freenode/#lisp - IRC Chatlog

21:24:36 sukaeto yeah, I was a vim user for about a dozen years before finally making the switch to Emacs+evil

21:25:32 Cthulhux now just get rid of evil

21:25:33 sukaeto Emacs is strictly better - better buffer management, integration with things like gdb and whatnot, of course SLIME, and the fact that you can easily extend it yourself with elisp

21:26:12 sukaeto but I don't want to start an editor war, here. I know plenty of people who use vim and are totally happy with it.

21:26:22 Cthulhux i use ed(1) ^^

21:26:25 ptdel I use vim and I am in a pit of dispair

21:26:39 sukaeto Cthulhux: but why would I do that? vi is clearly the best text editing language in existence. ;-)

21:26:44 Cthulhux emacs needs better concurrency IMO

21:26:55 Cthulhux sukaeto: because it's not

21:27:00 Cthulhux :p

21:27:15 Cthulhux if you want efficient editing, use ed(1)!

21:27:34 White_Flame M-x butterfly-mode

21:27:38 ptdel Cthulhux: what if I want to save children in Uganda

21:27:51 sukaeto I say use whatever makes you most productive.

21:28:09 White_Flame machine code with toggle switches it is then!

21:28:11 ptdel I feel more productive knowing the children in Uganda are getting daily contributions from vim users

21:28:25 sukaeto if you're really productive in ed then go for it. But that would be pretty weird.

21:29:45 ptdel whats weird about solving all of your editting problems with perl-style regex

21:35:26 edgar-rft ACTION was named after ed(1)

21:38:03 Cthulhux <ptdel> Cthulhux: what if I want to save children in Uganda << that's funny because you can do this in emacs's web browser while you run ed(1) in an emacs shell

21:38:04 Cthulhux :-)

21:38:29 Cthulhux (i ran vim in an emacs shell once. it felt wrong. evil, but wrong.)

21:52:51 phoe jackdaniel: Xach: https://github.com/muyinliu/cl-cpus/issues/2

21:53:56 phoe Cthulhux: evil is the right word here

21:54:20 Cthulhux i AM evil

21:54:22 Cthulhux 8)

21:56:31 edgar-rft i PM evil

21:56:41 Zhivago I don't think that Cthulhu is evil. It must be that x.

21:56:50 ptdel chaotic neutral?

21:57:25 Zhivago I don't think it's human enough to be anything.

22:03:27 Cthulhux human enough to fail to code

22:55:17 Xach phoe: it's too late by then.

22:55:47 Xach asdf has a :feature option that i've seen in a system or two

0:02:18 pjb` ** NICK pjb

1:17:35 stylewarning What method do I specialize if I want to do some actions after an ASDF system is loaded?

2:35:59 kilimanjaro ** NICK neophyte

2:46:57 neophyte ** NICK kilimanjaro

2:53:49 QualityAddict ** NICK Guest82516

3:01:16 beach Good morning everyone!

3:02:03 loke Hello beach

3:02:18 loke ACTION is on leave today, watching over workers installing a new air conditioner.

3:02:32 loke Actually, 4 air conditioners. That's a lot of work.

3:21:04 PuercoPop stylewarning: (defmethod perform :after ((op asdf:load-op)) ((system (eql (find-system <foo>)))))

3:21:04 PuercoPop *asdf:perform

3:23:24 stylewarning PuercoPop: do you know where an appropriate place to put this is?

3:33:43 PuercoPop stylewarning: place I put that is in the ASDF file of the config and place that in (in-package #:asdf-user)

3:35:39 PuercoPop This is how I use it to load a config file http://paste.lisp.org/+7JZM

3:47:57 stylewarning PuercoPop: Let me give it a shot!

4:00:26 stylewarning PuercoPop: worked perfectly, thanks!

4:15:57 jack_rabbit Is there a way to create a lexical-scope binding for a function?

4:16:15 beach (let ((x 10)) (defun ff (y) (+ x y)))

4:16:24 beach Is that what you mean?

4:16:46 beach Or maybe you mean (flet ((ff (x) (+ x 10))) ...)?

4:16:49 shka_ jack_rabbit: or you mean a local function?

4:16:50 jack_rabbit No, like a let, but for the function namespace rather than the variable namespace.

4:16:55 shka_ damn it

4:16:56 jack_rabbit Yeah, flet probably

4:17:01 shka_ beach: good morning

4:17:04 beach clhs labels

4:17:04 specbot http://www.lispworks.com/reference/HyperSpec/Body/s_flet_.htm

4:17:07 jack_rabbit Thanks, beach.

4:17:10 beach sure.

4:17:13 shka_ jack_rabbit: there is also labels

4:17:20 shka_ …

4:17:26 shka_ i guess i will just go to work

4:17:40 jack_rabbit Trying to temporarily shadow some functions. :)

4:17:49 shka_ flet is way to go

4:17:59 beach jack_rabbit: You can't do that with standard functions though.

4:18:10 jack_rabbit :/

4:18:35 shka_ what standard function, though?

4:18:57 jack_rabbit Guess I can traverse the body and replace instances of the symbols for the functions I want to shadow.

4:19:09 jack_rabbit in a macro.

4:20:54 beach jack_rabbit: That is generally very hard to do. It requires a thing known as a code walker, and such a thing can not be implementation independent.

4:21:44 jack_rabbit Why is it not as simple as recurring through the lists and identifying the relevant symbols?

4:22:10 beach Because there may be macro calls that, when expanded, introduce some of the symbols you don't want.

4:22:31 beach Also, the role of a symbol may be different.

4:22:47 beach Perhaps you want to replace a symbol when it is in a function-call position, but not when it is data.

4:22:55 jack_rabbit Yes.

4:23:03 beach To know the role of a symbol requires a code walker.

4:23:26 jack_rabbit Can't I macroexpand the body first?

4:23:54 beach But to know whether to macro expand, you need a code walker.

4:24:03 jack_rabbit Ahh.

4:24:06 beach You may have (macrolet ((bla ...)) ...)

4:24:27 jack_rabbit There's no built-in "fully expand everything in this form"?

4:24:36 beach It is known as a code walker.

4:24:39 stylewarning ql:quickload :macroexpand-dammit

4:25:09 jack_rabbit Gotcha. and there's not a standard built-in one that's accessible?

4:25:32 jack_rabbit I guess that's what macroexpand-dammit is for.

4:25:33 beach There are some libraries as stylewarning pointed out.

4:26:11 beach But unless they adapt to the implementation, they are going to be approximate.

4:26:42 jack_rabbit May give different results than the implementation's walker, you mean?

4:27:32 beach They may work for most simple cases, but may not work for complex things involving macrolet and such.

4:27:50 beach Then again, maybe you don't have such complex code.

4:28:51 jack_rabbit So, in that case, I wanted to do the following: create a macro such that arithmetic operations within the body printed what they were doing to *standard-output*

4:29:14 jack_rabbit Just a simple excercise to see how easy it was. I don't have a practical application for it.

4:29:43 beach Not simple and not easy if you want it to be general.

4:29:57 jack_rabbit Makes sense.

4:38:42 beach I seem to always be unable to find the Common Lisp HyperSpec page that explains how a conforming program is allowed to use symbols in the CL package.

4:39:10 |3b| 11.2.1.2.1 or something like that?

4:39:26 |3b| clhs 11.2.1.2.1

4:39:26 specbot Couldn't find anything for 11.2.1.2.1.

4:39:30 |3b| guess not :p

4:39:37 |3b| clhs 11.2.1

4:39:37 specbot Couldn't find anything for 11.2.1.

4:39:52 beach Yes, thanks.

4:40:11 beach I always look in the first few chapters. Never think to look in the chapter on packages.

4:40:16 |3b| ah, 11.1.2.1.2

4:40:28 beach Yes.

7:22:50 phoe I have a LPARALLEL question. Is there a way to submit tasks to a channel that does not cause results to be generated?

7:23:11 phoe I want to fire-and-forget, I want the results not to be accumulated.

7:34:20 phoe Because the results, when unhandled, will cause a memory leak, and when handled, will require a separate result-receiving thread or another thread to try popping the results until none are left.

7:44:44 _death phoe: results must always be received.. for example you can look at how lp:pmapc is implemented

7:45:56 phoe _death: okay - I will work around this by beginning each worker function with a call that pops and discards all accumulated results.

8:00:26 _death phoe: that seems more expensive than another thread that spends almost all its time waiting

8:01:35 _death it also won't work, as there's always at least one result not yet received

8:02:30 phoe well, damn.

8:03:26 phoe _death: I have try-receive-result though that will not block if there is no result.

8:04:55 _death phoe: yeah.. I mean it doesn't work if you want to prove the invariant that all results are received

8:05:38 phoe if there are 8 workers, then at most 8 results will be not received at any time.

8:06:02 phoe good enough.

8:07:33 _death but why not just have another thread to do that instead of increasing the cost for each worker

8:15:36 pjb` ** NICK pjb

8:15:52 phoe isn't the cost there to be paid anyway?

8:25:38 _death phoe: yes, but the cost of one thread waiting is less than the cost of each worker polling

8:29:18 phoe ACTION nods

8:48:07 Suzuran42 ** NICK Suzuran