freenode/#lisp - IRC Chatlog
Search
0:42:23
phoe
Implementors of #lisp: do we have anyone proficient with debugging ANSI LOOP? I am trying to fix ANSI-TEST failures on CCL.
0:42:48
phoe
One of them is the infamous (loop for x of-type (integer 1 5) from 1 to 5 collect x) that fails on safe code because (setq x 6) is happening behind the scenes.
0:48:12
White_Flame
"the loop terminates when the variable var passes the value of form2" Sounds like the bug is with the test. The actual variable is supposed to breach past the "to" value
0:51:21
White_Flame
hmm, no I think I might be wrong. That's from downto/upto. "to" specifically flags an ending value, not a threshold
1:04:24
phoe
I asked specifically about ANSI LOOP because this implementation of LOOP is what is used inside CCL.
1:10:43
phoe
pjb: dead code per se isn't invalid, it just won't be reached and a sufficiently smart compiler will optimize it away
1:15:55
White_Flame
I still think the test itself is badly written. Its declaration mislabes what ranges it asks LOOP to calculate for X
1:16:25
White_Flame
from CLHS: (let ((x 1)) (loop for i from x by (incf x) to 10 collect i)) => (1 3 5 7 9)
1:17:16
pjb
The value 6 is not of the expected type (integer 1 5). on "Version 1.12-dev (v1.12-dev.4-4-gd9740256) Darwinx8664"
1:18:21
White_Flame
in http://www.lispworks.com/documentation/HyperSpec/Body/06_abaa.htm is "stepped[1]" supposed to refer to the glossary with that [1]?
1:18:42
phoe
I ask because ECL also uses a modified ANSI LOOP and maybe it has fixed this issue already.
1:19:10
White_Flame
"step v.t., n. 1. v.t. (an iteration variable) to assign the variable a new value at the end of an iteration, in preparation for a new iteration."
6:37:40
jackdaniel
phoe: also I think you mean "MIT LOOP", not "ANSI LOOP", I'm not aware of any LOOP implementation released by ansi comitee ;)
7:46:27
no-defun-allowed
Is https://github.com/deadtrickster/safe-queue the repository that Quicklisp safe-queue is cloned from?
7:48:53
no-defun-allowed
There is a little typo in the lparallel backend: the word "mailbox" should be replaced with "queue" in the documentation and lambda list in queue-empty-p.
7:51:30
no-defun-allowed
Though this maintainer looks a lot more active then the last one I wrote a PR for.
7:53:38
no-defun-allowed
Would it be impolite to "bump" the last one if we haven't gotten a response in six months?
7:54:12
no-defun-allowed
https://github.com/jsmpereira/heroku-buildpack-cl/pull/9 is getting a little stale.
7:57:20
no-defun-allowed
Does this sound nasty in any way? "I hate to be rude, but it has been six months since I wrote this PR and haven't really gotten a response. It is fairly possible that users of this buildpack have had some troubles compiling programs as a lot of unportable extensions have changed significantly, as well as some portable ones too (uiop for example). Merging this would probably be a good idea."
7:58:30
jackdaniel
I've heard that practically everything before ", but" has its meaning reversed ;)
7:58:30
Cymew
I'd never heard of Heroku before. Does that mean it has been superceded or that I'm lagging behind? ;)
7:59:48
no-defun-allowed
Cymew: You put code on it, it passes it through a "buildpack" to dump an image, it runs the image. Not the worst serverless architecture I can come up with, but it took some adjustment to get used to.
8:00:47
no-defun-allowed
More so because only six languages, none of which I know how to write or wouldn't dare writing a server in, are properly supported and the rest of us just get API specifications.
8:01:21
no-defun-allowed
Those are good, but they don't make for good learning material. Like trying to learn a non-programming language by skimming a dictionary.
8:02:15
no-defun-allowed
It fills my niche of "host for broke people that want to host servers" fairly well.
8:36:35
minion
The URL https://gitlab.common-lisp.net/users/sign_in?secret=82a381b7 will be valid until 08:45 UTC.
9:07:01
gaqwas
I've started fiddling around with Common Lisp a few days ago and thought I could learn more by reading what's being discussed here and possibly asking questions now and then.
9:08:37
jackdaniel
from what I've been told you'd enjoy plenty of flamewars on the usenet around that time ;)
9:27:10
minion
The URL https://gitlab.common-lisp.net/users/sign_in?secret=8d02f622 will be valid until 09:30 UTC.
10:26:36
shidima
I set up a trace with sly, but I would like to step through the code line by line if possible
10:30:10
no-defun-allowed
You'll have the best experience with it if you evaluate (declaim (optimize (debug 3) (safety 3) (speed 1))) so all debugging information is preserved and no optimisations that could mess with what you expect to be evaluated are enabled.