log☇︎
24900+ entries in 0.014s
mircea_popescu: asciilifeform, did the fg / kernel thing ever make sense to you ?
mircea_popescu: i guess we try this crosslog thing see.
asciilifeform: ^ interesting thread re 'fg in kernel or not?' that sat in shadow of 'grr no logs' episode
asciilifeform: but iirc this also existed prev.
mircea_popescu: yeah, the one time that weird quirk came in handy
asciilifeform: ( protocol, ftr, is such that speakers do not actually see their own output. logger has to log it explicitly )
asciilifeform: iirc this worked in all prev. loggers also
asciilifeform: ok lemme know if you want it converted to classical 1line.
mircea_popescu: not what the spec ~actually~ satys, but i guess you went per tradition.
asciilifeform: mircea_popescu: want this in next rev ?
asciilifeform: what's with this 'guess our magic rate limit' thing.
asciilifeform: fucking tards, the fleanode people, if you wanna by synchronous, BE SYNCHRONOUS, say when ok to send next ln !!11
mircea_popescu: asciilifeform, i realise now i didn't even spec the bot correctly, it's supposed to multiline if multi links
snsabot: Logged on 2019-08-08 16:51:33 asciilifeform: for completeness, the www end, in case asciilifeform gets sepsis before gets to genesis. and example config (config is common to both proggies, give it on cmdline)
snsabot: Logged on 2019-08-08 17:54:51 mircea_popescu: freenode wants 1s delay between msgs, and bot seems to try and do that but maybe add a 50ms buffer on top or somethign ?
asciilifeform: perceptible enuff diff that it annoys asciilifeform , who developed long ago habit of refreshing log pg on his various machines 'as if it were local thermostat log'
snsabot: Logged on 2019-08-08 17:53:34 mircea_popescu: http://logs.nosuchlabs.com/log/trilema/2019-08-08#1926442 << seems pretty snappy to me, yeah.
asciilifeform: http://logs.nosuchlabs.com/log/trilema/2019-08-08#1926458 << well, take the current log tail; here on my box 'time curl ....' is 2.2s for mine vs 0.09s for phf's
mircea_popescu: freenode wants 1s delay between msgs, and bot seems to try and do that but maybe add a 50ms buffer on top or somethign ?
snsabot: Logged on 2019-08-08 17:20:05 spyked: http://logs.nosuchlabs.com/log/trilema/2019-08-08#1926408 <-- on it. gotta run to sleep afterwards, but will be back in the morning and get up to date with l0gz. meanwhile, is it okay if I remove trilema comment PMs from the bot's message queue? otherwise it'll keep sending when I bring it back up, lol.
mircea_popescu: http://logs.nosuchlabs.com/log/trilema/2019-08-08#1926447 << i guess i already saw them all so yes. but in general, kinda needs a permafix
snsabot: Logged on 2019-08-08 17:05:07 asciilifeform: snsabot and its www end are imho usably snappy, but you can tell that they run in interpreter, visibly slower than phf's
mircea_popescu: http://logs.nosuchlabs.com/log/trilema/2019-08-08#1926442 << seems pretty snappy to me, yeah.
snsabot: Logged on 2019-08-08 16:59:01 diana_coman: and ftr I have no idea why did the other bots go for lisp anyway, did I miss somewhere the rationale for it?
mircea_popescu: http://logs.nosuchlabs.com/log/trilema/2019-08-08#1926421 << iirc bv wanted to learn / practice it and everyone followed.
snsabot: Logged on 2019-08-08 16:54:53 mircea_popescu: spyked, feedbot choked on having to send me too many rss lines lol. can it come back ?
spyked: http://logs.nosuchlabs.com/log/trilema/2019-08-08#1926408 <-- on it. gotta run to sleep afterwards, but will be back in the morning and get up to date with l0gz. meanwhile, is it okay if I remove trilema comment PMs from the bot's message queue? otherwise it'll keep sending when I bring it back up, lol.
asciilifeform: snsabot and its www end are imho usably snappy, but you can tell that they run in interpreter, visibly slower than phf's
asciilifeform: python dun support machine threads.
asciilifeform: diana_coman: 1 reason to use cl, is that ~50x faster, vs pythonisms. and can use actual machine threads.
diana_coman: yep; but if you use gnat.sockets you already eat that too anyway; so lots of stink in already.
asciilifeform: there's a place, at least for so long as we're doing tcpisms, for langs with garbagecollector etc.
diana_coman: yes, my current understanding is exactly that - it's a stink anyway and unavoidably so might as well use python, at least it's quick and relatively clear
asciilifeform: diana_coman: principal weakness of ada in re this problem set (at least the ada 'specified' by asciilifeform) is stringism handling
asciilifeform: would have to write pg end also !
diana_coman: ultimately sure; currently I don't know; I even sketched quickly an Ada thing - main trouble though is lack of proper db interface (and otherwise relying on GNAT.Sockets and therefore the whole strand of streams etc)
asciilifeform: i actually started to write one, lol! when this particular episode began. but quickly realized that it'll carry on for month+
asciilifeform: currently there aint a 'tmsr lang' in which can readily write wwwisms. (i dun even have a tcp end for gnat atm)
asciilifeform: diana_coman: rationale is that each bot maker used the lang that was most fresh in head.
asciilifeform: and ultimately in tmsr-l
diana_coman: and ftr I have no idea why did the other bots go for lisp anyway, did I miss somewhere the rationale for it?
asciilifeform: but deliberately wrote this one so it can be with minimal effort rewritten in sumthingelse (i.e. did not use py 'irc libs' or whatevers)
asciilifeform: ever since 1st had to write a wwwistic proggy (phuctor front)
diana_coman: and I find out I didn't miss at all not using python for years; onth no idea what else would be better for such a task since it can't be really clean anyway
asciilifeform: so instead has configable delay there.
asciilifeform: unlike earlier ben_vulpes irc proggy, mine does not 'wait' for fleanode-specific warmup message (deliberate, i wanted to be able to stand it up behind e.g. znc , if desired)
asciilifeform: mircea_popescu: as you can tell from last coupla comments on your www, even there i missed initially a spot of retardation (fleanode lets you auth, but if you join ~immediately~ then won't be authed for purpose of +r chans. this is documented nowhere.)
mircea_popescu: spyked, feedbot choked on having to send me too many rss lines lol. can it come back ?
asciilifeform: PeterL: in order to actually attempt to 'fit in head' the barbaric protocol, i wrote bot on naked sockets, instead of using whatever-lib
asciilifeform: what does, is to prevent eternal hang on silent (ask the tcp committee why this is physically possible, not me) deaths
asciilifeform: PeterL: recv timeout doesn't disconnect. state machine goes to recv again.
asciilifeform: for completeness, the www end, in case asciilifeform gets sepsis before gets to genesis. and example config (config is common to both proggies, give it on cmdline)
PeterL: It's been a while since I coded a bot, is it possible to send a PING when you hit timeout rather then going straight to reconnecting?
asciilifeform: imho the current item should be left alone until mirror stands up , at least unless someone finds how to choke it trivially
asciilifeform: main problem is that right nao there is exactly 1 reliable logotron, so when i add knobs, will have to be cycled (will try to do when no one is awake)
asciilifeform: item is sewn such that theoretically can digest any barf thrown by fleanode. but this will have to be verified experimentally.
mircea_popescu: sounds pretty well designed tbh.
snsabot: 1000 results for "f:ifeform f:escu trb" in #trilema
asciilifeform: !q s f:ifeform f:escu trb
asciilifeform: ( idea was, master and slave required to sit on diff. fleanode shitpits at any given time )
asciilifeform: initially i considered to have two-headed bot, with a slave who is kept around strictly if master falls. but this not implemented yet.
asciilifeform: there is timeout on recv() so 'silent stall' also not wedges bot. at least in theory.
asciilifeform: bot is built to cycle through list of irc servers (in our case, fleanode) , if one throws, goes to next
mircea_popescu: yeah, kinda curious how this weathers.
asciilifeform: the minus is that bot search dunwork if wwwtron is down. ( but why should be down, for more than coupla sec it takes to reset when new ver. phuctor's has not 1ce fallen by itself since last rewrite. )
asciilifeform: so when changing www search algo, same will apply to bot
asciilifeform: i wrote the bot to use the www end's search function, so as to avoid having two duplicated searchtrons in there
asciilifeform: mircea_popescu: ^ is good example, paginator not in there yet
mircea_popescu: !qs the
diana_coman: !q s too short
snsabot: asciilifeform: time since my last reconnect : 0d 1h 1m
asciilifeform: btw ok to put spaces,
mircea_popescu: maybe. tnh i dunno. maybe we end up standardizing some kinda help format.
asciilifeform: mircea_popescu: think ought to 'per cmd' help eh.
snsabot: asciilifeform: time since my last reconnect : 0d 1h 0m
snsabot: asciilifeform: this command is not yet implemented.
asciilifeform: yw diana_coman . it is not difficult to add knobs, whole thing is <1000ln.
diana_coman: fwiw I noticed and very much liked that knob, asciilifeform !
asciilifeform: ( rationale was, there are chans where folx speak erry coupla days or so, and handy to be able to jump back immediately w/out flipping pages )
mircea_popescu: ahh, it's so nice to be back!
asciilifeform: tells you when was last line spoken (clickable, will go to it)
asciilifeform: right nao i have 1 knob that phf did not, the activity ticker on top of page
mircea_popescu: ima have to train myself into comprehension that indeed, we've finally come to sanity, can now add knobsd.
asciilifeform: the beauty of 'publish the fucking src' is that in principle can then add knobs.
mircea_popescu: asciilifeform, i asked for it ; and then i didn't find it that useful either. i'm guessing i just don't understand what wuld be useful exactly.
asciilifeform: but theoretically can be made
mircea_popescu: diana_coman, the plan was to first federate, then move. we're not progressing re federation.
asciilifeform: mircea_popescu: phf's had the activity graph. i did not clone it
asciilifeform: i recall the 'ro digg' aha
mircea_popescu: it might be nice, have something like that for loggers, you know. though i dun specifically know what.
mircea_popescu: it's a data visualization, back then i ran a ro digg ; that's users / activity.
diana_coman: tbh I also don't quite grok what's the status re irc as there was the plan to move and so on
asciilifeform: for instance right nao anyone leasing a box on piz, can hear the pw
mircea_popescu: we consider this.
mircea_popescu: as crazy joe davola once said, "i like to encourage intruders"
diana_coman: mircea_popescu: no idea yet; and on one hand there's merit to "other than pizarro" but onth ugh where "other"
asciilifeform: theoretically can sit it down behind znc, and run sslism to fleanode, but i fuckinhate sslism
mircea_popescu: asciilifeform, honestly i wouldn't worry about this, let's see who steals it.
asciilifeform: current bot -- the latter
diana_coman: well, I surely plan to run another bot myself, yes