#lisp - Thu 12 Apr 2007 between 20:42 and 21:23

NY Lost Funds



magic_use1http://www.sics.se/~joe/apachevsyaws.html
a la this one in erlang
michaelwjsnell: how is the hashtable growing done?
hefnerXach: what would tying cl-vectors to graphviz accomplish?
jsnelldouble the size, round up to something resembling a prime, put all the data in the old to the new one
Fareso :size the actual size is bad?
the size is dynamically found, btw
from the size of the a-list
jsnellhrm, I was going to say that it would be bad, since obviously the hash will need to be rehashed and grown before its density is 1
Fare(* size 2) ?
(* size 1.618) ?
jsnellbut apparently the default threshold is 1 (= density 100), so I must be misunderstanding something
I expect that :size size would be good enough, it's not like it needs to be grown more than once. but sure, (* size something) would be ok
archaelusmagic_use1: there are heaps of web servers written in common lisp, but not many lisp implementations have green/user-scheduled/non-os thread implementations
Farearchaelus, you only need one, though --- e.g. cmucl
luis<archaelus> just moving all the source into the cffi repo? ; sorry, got sidetracked
magic_use1lol
luisarchaelus: yeah, and documenting the bridge in the CFFI User Manual
Xachhefner: pretty graphs online.
magic_use1cmucl?
Xachhefner: without using PDFs (like cl-user.net) or the ugly builtin rendering (like CLAPPA)
magic_use1clisp or sbcl no?
oh bugger
archaelusluis: I'd be happy for that to happen - I guess I see cl-objc as only useful to mac os x users (it's more cl-cocoa really, I've no plans for gnustep support), so do we want to make everyone who wants cffi install a whole bunch of mac os specific stuff?
hefnerXach: writing a new graph layout engine sounds like it would be more fun.
luisarchaelus: it'd obviously have its own system
Xachhefner: i think that would be fun, but there would be a bit of a shorter turnaround to render the layout produced by graphviz.
gigamonkeylemonodor: I was only vaguely aware of the cyborganic thing
archaelusluis: oh ok - then sure, that sounds perfect :)
Xachhefner: i'm all for doing it completely in lisp, i just think this'd be easier short-term.
luisarchaelus: anyway, I haven't discussed this with James. And we probably should let let Luigi do his thing first.
archaeluson the other hand there are a tiny number of source files at the moment, so it'd be quite easy to roll it in now - and given cffi has a much higher standard of coding (tests, documentation) it might be a good role-model to follow anyway
luis: would you and the other cffi guys be interested in co-mentoring the cl-objc project? (I've got 2 SoC projects this year and am in an inconvenient timezone for real-time help UTC+12)
luisarchaelus: I'm doing a SoC project myself, but I'll be glad to help out with any issues that might come up.
gigamonkeyHas anyone here played with (or used in anger) the Mercurial version control system?
sbclrhello, I have a basic question about sbcl
how do I chdir before I sb-ext:run-program ?
nyefIf you must, (apropos "chdir") will probably give you sufficient clue.
(Okay, you'll need to load sb-posix first, but then it becomes fairly obvious.)
sbclrI've been having problems with sb-posix:chdir...

Page: 2 9 16 23 30 37 

IrcArchive

NY Lost Funds