luis changed the topic of #slime to: SLIME, the Superior Lisp Interaction Mode for Emacs | | |
<fiddlerwoaroof> margaritamike: did you figure it out?
<fiddlerwoaroof> I got it to work, but have issues with incompatible contribs, Ithink
<fiddlerwoaroof> (e.g. swank-presentations doesn't work in scheme)
<margaritamike> fiddlerwoaroof: I stepped away for a bit. Back now. Trying to figure out how to install 9.2 on ubuntu :p
<fiddlerwoaroof> We probably should have some way to specify hooks per implementation
<margaritamike> It's good to know it mostly works
<fiddlerwoaroof> One thing I do is compile from source and install in my home directory
<fiddlerwoaroof> I don't know the details for mit-scheme, but it would look something like ./configure --prefix=$HOME/mit-scheme
<luis> fiddlerwoaroof: yeah, or otherwise fail gracefully when something is unimplemented.
<fiddlerwoaroof> Yeah
<fiddlerwoaroof> Anyways, luis, I'm about to submit a PR for improved event handling in swank (the run-hooks thing we talked about the other day)
<luis> fiddlerwoaroof: that's great!
<margaritamike> fiddlerwoaroof: supposedly i need mit-scheme already to run ./configure o.o
<margaritamike> I'll see what happens if i install my 9.1.1 again and try to run ./configure
<fiddlerwoaroof> margaritamike: yeah, it's pretty typical for lisps to require themselves for bootstrapping
<fiddlerwoaroof> It makes it --- interesting --- to get a lisp running on a new platform, especially if it's also a different CPU architecture.
<fiddlerwoaroof> The JVM has the same exciting property, though, so we're in good company
<margaritamike> :o interesting indeed
<margaritamike> it's compiling now yay
<margaritamike> should be able to try soon
<margaritamike> bah compile failure
<margaritamike> i'll get there soon
<fiddlerwoaroof> Working on slime is sort of frustrating
<fiddlerwoaroof> The RPC code is really sensitive to mistakes and you end up unable to connect
makomo has joined #slime
<margaritamike> fiddlerwoaroof: how did you get 9.2 installed?
<margaritamike> this is driving me slightly off the rocker
<fiddlerwoaroof> margaritamike: I'm on osx
<fiddlerwoaroof> and brew distributes it
<fiddlerwoaroof> margaritamike: there are binary releases here:
<fiddlerwoaroof> nevermind
<margaritamike> :<
<fiddlerwoaroof> margaritamike: look at the README
<fiddlerwoaroof> There are special instructions for 'Building an incompatible compiler' that might be relevant
<margaritamike> yes indeed
<margaritamike> however
<margaritamike> I have no idea what lib/*.scm is
<margaritamike> lib/*.com**
<fiddlerwoaroof> On Debian testing, I just did ./configure --prefix=$HOME/mit-scheme && make && make install
<margaritamike> I think it's supposed to be, but i have no idea where to find that
<fiddlerwoaroof> and it seemed to work
<fiddlerwoaroof> I used the package from here:
<margaritamike> indeed i used it too
<fiddlerwoaroof> On website, this is the "Unix binary, x86-64" download, not the source code one
<fiddlerwoaroof> Anyways, maybe someone in #scheme knows how to get this to build better than I
<margaritamike> Indeed I did too
<margaritamike> I'll check, but so far no answers
<margaritamike> What are the benefits of this whole bootstrapping philosophy
<margaritamike> Wonder if I would not have this same annoyance without it
<fiddlerwoaroof> margaritamike: you don't have to implement your high-level language in a low-level one
makomo has quit [Ping timeout: 250 seconds]
PuercoPop has left #slime ["rcirc on GNU Emacs 26.1"]
scymtym has quit [Ping timeout: 244 seconds]
makomo has joined #slime
<luis> fiddlerwoaroof: did you mean to push the new contrib to the same pull request?
<luis> Ah, you've fixed it. Nevermind.
Shinmera has quit [Quit: しつれいしなければならないんです。]
_whitelogger has joined #slime
Shinmera has joined #slime
_whitelogger has joined #slime
scymtym has quit [Quit: ERC (IRC client for Emacs 27.0.50)]
scymtym has joined #slime
cage_ has joined #slime
jackdaniel has quit [Ping timeout: 244 seconds]
jackdaniel has joined #slime
jackdaniel has quit [Quit: ZNC 1.7.1 -]
<margaritamike> Ok tried on a mac
<margaritamike> seeing versions differ between slime and nil (swank). Continue?
<margaritamike> That normal?
<margaritamike> Ahh I see what you're talking about with presentations now, fiddlerwoaroof
<margaritamike> swank autodoc seems to not be working either
<margaritamike> Lot of things not working. Can't even eval (+ 2 2). Did you have this issue?
scymtym has quit [Ping timeout: 268 seconds]
<luis> margaritamike: maybe start by disabling all the slime-contribs
<fiddlerwoaroof> luis: your suggested changes look good, I've implemented them
<fiddlerwoaroof> luis: is it straightforward to update CI to use newer emacs versions?
<fiddlerwoaroof> I noticed it's only using emacs 23/24, it would probably be a good idea to have it run 25/26 as well
<luis> It should be, modulo test suite flakiness. We should drop Emacs 23, maybe.
jackdaniel has joined #slime
jackdaniel has quit [Quit: leaving]
jackdaniel has joined #slime
jackdaniel has quit [Client Quit]
jackdaniel has joined #slime
jackdaniel has quit [Client Quit]
jackdaniel has joined #slime
jackdaniel has quit [Client Quit]
jackdaniel has joined #slime
<margaritamike> If I wanted to bind slime-fuzzy-indent-and-complete-symbol to TAB, how would I do that :P
<margaritamike> Does anyone in here do that?
<luis> margaritamike: so, you'd do something like: (define-key slime-mode-map (kbd "TAB") 'slime-fuzzy-indent-and-complete-symbol)
<margaritamike> luis: Is that what you bind to tab?
<margaritamike> I wanna be as efficient as possible with SLIME
<margaritamike> link not found
<margaritamike> :<
<margaritamike> oh forgot the l
<luis> margaritamike: I don't. But TAB indenting, and C-c TAB completing is hard-coded into my muscle memory by now. I might give slime-fuzzy-indent-and-complete-symbol a spin, actually. :-)
* luis sends a note to his work e-mail
<margaritamike> :P
<fiddlerwoaroof> margaritamike: I use slime-company for completion, because I like how company works
<fiddlerwoaroof> I also have it set to complete on TAB or after pausing a second or so after typing
<margaritamike> fiddlerwoaroof: Is it fast?
<margaritamike> I used ac-slime and it was pretty laggy.
<fiddlerwoaroof> I've found it find
<fiddlerwoaroof> s/find/fine
<fiddlerwoaroof> There are a couple issues: in the repl it sometimes gets confused if the output includes an unpaired double quote
<fiddlerwoaroof> And it's a bit tricky to get the initialization sequence right in your init.el
cage_ has quit [Remote host closed the connection]
<fiddlerwoaroof> margaritamike: here's a screenshot (r-h-u)
<luis> fiddlerwoaroof: I'm going to try and upgrade the Emacs versions on travis.yml now
scymtym has joined #slime
<fiddlerwoaroof> cool
<luis> <-- this will either work correctly the first time or take 20 tries to get right. That's been my travis experience so far. :-)
<fiddlerwoaroof> Yeah, I spent a while trying to get travis to work for one of my projects and then I gave up and switched to circleci
<fiddlerwoaroof> which was a lot nicer
<luis> The travis bits seemed to work fine, but there are some REPL tests failing on the newer Emacs versions. Will have to look into that later this week.
<luis> The test suite behaves very differently when executed interactively and non-interactively and it doesn't always recover from previous errors. It's a bit of a nightmare.
<fiddlerwoaroof> sounds like a typical test suite :)
<fiddlerwoaroof> luis: CI is still running, but here's what I think is a minimal viable implementation of the output stream concept