freenode/#lisp - IRC Chatlog
Search
17:11:59
nosefouratyou_
does anyone know of a small project I could write tests for? I want to practice using testing in cl.
17:17:26
malice
nosefouratyou_: and if you want to start *really* small, you can add tests to cl-sandbox: https://github.com/MatthewRock/cl-sandbox/
17:17:53
nosefouratyou_
malice: I am definitely looking for something small to start with, thank you for that suggestion
17:23:30
nosefouratyou_
prove looks awesome! I really like the fact that it has colorized output, sort of like npm stuff
17:26:30
nosefouratyou_
is there a way in helm to make backspace delete the word instead of individual characters?
18:14:36
Bike
kang0: this isn't a general channel. you can get better help for a lot of questions elsewhere.
18:37:52
shka
does anybody knows why the heck armed bear common lisp was called armed bear common lisp?
18:44:13
malice
jackdaniel: I remember that you had some nice image about various implementations and thought that you had some knowledge about them
18:44:13
jackdaniel
old website shows a quote, which indicates it was a pun: " The right of the people to keep and arm bears shall not be infringed! "
18:44:47
jackdaniel
yes, I had a diagram, but I don't know why armed bear, I only suspect that it was the abovementioned pun
19:07:32
phoe
p_l: I know that ABCL ran on IKVM, but I had no idea that it was used for commercial purposes
19:08:42
p_l
I have a... view into certain project at one multinational. With the quality of source, rewriting in Lisp would probably fly unnoticed...
19:10:08
p_l
tl;dr version is that quite often companies don't care what language is used so long as it's on supported platform
19:12:59
p_l
and some of the "important LoB apps" is a) dead simple coding b) nobody looks wt it anyway except poor sods in support. So be your own support
19:18:00
p_l
an app that every month provides reports considered crucial in a certain multinational is some of the worst java code that had been decompiled with original source lost
19:18:59
p_l
this is combined with shell script that triggers an oft-failing file transfer tool, which is then processed by MSSQL stored procedures and SSIS packages
20:44:11
aeth
Programs that mostly just interact with a database can be written poorly because the real work is done in the database.
21:28:08
phoe
...that's going to be a proof that you can't get some friends without getting enemies along the way
23:46:13
thorondor[m]
hello. do you know if there's a portable way of handling extra initargs in CLOS?
23:46:34
thorondor[m]
LispWorks has :extra-initargs class option, but doesn't seem to be supported in other Lisps
23:53:51
thorondor[m]
I don't see it. And by inspecting SBCL code, it even seems to be fixed at MOP level.
0:02:15
Bike
let make-instance accept more keyword arguments that aren't in the lambda lists or slot initargs?
0:09:11
thorondor[m]
I know it is a rare use case, but I have it (it's the first time I feel I want it)
0:15:55
thorondor[m]
I want to use them on make-instance (initialize-instance) without having to define slots for them
0:18:11
specbot
Declaring the Validity of Initialization Arguments: http://www.lispworks.com/reference/HyperSpec/Body/07_ab.htm
0:19:51
Bike
(Defclass foo () ()) (defmethod initialize-instance :after ((instance foo) &key testing) (print testing)) (make-instance 'foo :testing 7) => prints 7
0:19:56
thorondor[m]
but CL has always been quite incredible in handling my corner cases quite gracefully
0:20:10
Bike
never used that feature before but it was a pain to implement, so i'm glad it works and is useful to you
0:39:38
emaczen
I think format can map 123 -> one hundred and twenty three. Is the reverse mapping built-in to commonlisp?
3:28:16
PuercoPop
Is there a good reason why SBCL doesn't provide the restart to finalize class when one calls class-slots on a class that hasn't been finalized?
3:34:17
axion
Is there a simple way to split a string delimited by #\Space into a list of strings, without using a library?
3:56:38
fiddlerwoaroof
What's the best way to split a pathname into a list of directories + the file name?
3:58:22
fiddlerwoaroof
JuanDaugherty: I actually can't find anything "off the shelf", although I've managed to hack something together that sort of works
4:03:11
JuanDaugherty
i was think about the end function, the path name elements and finally the file name
4:21:08
White_Flame
ugh, my old slime won't build on new sbcl, (*ASSEM-INSTRUCTIONS* not found in SB-ASSEM), and I need old slime to talk to long-running programs without version mismatches
4:21:34
White_Flame
guess I need a separate old sbcl to run that, too. And a separate emacs just to be configured for that
4:25:14
White_Flame
I would pay money to have somebody change SLIME such that the emacs side of it downloaded the instance-specific elisp into a closure only for that connection, instead of having a singular full slime client in emacs itself
4:28:19
White_Flame
if the swank end of things held a copy of the elisp specific to that swank, all would be swanky
4:29:38
White_Flame
heck, actually I guess I could still connect, but I can't build anything with swank at the moment
4:30:25
White_Flame
I don't know if there are variations for the emacs side of things, but still, the generic code can come along
4:31:18
SAL9000
and you're back where you started unless all Lisps that you want to co-host are on the same shim-version
4:31:38
White_Flame
I've written many thin launchers, and they stay put, because all the brains are in the downloaded code
4:32:00
SAL9000
my point is that eventually you'd need a new API which can't be backwards compatible for one reason or another
4:32:59
SAL9000
I thought you wanted to be able to use different SWANK versions concurrently in the one instance of Emacs
4:33:10
White_Flame
the idea would be that when you did slime-connect, it would download a copy of emacs-side slime specific to that connection
4:33:33
SAL9000
if it blindly executes the code blob it won't be able to evict old-swank when loading new-swank
4:33:36
White_Flame
instead of requiring installed emacs-side slime client code to version match the swank-side version
4:35:26
White_Flame
but it would solve the problem of having long-runnign servers with naturally older versions of swank
4:36:06
SAL9000
yes. If you don't care about evicting version 1 of swank to replace it with version 2 of swank -- all in the same Emacs instance -- then blind blob execution seems OK to me.
4:36:19
White_Flame
(either that, or bloat the crap out of emacs-side slime, retaining all older versions :-P)
4:36:52
SAL9000
well, I imagine that the "proper" way would be to formalise the SLIME/SWANK APIs more
4:37:17
White_Flame
yeah, I'm sure some of the eclipse or whatever packages that connect to swank would appreciate that as well
4:37:46
White_Flame
but still, I think you need to s/swank/slime/ in a lot of your statements, if you're understanding me
4:38:50
White_Flame
I can issue .lisp files to be eval'd into the system, but that wouldn't persist a restart
4:39:37
White_Flame
in any case, this has been a nuisance multiple times, but still a nuisance, not a breaking issue
4:40:25
fiddlerwoaroof
Hmm, I wonder if you renamed the swank package to SWANK-OLD (or the like) before loading new swank, if things would be alright
4:41:17
fiddlerwoaroof
That way you could keep old swank up and running until new swank has been tested
4:42:59
fiddlerwoaroof
Yeah, I was thinking about the situation where swank was part of some critical infrastructure
4:43:15
White_Flame
(well, and I can't start SBCL locally at the moment, because emacs is pointing at old slime, but whatever)
4:45:20
fiddlerwoaroof
But, in theory, renaming the package shouldn't affect the things in the package, unless the package relies on interning symbols in "PACKAGE-NAME" or the like
4:46:00
fiddlerwoaroof
So, it should be possible to run old and new versions of a package next to each other