freenode/#lisp - IRC Chatlog
Search
18:50:37
phoe
if they didn't nest, the comment would instead end on the first |#, leaving a contextless chunk of code full of mismatched closing parens
19:03:40
jackdaniel
it was one of first lex excercises we had at uni - implement nested block comment in c (non-conforming)
19:32:26
specbot
Notes about Style for Sharpsign Vertical-Bar: http://www.lispworks.com/reference/HyperSpec/Body/02_dhsb.htm
19:36:19
_death
seok: yes, you can use them together.. in fact it's sometimes a good idea when defining a macro, to determine the appropriate order of evaluation given the macro form
19:41:37
_death
also note that the default value is the symbol *auth*, not the value of *auth* at macroexpansion time
19:46:10
_death
I'm not sure what you're asking.. if your macro gets a symbol (like *auth*) you can use it in an evaluative context within the macroexpansion without trouble
19:47:12
aeth
On the earlier topic of architectures and compilers... Can RISC V replace ARM? Definitely. Can RISC V replace x86-64 on desktops/laptops? Probably not. Nobody (except Apple) is going to break backwards compatibility with 3-4 decades of software binaries. I don't think FASLs are typically portable across architectures, are they? So this even affects CL.
19:53:30
_death
aeth: for example sbcl fasl format has a version field, and I don't think it supports values other than the "current version"
19:54:29
aeth
_death: I wonder if that makes semver impossible, since every release would be a breaking one
20:05:49
lukego
beach: Just reflecting that Emacs is not as sticky as I often think. I've been writing a lot of Lua in Emacs and there it doesn't support much of anything beyond indentation and syntax highlighting. I've been writing R in RStudio which is a 90s style crappy IDE that just happens to make it easy to plot and render RMarkdown out of the box. Pharo smalltalk is its own world too. So McCLIM seems quite plausible in that sense.
20:09:28
lukego
I'd say it's an "it's complicated" but I'm only using it as an example of a system that I've been quite willing to "live" inside despite being an Emacs weenie and having to give up all Emacs features to work with it.
20:10:26
lukego
I think you have to choose your battles carefully when you're making a "parallel" computing world. I don't think you want to reinvent things like git.
20:10:49
lukego
well love-hate with smalltalk. I hack some smalltalk every ten years and then flame out after some weeks or months :)
20:11:34
lukego
I spent quite some time working with the GToolkit object inspector library recently. very love-hate. it's the opposite problem to Common Lisp world - anything written more than one year ago is hopelessly obsolete, anything newer is hopelessly unstable :)
20:12:15
lukego
maybe that's why I get nostalgic for Lisp now. Hard to get shit done without a stable foundation to work from.
20:13:06
lukego
I should try to get McCLIM up and running. just had a build error last time that was complicated by my very quirky usage of Quicklisp via Nix that prevented me from being able to tweak the code
20:13:58
lukego
also Emacs changes. I enabled company-mode, ivy, etc, that everybody is into these days and I don't know what the hell is going on anymore :)
20:16:26
lukego
pve: I do really like the way Smalltalk code is factored. I don't really like the IDE though. I like my code organized into files. my brain is linear :)
20:17:00
lukego
I haven't read the Blue Book and would like to one day see how it's implemented under the hood. I've gone down that rabbit hole with Forth and that was worth the trouble for sure :)
20:17:41
lukego
ok I have mcclim giving me the instllation error but running from a local checkout that I can actually patch. bedtime now but added to my procrastination vehicle rotation :)
20:18:21
lukego
pve: I don't think I've ever seen a Lisp-Smalltalk hybrid actually? seems to make potentially a lot of sense :)
20:20:04
pve
lukego: at this stage it's mostly a syntax thing, but I wanted to see what would happen if one could mix lisp with smalltalk style message passing
20:22:41
lukego
I also fantasise a little about doing LuaJIT style tracing in Lisp at least as an experiment/illustration. run code for a while, discover the in-practice types of some variables and definitions of some functions, and call COMPILE to create chains of specialized fast-path functions.
20:25:33
pve
the first picture defines a source file class for the loader, the second does some asdf integration
20:27:54
pve
as you can see, it's not exactly smalltalk, but if you squint you can kind of see where the inspiration came from.. and besides, it's in a pretty early stage
20:30:50
aeth
lukego: Imo, people don't use tools based on the merits of the software itself, but based on the ecosystem. Emacs is Emacs, but any CL-based editor with tight integration to a popular CL platform would win over most of the CL community.
20:33:32
lukego
aeth: Maybe. CMUCL had Hemlock but it felt like I was one of maybe three people in the world using it :)
20:38:18
lukego
oh hey I have the clim listener running now :) just kludged in another hard coded path to look for truetype fonts
20:41:52
aeth
lukego: yeah, but the thing is, it's not Emacs, it's GNU Emacs. GNU Emacs has GNU, which is a little more popular than CMUCL, especially in the 90s
20:45:41
p_l
lukego: seems like a lot of people like and use Hemlock, though not necessarily under that name ;)
20:49:03
aeth
I mean, being more popular than the official editor of GNU-as-a-platform isn't really an impossible advantage to beat these days. Emacs' moat has probably never been smaller.
20:51:30
aeth
I think these days, the main disadvantage is getting general users to get used to the Emacs keys that most CL programmers are used to by now...
21:35:50
aeth
p_l: well, if there are so many laptops, then just use a second laptop instead of foot pedals!
21:36:25
p_l
unwieldy and too expensive, not to mention it would also have anti-emacs keyboard of intel/apple source
21:59:10
aeth
edgar-rft: But everyone who uses BCI can tell you as if with one voice that it's the greatest thing ever and definitely doesn't control them.
0:34:44
iissaacc
does anyone know why (setf (return-code*) +http-not-found+) gives me the error "the value 404 is not of type sequence"
0:35:18
iissaacc
So it seems to be expecting a sequence rather than an integer return code. But the manual and the source code do it the way I have done it
0:59:08
Gnuxie[m]
iissaacc: if it's the last thing you're doing hunchentoot will try write the new value from setf to the response stream
1:09:08
iissaacc
so if i do the setf and then some other expression after, it should work? its to signal "user not found in db"
3:38:35
beach
I am fine at least. Today, I need to finish part 2 of my series of presentations for phoe's online Lisp meeting to be streamed on Wednesday.
3:39:34
beach
The series is called "Creating a Common Lisp implementation" and it is about different strategies for structuring such a thing.
3:40:10
beach
Part 1 (along with the other past presentations for the online Lisp meeting) is available somewhere.
3:43:48
beach
I hope to be able to explain the strategy used by SICL in the last part of the series.
3:44:06
minion
iissaacc: SICL: SICL is a (perhaps futile) attempt to re-implement Common Lisp from scratch, hopefully using improved programming and bootstrapping techniques. See https://github.com/robert-strandh/SICL
3:44:22
minion
iissaacc: Cleavir: A project to create an implementation-independent compilation framework for Common Lisp. Currently Cleavir is part of SICL, but that might change in the future
3:49:25
iissaacc
can anyone attend this online lisp meeting? ive just handed in my thesis so ive got plenty of time for stuff like that
3:49:54
iissaacc
more to spectate than contribute. ive only just started using lisp and im trying to immerse myself in it
3:51:09
beach
So you can type questions while the presentation is streamed, and the person presenting will hopefully answer in real time.