log☇︎
451400+ entries in 0.293s
decimation: as in, what type?
ben_vulpes: but how does it know what to serialize the block message into
decimation: yes, that's what the template does
ben_vulpes: just strikes me as a bit of black magic that a CDataStream can deserialize itself into a CBlock with just an overloaded right shift operator
decimation: ugh template shit
ben_vulpes: asciilifeform, trinque, decimation: how does http://btc.yt/lxr/satoshi/source/src/serialize.h#1141 implement serialization such that it writes correctly-shaped data to CBlock here: http://btc.yt/lxr/satoshi/source/src/main.cpp#2193
mircea_popescu: the fsf, and the "foss" concoction, are going to be gone before the usg.
assbot: Logged on 28-06-2015 04:59:46; decimation: I find it unlikely that the fsf will outlive him
mats returns to the shadows
assbot: Logged on 28-06-2015 04:49:18; asciilifeform: and that we're more or less stuck with this until mythical 'sanity fork'
mircea_popescu: http://log.bitcoin-assets.com/?date=28-06-2015#1180042 << pretty much. HOWEVER, more movement has been made towards sanity in this chan over the past six months than since 2009. ☝︎
assbot: Logged on 28-06-2015 04:31:39; trinque: ben_vulpes: y'know no sooner do I even consider relocating to the land of capitalism and Portland's like "but dude! an army of titties on bikes!"
mircea_popescu: http://log.bitcoin-assets.com/?date=28-06-2015#1179994 << topless ? ☝︎
mircea_popescu: ben_vulpes just more nonsense to filter.
asciilifeform: it's about 1994 here in this room
asciilifeform: posting from them right now, l0l
asciilifeform: i'm still in the 90s
mircea_popescu: asciilifeform you're thinking the 90s.
ben_vulpes: what is the 10x to 100x refer to - not the ip cost at a 'home' terminal?
asciilifeform: yeah, that this.
ben_vulpes: pardon my thickness, but what is "this" now?
asciilifeform: do this at a serious dc and you have half a dozen 'bofhs' going postal
assbot: Logged on 28-06-2015 05:28:41; mircea_popescu: understand - it's 10x to 100x more expensive to do a 213.130.38.26 on random home cable than it is on servers
asciilifeform: http://log.bitcoin-assets.com/?date=28-06-2015#1180198 << can't entirely agree here. every so often folks discover carriers diddling 'consumer' traffic in obscene ways (adding tracking crud to http, even) having gone by for years at a time with not a squeak from the chumps ☝︎
asciilifeform: but of necessity many of the boxes involved in the ramp-up will be pc.
asciilifeform: unquestionably the whole 'p2p' thing is the right and proper objective
mircea_popescu: understand - it's 10x to 100x more expensive to do a 213.130.38.26 on random home cable than it is on servers ☟︎
asciilifeform: (it'll have to swap.)
asciilifeform: until we actually get limited mempool, shoot fragging in the head, etc. pogo will be a bit slow
mircea_popescu: but on consumer internet terminals.
asciilifeform: but overall i like the idea of a dc box with bitcoind and LITERALLY NOTHING ELSE
mircea_popescu: pogo = dedicated node ; this = stick everywhere
mircea_popescu: also, kernel is a good idea maybe for pogo. not for a proper server. you want people to add it to all their noncritical dedis. ☟︎
asciilifeform: y'know how some hoster folks let you drop an iso in when you get the thing
mircea_popescu: so that can be v 1.1
mircea_popescu: we'll have to bootstrap ourselves.
asciilifeform: e.g. a kernel that boots right into, this.
asciilifeform: mircea_popescu: i was actually thinking of doing for ibmpc something like what i did for pogo
mircea_popescu: doesn't help us on pogo, but that's a diff story.
mircea_popescu: asciilifeform to answer earlier q : i think for thius purpose (running on dulap etc) static glibc is exactly right.
mircea_popescu: properly made sausage actually floats. therefore not a fish
asciilifeform: 'the most delicious fish is the sausage.' --my grandfather
mircea_popescu: "you mean sole ?" "o yeah that's the one!"
asciilifeform: (no patches necessary to the bitcoind proper beyond what was published previously)
asciilifeform: and would sorta like other folks to read the damn thing
asciilifeform: mircea_popescu: and yes, was gonna do this after certain tests. such as the one that is presently failing
assbot: Logged on 28-06-2015 02:30:30; asciilifeform: if mircea_popescu asks, this turd will be deployed to dulap (s.nsa box)
asciilifeform: doesn't take high tech magic
asciilifeform: pull down the tarball from ml, try
mircea_popescu: you actually managed to get glibc to static ?
asciilifeform: but the thing is running on my box now.
assbot: Logged on 28-06-2015 02:04:04; asciilifeform: There is no dynamic section in this file.
asciilifeform: dropping the connection would get bitcoind rotating to next node
asciilifeform: was just thinking this!
mircea_popescu: this entire experience painfully reminds me of the famous phoneline click somehow
asciilifeform: wouldn't be the first
ben_vulpes: that'd be threadsockethandler?
asciilifeform: ben_vulpes: but if you can write your own asynchronous everything, you can finally get the coveted msdos port...
ben_vulpes: http://btc.yt/lxr/satoshi/ident?_i=fShutdown << for log readers interested in how precisely bitcoin shits its pants when told to go to sleep
asciilifeform: ben_vulpes: is able to take new connections while doing said loop
asciilifeform: and i die of old age 3-5 times
asciilifeform: at this rate, the thing will sync some time after the eeproms in my computer lose 9/10th of their bits
asciilifeform: but if you ask me, this falls squarely under 'do this after it fucking ~works~' as described by mircea_popescu earlier.
asciilifeform: i was reluctant to throw gccisms in there, but if we go and start building on mircea_popescu's soviet pdp11 we can always roll equivalent macros for that
ben_vulpes wonders at actual utility of a patch shaped like this
assbot: Common Predefined Macros - The C Preprocessor ... ( http://bit.ly/1JqXXxK )
assbot: Standard Predefined Macros - The C Preprocessor ... ( http://bit.ly/1JqXV95 )
asciilifeform: (can't recall off top of head what, precisely. rtfm...)
asciilifeform: ben_vulpes: there's gcc-specific stuff that folds the line and # into the print
ben_vulpes: is there a sane cpp logging thing that could be used in place of these print statements?
ben_vulpes: (curse these jackholes who littered the source with printstatements that include function names)
decimation: not much use to a license if you don't intend on enforcing it
decimation: yeah, as far as I can tell gpl only really 'sticks' to fsf software directly
asciilifeform: it is already seeping through the cracks.
decimation: really, he's already lost his war for the reasons you harp on re: bedrock
asciilifeform: on something which in turn is buildable ?
asciilifeform: won't ask for it to be buildable ?
decimation: but I recognize that rms' stand is probably the only way
decimation: gpl is pretty weak sauce unless you have a barrel of lawyers to help you enforce it
asciilifeform: in the sense of keeping the 'reasonable' folks at bay? doubtful
asciilifeform: in the sense in which it is now alive? sure
decimation: I find it unlikely that the fsf will outlive him ☟︎
asciilifeform: it exists, was built, to help him die.
asciilifeform: i will note that llvm (since i mentioned it again) is upon the most cursory examination a piece of shit, that wasn't even built to 'be good' - but simply with the one and only purpose of killing gcc.
asciilifeform: but this had a cost, yes.
asciilifeform: the one thing the fucking parasites couldn't effectively plagiarize without leaving trace
asciilifeform: gcc is the original lonely fort keeping the whole gpl thing alive
decimation: sure, he wants to keep people from mkaing a bsd gcc
ben_vulpes: erm i thought bitcoind *did* service one at a time, looping over sockets
asciilifeform: decimation: rms did not want it to break apart into readily-plagiarizable pieces.
assbot: Logged on 30-08-2014 03:37:06; decimation: "The existence of LLVM is a terrible setback for our community precisely because it is not copylefted and can be used as the basis for nonfree compilers — so that all contribution to LLVM directly helps proprietary software as much as it helps us.'"
asciilifeform: http://log.bitcoin-assets.com/?date=30-08-2014#816009 << see old thread. ☝︎
decimation: sure, to keep folks from re-writing gcc
decimation: no and as I recall there's hand-written asm in openssl
asciilifeform: rms ~deliberately~ nudges it, whenever he can, in the direction of being as opaque as possible
asciilifeform: quite the contrary.
asciilifeform: gcc, as one might imagine, does not go out of its way to help this be a readily practical thing
asciilifeform: had to put it down, too many other things going on
asciilifeform: tried to come up with a mechanical mapping back from bin to what produced each section in the elf
asciilifeform actually began there