Conversation with #inferno at Tue May  3 15:06:15 2011 on powerman-asdf@irc.freenode.net (irc)
(16:30:47) vpm left the room (quit: Quit: bbl).
(17:32:13) bvalek2 [588054f4@gateway/web/freenode/ip.88.128.84.244] entered the room.
(18:09:40) anth_x: i don't know for sure, but i've always assumed assignment was atomic (it's hard to imagine otherwise)
(18:10:25) anth_x: but how's that useful for IPC? what you've described should *work*, but doesn't really tell you when things are happening.
(18:10:52) anth_x: (i mean, i get how you could build IPC around that, but it seems like channels are much less work)
(18:13:45) powerman-asdf: anth_x: that's not to replace channels, of course. but sometimes it's too much work to send signal on channel from one process to another to just switch some flag in adt which both processes share
(18:14:44) powerman-asdf: if we stick to 'only one process own this peace of data and may change it' then we shouldn't modify that data from another process
(18:17:36) anth_x: without synchronization (which you can only really reliably do with chans), that's the only safe way to do it.
(18:18:09) anth_x: but reading should be fine; what else could your "d.i < 3" possibly return? either it is or it isn't.
(18:44:12) powerman-asdf: anth_x: with int - yeah, probably. but what's about, say, string? if one process does "s[len s]='\n';" while other "if(s==other_string);"
(18:44:55) powerman-asdf: i'm afraid some of such things may result in segfault
(18:45:05) anth_x: i've always assumed it's atomic (and the debugger implies it), but can't say for certain. i'd ask charles or the list.
(18:49:18) The account has disconnected and you are no longer in this chat. You will be automatically rejoined in the chat when the account reconnects.