freenode/#lisp - IRC Chatlog
Search
8:48:57
adlai
https://freenode.irclog.whitequark.org/lisp/2020-10-15#28144452 << pillton, I once heard a theory claiming a paradoxical effect, wherein the surprise bestsellers that were failed by several editors, yet benefitted from -- grey-hat byte-stuffing, for lack of words -- still get incredible reviews because the critics just can't get past the tedia
8:51:12
adlai
; although that does not imply that the book is as terrible as you'd expect, from a glowing review by a paid professional critic.
12:35:06
jackdaniel
let's say taht my condition inherits from simpe-error and adds a new slot/initarg foo/:foo, is there a way to propagate value of foo to :format arguments? (without assuming that condition is not a standard class)
12:36:05
jackdaniel
(of course not by the caller of make-condition/error in the place where the condition is defined)
12:37:42
jackdaniel
as in (define-condition foobar ((foo :initarg :foo)) (:default-initargs :format-control "foo ~s not found" :format-arguments (list <whatever-was-passed-as-:foo>)))
12:39:46
phoe
AFAIK you need a custom constructor for that, so, (defun make-foobar (foo) (make-condition 'foobar :foo foo :format-arguments (list foo) ...))
12:41:00
phoe
I mean, setting portability aside for a moment, you can use INITIALIZE-INSTANCE :AFTER on a condition object but only if you use MAKE-INSTANCE instead of MAKE-CONDITION
12:44:29
phoe
we could try to turn this behavior (MAKE-CONDITION that obeys the MOP initialization protocol) into an implementation extension and then complain at SBCL that CCL, ECL, Clasp, ABCL, CLISP, LW, and ACL already support it
12:47:08
adlai
Xach: a potential collaborator wants to use the AGPL; there is a remote possibility that chanl, which is a dependency, will be the subject of patent trolling lawsuits after I, and a potential collaborator, do an unhealthy amount of studying from various open-source implementations of similar ideas.
12:47:49
adlai
there should be an "and" after the semicolon there; those are two separate instances of the risk of patent trolling going from zero to epsilon.
12:48:44
adlai
ACTION is tempted, around the two-thirds likelihood, to add some sorta "Please neither copyleft nor copyright your proposed contributions", although this is quite off-topic for here.
12:49:39
adlai
ultimately, what seems best to me right now is for scalpl itself to be outside of the default dist, although it should not require any software outside of the default dist.
12:50:20
adlai
as I've mentioned before, the current code is also replete with practices ranging from mediocre to worst.
13:06:31
adlai
ACTION should also take about a decade or few to get enough familiarity with the retail investor platforms so that scalpl can speak the protocol of at least one regular stock exchange
13:08:20
adlai
so far, it has been mostly speaking to various 'greysuit' digital currency exchanges, during the happy years before they get love letters from government agencies; it is a fun game, and quite useful for a certain class of people, although still - quite a narrow niche.
13:21:39
adlai
phoe: I explicitly wrote "public domain", and never included any copyright notice in the scalpl files, because I do not consider that tool an innovation, by this point in human history.
13:23:56
adlai
if the specific names of classes that I divined out of smoke, coffee grounds, and conversations with my dog, along with the various inconsistencies of imprecise arithmetic found in that repository, happen to be considered an invention... that's someone else's problem :)
14:56:24
dbotton_
phoe - will your book come out with a kindle edition when it is released in Nov? (I want to get a copy but want digital)
15:11:20
adlai
of course, I forgot that other hard problem: "... and the baud rate of the patent-pending low-power high-contrast pixel-based paper-mimicking anisotropic crystal array"
16:26:34
adlai
well here's a cool argument, if there ever was one: https://github.com/fukamachi/event-emitter/pull/11#issue-471842984
16:29:40
adlai
ACTION is deep within "usual approach useless here" territory; "fail early, often" + condition system = good thing; lighter diff = good, along different ethics; maintainer ... has not abandoned lisp, although, is busy.
16:34:50
adlai
ACTION has edited the unwitting competitor, "pull/8", to remove the link to the author of the issue reporting the problem, to minimize the number of links.
16:35:54
adlai
the author is still named in the commit message itself, and is likely to receive an automated notification if and when either one of the pull requests is accepted.
16:37:44
adlai
I intentionally do not include the author's name in these messages, to avoid a distracting notification; the author can hunt down this conversation in the maze of automated notifications, if the issue is worth voting upon.
16:40:08
adlai
furthermore, if my pull request is preferred over Zulu-Inuoe's primarily due to the commit message, instead of the code itself, I will be significantly less likely to perform such actions again.
17:24:47
dbotton__
phoe great, just no presale option on amazon, so hopefully will be there mid-nov when book comes out
17:26:06
dbotton__
What is the easiest gui library to use to do some playing with on windows? starting to work on so less trivial things to bring me to speed
18:18:00
rpg
@phoe: for some reason when I googled your book, the site also thought I was in Spain. I suspect it has something to do with the redirection from Google.
18:38:45
Xach
I don't know what's available for racket-specific help, but I think there is a #scheme channel
19:12:38
_death
so it appears cffi now takes responsibility for defining a size_t-compatible type :size.. nice
19:29:52
luis
_death: another long overdue thing someone should add to cffi is an API to zero memory
19:39:09
_death
but these can easily be defined by a user.. while size_t needs a grovel (although it looks like the current implementation makes "pretty safe bets")
20:03:55
jmercouris
what I don't like is the prevalent usage of it, and all the problems it leads to
20:04:17
jmercouris
C is fine when you MUST be on the bleeding edge of performance and if you are a genius like Dennis Ritchie