freenode/#lisp - IRC Chatlog
Search
15:01:09
jmercouris
so let's say I make a function call that doesn't provide a literal value, but instead a symbol
15:01:42
jmercouris
but what if I want somethings to remain as symbols for the purposes of callbacks
15:02:15
jmercouris
the foreing code needs to be able to send a message to the lisp side and know to invoke call-back-to-invoke
15:09:36
beach
What Emacs does to indent DEFCLASS is really not very good. I am suggesting this way instead: https://pastebin.com/ThBGvF3v Opinions?
15:15:57
Shinmera
beach: How do you feel about inline indenting? As in, providing spacing between items in a line.
15:19:29
jcowan
so it is just a matter of giving the value of a keyword (pseudo-)argument an extra indent when it is fat
15:20:54
beach
Shinmera: I would have to think about that. I assume you mean aligning things in columns, right?
15:22:31
beach
jcowan: Indenting the slot options with respect to the slot name, and the value of the slot option with respect to the name of the slot option.
15:23:26
beach
Pierpa: In that case, nothing will change. But when the slot name has many characters in it, and when the initform is complicated, then something better could be done than what Emacs does.
15:25:53
Shinmera
beach: I generally avoid doing it, but only because the cost of fixing the indentation on a change is too high. If the editor could automate it, then that would fall away.
15:26:11
jmercouris
Shinmera: yeah, but anyone who doesn't have that editor has that same pain point
15:26:21
beach
Shinmera: It gets complicated then, because you may have one line that says `x very-long-value-for-that-very-short-name' and one line that says `very-long-name-for-a-really-short-value y'
15:27:50
Shinmera
beach: The column aligning would also be helpful for things like a series of let bindings or defvars, etc.
15:30:21
beach
Shinmera: Oh, I agree. But again, because of the risk of significantly longer lines when columns are aligned, it has to be a choice.
15:31:28
beach
jmercouris: If we do nothing better than the lowest common denominator, we won't make any progress.
15:37:14
beach
Shinmera: I think an indentation command like that would have to be applied to a region containing lines with the same number of "things" in each one, except possibly for terminating comments.
15:38:46
jmercouris
I get that it is still kind of catering to the lowest common denominator, but it will help ease the transition, assuming climacs may one day be the dominant ediitor
15:40:43
beach
jmercouris: Writing two or more editors in one would be too hard for me. So I don't think I will try any compatibility mode, neither for Emacs, for VIM, or for gedit.
15:41:50
beach
jmercouris: Again, I don't care if I am the only person using it. Therefore, making it easier for others to adopt is not a priority.
15:45:39
jdz
beach: Sorry, I have not read all of the backlog here, just looked at your paste link. Not sure what you mean by "use something like this"?
15:51:11
beach
jmercouris: Thanks. But my effort is compensated by papers that explain the innovative parts of what I write.
15:51:52
jdz
I don't remember having multiline or long initforms, but I'd still initform code on the same line as the :initform keyword.
15:52:46
jcowan
I also am writing an editor primarily for personal use, though perhaps some other people will like it
15:53:27
beach
jcowan: What technique does it use in order to indent Common Lisp code, and what information does it present the Common Lisp programmer with? :)
15:53:28
jcowan
It will have many non-features, such as syntax coloring, full-screen editing, single-keystroke commands, and so on.
15:53:58
jcowan
I am currently working out the command set, and haven't decided on an implementation language yet
15:54:54
beach
jcowan: I was talking about the user being a Common Lisp programmer who would want to see the editor help with information about the program.
15:55:47
jcowan
The most interesting thing will be how to do parenthesis-bouncing in the context of CLI editing
15:57:07
jcowan
I currently use the ex mode of vim, and will stay fairly close to that, but I also intend to add the structural regular expressions of sam, and I do drop into vi mode when I want to do paren-bouncing, so I want a replacement for that
16:22:42
jmercouris
when I run the following in a standalone sbcl it works https://gist.github.com/8210dad39133fa1f45cb981969cfb3de
16:23:15
jmercouris
this is obviously unacceptable, as I need to be able to test in slime, any idea why this might be?
16:25:34
flip214
jmercouris: can you get a thread list at that point, and look where they're hanging?
16:38:25
_death
I liked the s-libraries when I started writing lisp.. apparently he writes smalltalk now
17:49:35
whoman
jmercouris: btw, i wanted to ask you for a week or two, what project you were working on? i cant remember what you said about it exactly but it sounded very cool
18:33:34
jmercouris
right now I am making a very deep changing removing the FFI by implementing a server/client interface between a lisp core and a "dumb" gui frontend
23:22:14
fouric
Is there a way to get a bit more detail about *what* is taking up all that space in a SBCL (or any other implementation, really) core?
23:57:25
attila_lendvai
fouric: if you get to that point then there are even path-to-root functions on some implementation that can tell you about a specific object what is holding it from garbage collection
23:57:25
Colleen
attila_lendvai: frgo said at 2018.03.07 22:35:24: I just pulled from dev and now I see some strange messages regarding "git submodules" - please see https://gist.github.com/dg1sbg/ccebad1c7702fe6a359c344bce50368c - How d I deal with those? Any idea what I am doing wrong? Thanks for looking into this.
0:03:21
fouric
Do you have any particular links you can give me? All my searches are just yielding normal list-traversal functions
0:03:37
attila_lendvai
fouric: there's something new in sbcl that I haven't seen, iirc in sb-introspect, and a few years ago we wrote our own in hu.dwim.debug (also sbcl specific, but probably somewhat bitrotten by now)
0:08:18
attila_lendvai
fouric: actually, I think this is the more public API: SB-INTROSPECT:MAP-ROOT
0:10:15
attila_lendvai
scratch that, map-root seems to do something else. no idea what inspired its name, though...
1:27:31
trocado
Hi! I wonder if there's a simple way to tell split-sequence not to split sub-strings.