL at all, even though that's an overreaction.
8:34:35
jackdaniel
loke: check out this post: https://common-lisp.net/project/ecl/posts/ECL-license.html
8:36:07
beach
loke: That would be a problem for those companies only if they don't intend to make their software free.
8:36:11
jackdaniel
in this post I argue, that it doesn't really matter, because any alternative interpretation than what llgpl provides is pure nonsense and is no different than GPL (hence, why would someone add lgpl license)
9:32:43
loke
It would be neat if it was possible to just load the ASD and be done with it, but it isn't that easy. Right now I ahve to run their makefile and install all the datafiles. Then I have to comment out a form in the ASD file before I can load it.
9:34:33
loke
and admittedly, the fact that I _am_ able to load it and run it suggests that it's not too bad.
9:34:56
jackdaniel
well, ASDF is not pure Lisp (i.e if you have conforming implementation downloaded from the moon - it won't work)
9:35:08
loke
Look at the DEFMTHOD on line 15. I don't know what is the purpose of that. https://gist.github.com/lokedhs/e32ff6b1e9fedd851c25558847f86660
9:36:05
loke
Also, Maxima runs on a huge range of CL's. In fact, I beliebve it runs on literally all of them (even GCL, I htink!)
9:36:39
loke
Maybe they dropped GCL support... Because I'm pretty sure I've seen defmethod calls in there (outside of the .asd)
9:43:34
loke
This months, there have been a handful of fixes that applies to Alpha and S390 platforms
9:44:08
jackdaniel
in that case ECL and CMUCL are as dead as they are. I'm sorry that you think that way
9:46:17
loke
I've had this discussion with pjb in the past. With his definition of “dead”, an open source project can never die.
9:46:40
jackdaniel
I think I don't want to have this discussion, just signalling that declaring projects dead despite people working on them is simply rude
9:49:04
loke
That made me wonder how we can look at this in such different light. I think that it's simply a case of me not attributing such negative value to the word “dead”. I have projects that I occasionally fix bugs in, and even push new releases (gnu-apl-mode for emacs is the best example). I'd consider that project to be for all intents and purposes dead at the moment.
9:50:39
jackdaniel
my point is that you have projects which have contributors devoted to keep thing alive. declaring these projects (despite having people working on them) you basically claim, that their effort is insignificant and worthless
9:51:10
loke
jackdaniel: Hmm... Not at all. I _think_ I can see how you came to that conclusion though.
9:53:07
jackdaniel
well, if someone would tell me, that ECL is practically dead (despite me making my best effort to improve it) - I'd feel offended
9:53:23
loke
What I'm saying is that if people are actually working on CLISP, they've taken an active decision to not make them undead (yet?). They may have a good reason for it. Perhaps working on a project while keeping it dead is a good way to not put pressure on themselves.
9:54:37
loke
jackdaniel: I think it's best if I simply don't use that word then, because it obviously causes misunderstanding. I'm struggling to find a better word that describes the state of a project which receives no public updates.
9:56:39
loke
jackdaniel: In any case, please be assured that I have no intetion of being rude to anyone who devotes time to the public good. I'm an open source developer too.
9:57:54
jackdaniel
sure, I didn't think that was in purpose, just a carefree remark - maybe *that* is the problem?
9:58:24
jackdaniel
people declare things "dead" too easy - "Lisp is dead" - how many times did I hear this phrase from outside cl community?
10:03:15
loke
Far too many times. Although for some definitions of the word, perhaps Lisp is dead in somce contexts? :-)
10:32:37
scymtym
loke: the state you describe w.r.t. your apl emacs mode is sometimes called "maintenance mode". i don't know whether it also applies to gcl or clisp, but when applicable, it seems like a fair description
14:38:27
fittestbits
I was looking at the polyline/polygon code and was wondering why classes standard-polyline/polygon use a sequence of points for their internal representation instead of two coordinate sequences?
14:39:16
fittestbits
Seems like coordinate sequences would make drawing faster at the expense of creating points when when polygon-points is called.
16:32:17
karlosz
beach: i'm interested in adding SICL to clisp as a gsoc project with the clisp people. having seen clasp's integration, how much work do you think this would entail?
17:12:33
beach
karlosz: What parts of SICL are you thinking of adding? I would think the most useful part would be a Cleavir-based compiler, but perhaps you are thinking of some other things? A simple Cleavir-based compiler should not be too much work, but it would be wise to consult with Bike and drmeister to find out exactly what needs to be done.
17:27:14
onion
how come there is no macro being used here? : https://github.com/robert-strandh/CLIMatis/blob/master/X11-colors.lisp ^_^
17:30:09
onion
which could also have written the macro , and (use-the-macro '((black 0.0 0.0 0.0) (snow 1.0 0.98 0.98))) that makes (defconstant +SYM+ (make-named-color SYM VAL)) for each elem..
17:32:34
jackdaniel
and what would be the gain? a little more complicated sed/awk script for no real benefit (imho)
17:33:51
jackdaniel
(I mean - we all know macros, sure, but this is not code which is shared across basically anything, and having macro for a sake of having a macro and shaving a few bytes isn't very convincing)
17:40:03
jackdaniel
I think it boils down to the question what is the code and what is an intermediate representation
17:45:45
onion
heh why bother ? the original rgb.txt is a lot more simpler, why not just read that in directly ? just add sed and awk *and* cl to it... ?
17:46:35
beach
I for one would like to use it for an OS that does not have sed and awk. Perhaps the Mezzano people feel the same way.
17:48:05
onion
that would make sense. can't rgb.txt be parsed easily with CL? rather than also keeping another copy of it, say if the file gets out of sync for whatever reasons
17:50:51
beach
nyef: By the way, the partial inlining paper was accepted. And thanks to you, one of the referees remarked on the "elegant proof of correctness".
18:08:56
karlosz
beach: since the goal of gsoc is to have a deliverable by the end of the smmer, im thinking of splitting the work into achievable goals
18:09:21
karlosz
i know that getting the modular library functions like the sequence functions would be pretty easy
18:11:00
karlosz
i think that type informed optimizations would be very useful to have, since clisp doesn't have any