IRC channel logs


back to list of logs

<leoprikler>dunno, but lang as a feature seems pretty broken imo
<leoprikler>like, you have to ensure it never means anything else in any implemented language (and that the reader for that language can handle it)
<mwette>leoprikler: Look at the test case. It's on ecmascript for which # would generate a parse error, if the ecmascript saw that line. I agree it needs more discussion, hence never submitted as an official patch.
<drakonis>leoprikler: i agree that lang as racket does it, is pretty broken
<drakonis>since you do not have the ability to mix langs within a file
<drakonis>you would however want to include the ability to call out to other languages in guile
<mwette>drakonis: nyacc's examples/nyacc/lang/nx-lib.scm provides (install-inline-language-evaluator); so, you can (define foo #<ecmascript: 1 + 2;># ) where foo -> 3
<drakonis>that would be quite the solution
<leoprikler>I think for Scheme and Scheme-esque languages, reader extensions might work, but how would you embed Scheme code into ECMAScript?
<leoprikler>probably eval-string combined with backticks
<dsmith-work>Morning Greetings, guilers
<dsmith-work>New Inet connection here. Looks like they are blocking irc. :(
<dsmith-work>sneek: botsnack
<sneek>I've been running for 14 days
<sneek>This system has been up 2 weeks, 6 days, 1 hour, 56 minutes
<dsmith-work>Hey hey.
<dsmith-work>sneek__: botsnack
<dsmith-work>sneek: botsnack
<dsmith-work>Hmm. I bet the bot it not allowed to change it's nick. Because the sneek__ is not registered. (sneek_ is)
<RhodiumToad>sneek pinged out at about 16:25 BST
<RhodiumToad>oh, I see
<dsmith-work>Just to register sneek__ with nickserv too. IF that's the problem.
<dsmith-work>Hmm. I wonder. If someone kicks the bot, maybe it will get it's nick back.
<dsmith-work>wingo, civodul: Can you become ops and try kicking sneek__ . Hoping it will come back as sneek
<RhodiumToad>maybe the bot could be taught to do that itself via /msg nickserv ghost ?
<dsmith-work>Yeah, I think it does for sneek_ just need to make it happen for sneek__ too.
<manumanumanu>Everytime I think "This seems like an easy thing to do using read-hash-extend" I always regret it :(
<dsmith-work>Now that Guile has a Scheme reader, it might be easer to localize those kinds of things.
<manumanumanu>is guiles reader in scheme now?!!
<lampilelo>since 3.0.6
<manumanumanu>I wonder what it would take to make reader extensions module-local...
<manumanumanu>everywhere I go thinking about it, I see obstacles.
<dsmith-work>manumanumanu: civodul had something that allowed you to use a different custom reader. Was a while ago. Not sure it it's still relevant.
<manumanumanu>guile-reader, sure
<manumanumanu>that is the easy part
<manumanumanu>but what if I define a reader-extension in my module (bjoli literal-json) and any code importing that module could do #json[1, 2, 3]
<manumanumanu>I would need a "merge-readers" like merge-generics. Say I import (ArneBab wisp) and (bjoli json). What kind of voodoo would I need? I don't think it could be much else than a hack.
<manumanumanu>I have megacut: which does a read-hash-extend, but that makes global changes to the reader. Which is icky.
<manumanumanu>oh my, look at the time. good night folks.
<civodul>manumanumanu: my take is that you should write (json #(1 2 3)) or whatever instead of using special read syntax :-)
***ChanServ sets mode: +o civodul
***sneek__ was kicked by civodul (Kicked by civodul)
***ChanServ sets mode: -o civodul
<civodul>dsmith-work: i kicked sneek__ but it didn't have the intended effect :-)
<dsmith-work>civodul: Thanks. Yeah, figured it wouldn't work but worth a try.
<dsmith-work>I don't see join/part messages. I assume it just re-joined with the current nick?
<dsmith-work>sneek__: botsnack
<dsmith-work>Well, I'll be home soon.
<civodul>yes it immediately rejoined with the same nick