08:06 <fiddlerwoaroof> luis: can you think of anything else to do before merging the buffer-streams PR?

08:26 <fiddlerwoaroof> If not, I think I'm going to use my copious spare time to tackle input streams

12:53 <luis> fiddlerwoaroof: I was thinking we could solve the memory leak issue by using buffer-local variables to store the markers.

12:56 <luis> Meh. I'm confused by GitHub's code review system. I left this comment there, but I'm not sure if it's buried away. I really prefer Gerrit these days. :-)

16:52 <fiddlerwoaroof> luis: I did that

16:52 <fiddlerwoaroof> Well, maybe differently than you think.

16:52 <fiddlerwoaroof> I stashed the target name in a buffer local variable and used a hook to cleanup the hash-table

16:54 <fiddlerwoaroof> Just using buffer-local-variables would be suboptimal because it would be rather expensive to "reattach" to a pre-existing target

18:16 <luis> fiddlerwoaroof: João just pointed me to SLIME's channels. It's not a drop-in replacement for what we're doing with *event-hook* but I think that, with a few tweaks, it's a more interesting way of extending swank/slime.

18:19 <luis> fiddlerwoaroof: I'll prototype that in my https://github.com/luismbo/slime/tree/split-dispatch-event branch so you can see what I mean

18:30 <fiddlerwoaroof> cool