log☇︎
380400+ entries in 0.254s
BingoBoingo: But seriously none of the republicans are going to stop before the first primaries because they gotta build their profile in public to become TV hosts
BingoBoingo: ^ That CLinton was well known
gribble: George Clinton Parliament Funkadelic official website: <http://georgeclinton.com/>; George Clinton (musician) - Wikipedia, the free encyclopedia: <https://en.wikipedia.org/wiki/George_Clinton_(musician)>; George Clinton and Parliament Funkadelic - Facebook: <https://www.facebook.com/georgeclintonpfunk/>
mircea_popescu: let's limit this discussion to "after tv"
BingoBoingo: George Clinton was well known at the time though
mircea_popescu: "that guy with the hot wife"
trinque: Trump and Carson practically fist-bumped on stage in that CNBC "debate"; I wouldn't be surprised to see them end up running together
BingoBoingo: Obama, Clinton, Truman...
mircea_popescu: i would imagine the republicans understand enough of their small penis issues to not imagine they can pull democrat tricks.
BingoBoingo: NEarly all the repubes are still ruining
BingoBoingo: <mircea_popescu> who the fuck are they gonna pick ? jeb ? the doc nobody heard of ? << Doctor Stoner is actually leading now. Rubio appears to be the establishment alternative now that Jeb is dead in the water
trinque: bragged about HP's role helping the NSA under her watch, too
asciilifeform: the one that killed hp
asciilifeform: wat about the chick
mircea_popescu: who the fuck are they gonna pick ? jeb ? the doc nobody heard of ?
mircea_popescu: 4:1 on trump ? cmon.
liquidassets: Exactly BingoBoingo except I still have to go back and read more to get the full context
deedbot-: [BitBet Bets Bets] 7.47500000 BTC on 'No' - Donald Trump gets Republican Nomination - http://bitbet.us/bet/1206/donald-trump-gets-republican-nomination/#b40
BingoBoingo: For honeydicking, I am less certain the proper course of action
BingoBoingo: <liquidassets> ... << full disclosure via logs is most correct thing for honeytrapping no? << Yes, like that time I blogged the FBI visit to my front door
liquidassets: anytime you're opened up to liability, the one who documents best seems to win
liquidassets: “only the Übermensch will have the strength to fully accept all of his past life, including his failures and misdeeds, and to truly will their eternal return” grepped straight from the logz
trinque: whole thing is a massive sin
assbot: Logged on 04-11-2015 18:50:13; *: asciilifeform would like input from his WoT re: what is the correct thing to do with obvious usg honeytrapping as described earlier.
liquidassets: http://log.bitcoin-assets.com/?date=04-11-2015#1316034 << full disclosure via logs is most correct thing for honeytrapping no? ☝︎
mircea_popescu: what the fuck these fucking idiots and theyr css. should be hanged.
mircea_popescu: so the fucking web idiots have come up with a novel css hell that chokes my browsers.
asciilifeform: why ever any time else?
trinque: asciilifeform: thinking the kernel's automount mech. for pogo's external disk slot?
asciilifeform: (see old threads where we beat this to a liquid pulp)
mircea_popescu: but not implemented thus currently.
asciilifeform: of a mechanism which ought to work this way
asciilifeform: everything on your box that can safely behave this way, pretty much already does.
asciilifeform: mircea_popescu: an allocator that knows that it is acceptable to lose some old bits is called a circular buffer and behaves like a looped tape, yes
mircea_popescu: back to tape : not to mention that since these these things are virtualized, you could in principle specify how you wish the available mem to be split among the various rotating tapes.
phf: apropos this conversation, emacs as init http://informatimago.com/linux/emacs-on-user-mode-linux.html, prototyped using "user-mode-linux", http://user-mode-linux.sourceforge.net something that could be used for bitcoind
mircea_popescu: kinda have a) a nix ; b) a bitcoin and c) an obvious pogo-merge for these two.
mircea_popescu: trinque the eventual end goal is to have bitcoinix. as alf says, no userland at all. nor really a kernel in the common sense.
phf: mircea_popescu: i honestly thought it was a thin wrapper around call to unix mkdtemp, glad i checked
trinque: I suppose given you can fire up networking entirely from within the linux kernel, init=bitcoind may be all that is needed at some point
mircea_popescu: and for this same exact reason, as alf would say, "cement".
assbot: Logged on 05-11-2015 03:16:02; phf: mod6: you might be right not wanting to `use`. i decided to look at file::temp to see how they do ffi, but instead it's a custom perl written blob. probably reasonable to use, but.. in any case i recommend at least conforming to the api of mkdtemp("/tmp/fooXXX") => /tmp/fooAj5. i took a stab at a sample code, http://paste.lisp.org/display/158520, but there are some other things to keep in mind,
mircea_popescu: http://log.bitcoin-assets.com/?date=05-11-2015#1316780 << i have no idea why reasonable to use. ☝︎
phf: system message buffer operates that way (i.e. what you get with dmesg)
mircea_popescu: actually the rotating tape is not such a bad idea. not just for pogo-log, but in general. wild notion, but, asciilifeform how about memory allocator that simply overwrites the beginning ?
assbot: Logged on 05-11-2015 03:00:13; trinque: no, but my shits do not flush the toilet for themselves....
assbot: Logged on 05-11-2015 02:54:26; asciilifeform: on a machine like pogo it does not make sense to produce the log at all
mircea_popescu: http://log.bitcoin-assets.com/?date=05-11-2015#1316749 << the principle of this is sound. if you're debugging on pogo you've got more pressing problems than the log. such as your psychotherapy appointment. ☝︎
trinque is curiously googling what init=bitcoind would take
mircea_popescu getting there
asciilifeform: mircea_popescu: then you will love the current thread...
mircea_popescu: http://log.bitcoin-assets.com/?date=05-11-2015#1316733 << we applaud this notion. ☝︎
asciilifeform: trinque: neither really belongs there.
trinque: init on the same device is taking about 300kb more than runit, as on this thing I have the former starting the latter
mircea_popescu: this is better than daytime dramaz
gribble: Bitfinex BTCUSD ticker | Best bid: 445.21, Best ask: 445.3, Bid-ask spread: 0.09000, Last trade: 445.2, 24 hour volume: 285121.27754096, 24 hour low: 366.66, 24 hour high: 504.0, 24 hour vwap: None
asciilifeform: ... not that the thing is in any way close to being pogoizable
asciilifeform: but it does NOT make sense for it to evolve in the direction of being 'a good unix util'
phf: in that case you'd better truncate it doing run too, because if that pogo doesn't go down, there's no cap on the log
trinque: bout two orders of magnitude north of that
asciilifeform: moreover, since unix has no future, but unfortunately we are stuck with a junkyard of rusting x86 and arm boxes, it makes sense for bitcoin to evolve in the direction of being an os
asciilifeform: if that.
asciilifeform: phf: log in the airplane 'black box' sense only.
phf: should there even be logs on pogo?
asciilifeform: my argument here is not that the behaviour of the classic debug log is a beautiful thing, but that any change to bitcoin that MAKES MANDATORY the presence of so much as an 8kB proggy on the machine, is LETHAL
asciilifeform: if it's 8kb that's too much.
trinque: one sec while I go find the ram usage of runit on a similar device
asciilifeform: (why waste the precious ram on that ??)
asciilifeform: because ANYTHING is too big
asciilifeform: trinque: i don't even know how big it is, but will tell you that it is too big for pogo.
asciilifeform: see, the whole 'unix philosophy' thing would be considerably more interesting if we had a machine where it actually... worked.
trinque: runit runs things in precisely this manner, and is the smallest init guy I know
phf: e.g. http://search.cpan.org/~dagolden/File-Temp-0.2304/lib/File/Temp.pm#safe_level (particularly if you read the paragraph under HIGH)
phf: mod6: you might be right not wanting to `use`. i decided to look at file::temp to see how they do ffi, but instead it's a custom perl written blob. probably reasonable to use, but.. in any case i recommend at least conforming to the api of mkdtemp("/tmp/fooXXX") => /tmp/fooAj5. i took a stab at a sample code, http://paste.lisp.org/display/158520, but there are some other things to keep in mind, ☟︎
trinque: asciilifeform: why not have the thing barf everything to stdout and stderr and let whatever's out there do with it what it pleases; opposed to unix-y small proggies that do one thing here?
asciilifeform: i have machines here where the size of the log approaches the size of the blockchain.
trinque: I was thinking of a lower log level than whatever I'd expect from "debug" but in that case yes
asciilifeform: a) a wedged node still pisses into the log b) log contains tx ids, many of which correspond to nothing that ends up in the blocks
trinque: this log file is a small fraction of the size of the blockchain which is also constantly being written
asciilifeform: and it was not deemed acceptable for the box to ever not be recording.
asciilifeform: because guess what, finite tape fits in box.
asciilifeform: trinque: you do know that aircraft crash boxes traditionally ran the tape in a loop, right ?
asciilifeform: but this is pure gold,
asciilifeform: how do you propose to do this ?
asciilifeform: yet it still makes sense to have a small 'crash recorder' sort of log, for when it shits itself.
trinque: no, but my shits do not flush the toilet for themselves.... ☟︎
trinque: when it is sitting on a system where the user can decide to do this if he chooses
asciilifeform: trinque: you don't ever flush your toilet ?
trinque: I am commenting on the idiocy of a thing *ever* taking it upon itself to erase the history of its own operation
asciilifeform: we threw it out, because it is really a harmful thing, but later ended up with an impression that it will have to be - in a separate process, yes - put back, if pogo is to be deployed to civilian households
asciilifeform: trinque: this is reminiscent of the thread concerining upnp.
asciilifeform: on a machine like pogo it does not make sense to produce the log at all ☟︎
trinque: how does the fact that it was written for windows argue for the retardation staying
asciilifeform: thing was written for winblowz, in its heart, by a winblowz programmer, this is beyond dispute.
trinque: if it were the ADA version or even one written with as few deps as possible, I'd grant you're right
asciilifeform: if you recall, the original wasn't even assured of running on a unix
asciilifeform: i must confess that i disagree. the thing runs, as it might, self-contained, to the extent practical.
trinque: the rotation was put in the debug log; that is mega-autism
phf: mkdtemp is a canonical way of creating a temporary directory on unix, that'll also ensure a) no collisions b) proper permissions, i.e. 700
trinque: asciilifeform: yes I would prefer that it only do one thing, but you are correct that it calls this only if !fDebug