freenode/#lisp - IRC Chatlog
Search
1:59:35
earl-ducaine
no-defun-allow: I was hoping to avoid hard-coding ASCII codes, which on reflection is silly. The (horizontal) tab character along with vertical tab, form feed, etc. is only meaningful within the context of ASCII
2:02:18
Nilby
In case of reality shifts one could say, #9=#.(loop for i from 0 below char-code-limit when (string-equal "TAB" (char-name (code-char i))) return (code-char i))
2:03:31
aeth
Nilby: tab must be present. http://www.lispworks.com/documentation/HyperSpec/Body/13_ag.htm
2:14:46
Nilby
The sad thing is I have trouble fitting my characters in 64 bits, so I imagine the future characters will be better as a digest hash or something.
2:17:39
no-defun-allowed
(name-char "Dog") in SBCL is funnier than it should be. Sure, lispms have Super, Hyper, Top, Greek, etc keys, but what about the Dog character?
2:22:12
earl-ducaine
aeth: Only #\Newline and #\Space must be present. The others are 'semi-standard';
2:23:45
aeth
It seems to me that they might not be mandated because they might not exist in some encodings
2:25:32
Nilby
#\dog is currently more tragic than funny for me, since it only come up properly in 2-3 out of like 15 different things, and when it does come up, it looks different an sometimes pathetic.
2:36:42
earl-ducaine
aeth: The motivation for the question was writing a lisp function compatible with the C function isspace. That requires \t \v \f and \r. So, \v (vertical tab) is the one that's problematic.
2:39:40
earl-ducaine
But in my case, there's no good reason I can think of for not using (code-char 11) as suggested above.
2:42:26
aeth
I personally rely on code-char being an ASCII superset (probably Unicode) even though it's not guaranteed. I'm not sure there's any real alternative.
2:45:57
earl-ducaine
moon-child: isspace is widely used in C parsing to determine whitespace characters, especially parsers. In my specific case writing writing a cl scanner that accepts the wayland protocol in exactly the same way as the model implementation.
3:05:15
earl-ducaine
aeth: cl-babel provides a compatability layer with respect to encoding. I was checking it out when I was investigating #\Tab. There are indeed comments with regard to implementation idiosyncrasies.
3:05:33
no-defun-allowed
ACTION uploaded an image: notfunny-abelson.png (125KiB) < https://matrix.org/_matrix/media/r0/download/matrix.org/QgqTdYHvSVqRQjrfslNsAakd/notfunny-abelson.png >
4:00:47
beach
contrapunctus: I think there is a better way to do optical music recognition. Trying to do it automatically is going to introduce too many errors. I would prefer an interactive approach where the operator is in charge, but the software helps out. We can discuss this later if you like.
6:50:22
contrapunctus
flip214: nice š³ Would you like to give it another go? I might be more motivated if I had someone working on it with me š
6:51:27
flip214
contrapunctus: some interest is still there, of course... time is limited, but we can discuss ideas via email (make a mailing list?)
6:51:39
contrapunctus
Although a mostly-interactive solution might actually be less work than a batch scanning + interactive proofreading solution.
6:52:19
flip214
well, I'd just write midi files and have the operator use whatever midi editor is prefered for fixing things up.
7:08:26
remexre
oof, and archive.org is giving me "The search engine encountered the following error: invalid or no response from Elasticsearch"
11:40:01
Xach
borodust: I'm having trouble using the stable tag of bodge-blobs-support. I'm getting "fatal: not a valid object name"
11:59:03
phoe
Xach: I suggest that you remove links to Quickdocs from http://blog.quicklisp.org/2020/09/september-2020-quicklisp-dist-update.html
12:00:06
phoe
I fixed a bug in someone else's code who downloaded a project from quickdocs instead of git
12:00:53
phoe
at some point I might become annoyed enough to actually go to war with all the abandoned fukamachiware that is still prominent all around the net
12:01:17
beach
Isn't quickdocs the one that gathers all the documentation strings (including those for internal stuff) but doesn't take into account real documentation?
12:01:18
phoe
quickdocs being one very good example of a website that was once goodā¢ and then got abandoned, making it horribleā¢
12:02:38
mseddon
and yeesh. can we all start using https, it's 2020, ssl certs are free and easy to set up. :P
12:04:02
jackdaniel
while I agree that https is rather a good thing, it is not that there are no issues with making it omnipresent standard
12:04:26
Xach
ACTION is still adding to https://github.com/search?o=desc&q=type%3Aissue+author%3Aquicklisp&s=created&type=Issues
12:06:09
mseddon
jackdaniel: that's a crap argument for not employing it though. Safety belts aren't 100% either.
12:06:20
jackdaniel
energy efficiency, bandwidch increase (also speed), complicated specification (i.e hard to get right to implement from scratch)
12:07:15
jackdaniel
the last one applies also to so called "web"; adding more and more crap to the "standard' makes it impossible undertaking to anyone who is not a corporation with deep pockets
12:08:06
jackdaniel
(and I don't think of my arguments as being crap, in fact I took some time to think about this topic)
12:14:46
mfiano
Also good, because it's one more reminder and one more straw to break more camels' backs and not to rely on fukamachiware
12:15:00
jackdaniel
virtual pointer in charming in action: http://turtleware.eu/static/paste/c9d36e0f-terminal-pointers.webm
12:19:01
mseddon
hopefully it's relatively easy to also add support for gestures like pinch etc that use multiple pointers.
12:32:21
jackdaniel
then I use escape codes (as for ecma-48) + few extension escape codes as for xterm
12:34:04
jackdaniel
here: http://turtleware.eu/posts/Charming-CLIM-tutorial-part-2--Rethinking-The-Output.html
12:37:16
jackdaniel
the target goal is to write a clim application and be able to run it over: terminal, web and natively (or even switch at runtime;)
12:37:32
Nilby
jackdaniel: Wow, Thanks for that blog post. I wrote a lisp-only curses-like thing but it's very slow, so I'm envious.
12:38:38
Xach
jackdaniel: do you have a cooked terminal for the debugger? is the animated ui terminal separate from the rest or is it the only interaction terminal?
12:41:12
Xach
jackdaniel: say you have a bug in your code when you move the cursor somewhere, and it would normally trigger the debugger. what happens?
12:49:07
Xach
schweers: cooked means the terminal interprets input and output into higher level control sequences
12:50:41
Xach
jackdaniel: "animation" is a poor choice of words. watching it as a movie made me think of an animated movie, not a recording of an interactive session.
12:51:57
jackdaniel
I've recorded my interactions with the terminal with the application "peek" which captures part of the X11 screen, the rest are interactions directly with the terminal