<orbitz>
Kakadu: Anyways,if you use lwt, make sure to get the syntax highlighting to work
<orbitz>
If oyu use async, for long sequential sequnecs I would use pa_monad
Submarine has joined #ocaml
Submarine has quit [Changing host]
Submarine has joined #ocaml
<orbitz>
bitbckt: You know of any work to get GC help at the hardware level for comodity PCs?
osa1_ has quit [Ping timeout: 260 seconds]
<adrien>
considering how litte widespread it is for specific tasks, I'd be surprised to find that in comodity computers
<adrien>
plus you'd need to decide which model the GC would have
<orbitz>
I'm curious more about commoidy CPUs 10 years from now, I assume the work is being done now
<orbitz>
adrien: Well, I'm thinking maybe Intel supports a way to get an interrupt whenever a poitner is accessed that you hav emoved
<orbitz>
I beleive Azul hardware does this
<Anarchos>
orbitz interrupt when a pointer have moved, isn't it like page fault ?
<orbitz>
Anarchos: Maybe?
<orbitz>
although this would be pointer level not page
<Anarchos>
orbitz in Sparc ISA you have specific integer instructions supporting a bit tag , specifically intened to support box/unboxed integers for lisp
Yoric has quit [Read error: Connection reset by peer]
Submarine has quit [Ping timeout: 248 seconds]
ftrvxmtrx has quit [Quit: Leaving]
Yoric has joined #ocaml
ftrvxmtrx has joined #ocaml
<mrvn>
orbitz: you can write protect pages and catch the segfaults in the GC, mark the page read-write and mark all objects in it for another scanning run.
<mrvn>
orbitz: But that is expensive.
<orbitz>
Yeah
<orbitz>
I'm thinking something more streamlined, like how Azul does it
<orbitz>
I'm not sure how it does it on the back, but when teh GC runs the hardware catches an interrupt ro move the pointer as it's accessed
ulfdoz has quit [Ping timeout: 244 seconds]
netrino has joined #ocaml
<mrvn>
orbitz: which is expensive. Both changing the permissions and catching the access.
ulfdoz has joined #ocaml
<orbitz>
mrvn: Azul is custom hardware so I don't think it's that expensive there
Yoric has quit [Ping timeout: 244 seconds]
Yoric has joined #ocaml
ulfdoz has quit [Ping timeout: 240 seconds]
Submarine has joined #ocaml
Submarine has quit [Changing host]
Submarine has joined #ocaml
Submarine has quit [Read error: Operation timed out]
ulfdoz has joined #ocaml
ulfdoz has quit [Ping timeout: 245 seconds]
Yoric has quit [Ping timeout: 260 seconds]
ulfdoz has joined #ocaml
osa1_ has quit [Ping timeout: 260 seconds]
struktured has quit [Remote host closed the connection]
struktured has joined #ocaml
ulfdoz has quit [Ping timeout: 245 seconds]
ulfdoz has joined #ocaml
Anarchos has quit [Quit: Vision[0.9.7-H-090423]: i've been blurred!]
Yoric has joined #ocaml
Yoric has quit [Ping timeout: 244 seconds]
zolk3ri has joined #ocaml
Yoric has joined #ocaml
ulfdoz has quit [Remote host closed the connection]
ulfdoz has joined #ocaml
<zolk3ri>
Hi.
<Kakadu>
hi
Yoric has quit [Ping timeout: 264 seconds]
hongboz has left #ocaml []
hongboz has joined #ocaml
Yoric has joined #ocaml
Yoric has quit [Ping timeout: 264 seconds]
fraggle_ has quit [Read error: Connection reset by peer]
fraggle_ has joined #ocaml
Submarine has joined #ocaml
Submarine has quit [Changing host]
Submarine has joined #ocaml
mjonsson has joined #ocaml
sepp2k has quit [Remote host closed the connection]
Submarine has quit [Ping timeout: 240 seconds]
slash__ has joined #ocaml
slash__ has quit [Client Quit]
slash__ has joined #ocaml
slash__ is now known as slash^
gnuvince has quit [Ping timeout: 250 seconds]
Progster has joined #ocaml
gnuvince has joined #ocaml
<adrien>
I've been using sexplib in yypkg and now I use hand-written (de)serializers
<adrien>
(while I was using the type-conv-based ones before)
<orbitz>
How ocme?
<adrien>
the error messages suck just as much
<adrien>
but I have no idea how to make proper ones
<adrien>
orbitz: you're wondering why I stopped using the type-conv-based ones?
<orbitz>
yes
<adrien>
one reason is that I use it on windows and camlp4 has some issues there so it's better to just avoid the dependency completely
<orbitz>
ok
<adrien>
then, I think my code is much smaller and it's noticeable in the size of executables, but that came as a bonus