libera/#clim - IRC Chatlog
Search
9:56:40
selwynning
which is not my favourite task, but hopefully much of the code can be reused for more interactive things later on
11:00:25
selwynning
is there a convenient way to draw a filled white circle with a black outline? or are two calls to draw-circle needed
11:04:52
gilberth
selwynning: You need two calls. In CLIM a shape is either filled or stroked, never both at the same time.
11:13:30
gilberth
I see. I am not a fan of "modern" IDEs. I am quite comfortable with Emacs and SLIME.
11:15:07
beach
But what you are seeing in that demo is essentially an Emacs-like editor with a very detailed analyses of the code, at typing speed.
11:18:03
gilberth
Oh, I have a whole host of things. Begins with my RE engine and scanner generator, which is a major roadblock as most other things depend on that. Then I have my C parser, which can deal with all header files I can find, fine. An updated CADR emulator. And then of course Free-CLIM. And besides a new HTMLified ANSI-spec. Quite a list.
11:22:03
beach
You might want to know that scymtym managed to parse the dpANS and he can generate HTML too. The point was to make it possible to add "issues" for WSCL.
11:22:35
gilberth
For the most part, it is perfectionism. And me not having the energy to defend my design. Or the fear that I may face a need to do that.
11:23:55
gilberth
beach: Well, that's easy. I focused on typography to begin with. Adding links and annotations shouldn't be hard. Here is chapter 3: <https://lispcafe.org/z.html>
11:25:35
beach
No, I don't. I downloaded your HTML files at some point when it was not clear whether you were going to stick around. But I don't think it's public. And even if it is, it is not annotatable.
11:27:25
gilberth
It's public and Google finds it, sometimes even as the top result. And the latter concerns me, hence I ask you kindly to exclude that by means of robots.txt.
11:29:34
gilberth
Meanwhile, I will place <LINK REL="CANONICAL" ..> to tell Google that my version is the original. In case you rewrite URLs, I ask you not touch those. And then, only then, I can move the URL to something not depending on dyndns.org and not being on port 8000 and being HTTPS on a proper host.
11:30:12
selwynning
i am not familiar with the issue but it seems as if you are expecting google to behave nicely
11:31:12
gilberth
selwynning: Google respects both robots.txt an LINK REL=cannonical. I just ask beach to respect that and play with the rules.
11:31:19
beach
Whenever something web-related comes up, I am totally lost, and if I try to actually do something, it means days of frustration because I don't know what I am doing.
11:32:54
gilberth
beach: robots.txt is a file which tells web crawlers which pages they may crawl and index. See e.g. http://bauhh.dyndns.org:8000/robots.txt
11:33:42
gilberth
It excludes the /edit/ URLs for one thing because that is where you edit annotations and not for robots. And the hemlock manual, which I didn't what to be indexed.
11:35:52
gilberth
beach: You place it at metamodular.com/robots.txt and say "User-Agent: *\nDisallow: /CLIM-Spec/" This would tell Google to not index that.
11:37:02
gilberth
It's not that urgent. And it would need some days anyway to make it to the databases.
11:40:03
gilberth
beach: You sure have some tool to download the spec, do you process the files otherwise by any means, e.g. fixing URLs or something?
11:45:48
gilberth
Ok, because I need to place those <LINK REL=CAONONICAL ..> thingies, which may not be touched. I'll just wait and see how your /robots.txt fares. Hope that will do.
11:47:03
gilberth
This clim-spec of mine is online for about two decades now, I plan to keep it online for at least as long.
11:56:12
beach
I think your stuff could benefit many people, so it's a shame that it's not available.
11:58:08
gilberth
beach: Not much. If you care, you may tell me what you think of my regular expression submatch addressing draft paper. Which I should have published years ago.
12:03:53
gilberth
Anyhow an implementation of that is the first thing that is on my list of things to release. I have it in "production" here for quite some years, silly to not release it. And hayley has another implementation of the algorithm.
12:18:05
gilberth
beach: I invented the algorithm. Or rather found it. I'm the first guy who could compile RE with submatch addressing to DFAs.
12:29:41
jackdaniel
I'd say that the source shows excessive drawing state configuration, but since I know why is that I'll keep my silence ,)
12:32:26
jackdaniel
it is not a drawback of clim, it is a drawback of me trying to make the backend code as easy to read as possible
12:35:19
jackdaniel
actually some people are crazy with bandwidth; that said if someone feels that they will save the world with smaller svg, they may always hire me to fix that ;p
12:36:59
jackdaniel
then I've went to some lengths to make some exotic designs work; you are all settled
12:42:09
contrapunctus
This is related to `anathema` - what's the best way to change the default-initargs of an object, specifically `basic-pane` and all of its subclasses? I could make an :after method on `initialize-instance`, but that's prone to getting overridden by someone else defining one...
12:43:27
jackdaniel
you create a pane anathema-pane with :default-initargs and use that as a superclass
13:33:44
contrapunctus
jackdaniel: I tried this - https://codeberg.org/contrapunctus/anathema/src/branch/set-theme/src/core.lisp#L13 - but it doesn't seem to have affected the default initargs, at least not according to `(closer-mop:class-default-initargs (find-class 'clim:basic-pane))`...
13:40:13
jackdaniel
if you make pane-mixin a superclass of the basic-pane, then the basic-pane overwrites the pane-mixin, not the other way around
13:44:33
contrapunctus
Right...I just defined an :after method on `initialize-instance`, specializing to `pane-mixin` - it worked! (Hopefully it doesn't go up in flames later.)
14:06:57
contrapunctus
Thanks :) I just hope it's not a problem to rely on the `clime:foreground` slot. Why is `clim:pane-background` an accessor, but `clim:pane-foreground` a reader...?
14:25:04
jackdaniel
if I were to change that to an accessor, I'd first grep for all uses of both to see whether there is something indicating that this is a deliberate choice
15:22:08
gilberth
Especially since I first tried to implement Laurikari's idea. But that just cannot work.
15:25:49
gilberth
Ok. As I said it's a draft-draft. Depression took its toll and so I never had the energy to go and publish it. A shame.
15:47:35
gilberth
beach: I perhaps won't make it by next ELS. And yes, we can say it's Lisp related because the primary implementation is in Lisp. Though, strictly speaking, that algorithm is of general use and not tied to Lisp.
15:51:33
gilberth
Anyhow, formal publication is my least priority these days. I'd rather need to get the implementation out the door. I could also imagine making it an article and just putting it onto my (which one actually?) website. idk. After all, although some reputation is nice, I have no practical use for it.
15:53:32
beach
contrapunctus: Yes. We have increased the level of ambition for SICL so that we should be able to rely on separate libraries, not particularly written for SICL. So then, it becomes natural and desirable to turn former SICL modules into independent libraries.
15:53:33
beach
We have done the reader, the printer, the pretty printer, LOOP, FORMAT, streams, and perhaps some I forgot. The package system is next.
15:54:17
beach
gilberth: ELS is a good forum for getting things out there. Plus, you could go meet some old friends.
15:55:16
gilberth
beach: That would imply that I travel. I am not yet fit to actually travel anywhere.
16:04:07
contrapunctus
beach: ah, I see...I hope this one will support mutually-dependent (?) PLNs... (i.e. package A has a PLN for B and package B has a PLN for A)
16:10:34
contrapunctus
When I tried it with SBCL, I was told that package B didn't exist (it was defined after A). I resorted to using `trivial-package-local-nicknames:add-package-local-nickname` after the definition of B.