Show Idle (>14 d.) Chans

← 2018-02-24 | 2018-02-26 →
deedbot: << Qntra - Chicoms Set To Lift Presidential Term Limits, Attempt To Putinize Xi Jinping
mircea_popescu: putinize lol. what a great word.
mircea_popescu: and in other news, diana utterly owns "ossasepia" in teh google. wd eh,
asciilifeform: mircea_popescu: sounds quite good to me; let's do it
asciilifeform: and i was thinking nearly same thing re the price.
asciilifeform is on a remote console where he doesn't see PMs; but will check back later today.
mircea_popescu: ben_vulpes how about you lot take three dozen FGs at .45 ? (cash ; or at your option .5 BTC worth of convertible bonds for s.nsa).
jhvh1: shinohai: The operation succeeded.
deedbot: << Qntra - Apple Hands Over More Keys
BingoBoingo: shinohai ty, edited "Ya Socialismo" to "Oriental socialism", different meat substrates and gotta keep the Chicoms burning time on Engrish lessons.
shinohai: aha ty
BingoBoingo: "Ya Socialismo" is what stoner latinos aspring to Jay and Silent Bob want. Oriental Socialism is more or less an anthill.
mircea_popescu: is this "socialismo ya!" ?
BingoBoingo: Here they flip the words. Apparently the Uruguayos suck at grammar
BingoBoingo: Or so the other immigrants tell me
BingoBoingo: I am just the fellow who reads the graffiti
BingoBoingo: Thinking about it I recall seeing it both ways
mod6: Another day, another 7" of snow.
mircea_popescu does not envy his northern friends.
phf: << i thought that it was just a helpful warning, but upon reflection i realized that this is actually a bug. investigating it further i took a wrong direction on a diff's command line flag switch, and as it stands if you see this warning it
a111: Logged on 2018-02-25 01:46 hanbot: phf : patched fix for vdiff produced a genesis for me but complains about 'no newline at end of file' for some 67 files. not clear to me whether i oughta be concerned/what gives.
phf: means that vdiff is producing invalid vpatches
mircea_popescu: nice find then.
mod6 looks
phf: i have a btcbase based test harnes, but it of course also didn't catch the issue, because it doesn't know about `\ No newline ...' in band thing either. it treats that line as a textual prelude to the next hunk which is a valid patch format
phf: in any case i'll produce a fix by wednesday, but not before. this requires careful work
phf: mod6: an important thing to add to v might be a check ~after~ the press, which ensures that final states of files have shasums that the vpatch chain claims they should
phf: i think right now the way things are implemented (that goes for btcbase anyway) you can have a chain, that links correctly, but actually doesn't press to anything at all like what the hashes claim
phf: and in vaguely related, while writing a chain verifier for btcbase, i found that we have utf-8 in genesis
mod6: what my vtron does is this: as it iterates through the list of patches to be pressed, after each vpatch pressed, it checks that the output file sha matches the expected.
mod6: Here's a snippit of that code:
mod6: gah, UTF8!?
mircea_popescu: 23962 /*Boost has a year 2038 problem— if the request sleep time is past epoch+2^31 seconds the sleep returns instantly.
mircea_popescu: who put that abomination in there and why!
mod6: for crying out loud
mod6: I can confirm.
mircea_popescu: whose process even produces microsoft publishing dashes
phf: mod6: right, that's the right thing to do; for a longest time i was still pressing with the original, but i suspect most people (except maybe ascii) use at this point
phf: i still have muscle memory for dashes, quotes and a few math symbols on mac en layout. (i would even differentiate between russian << .. >> and en `` .. ''). this stuff seemed like the future in the early 2000s..
mircea_popescu: the guillemets are actually french not russian eh!
phf: easiest way to get a rise out of my iranian colleagues is to suggest that samovar and olivier salad were not "invented" by iranians
mircea_popescu: so is it your comment typed out on a mac, is the idea ?
phf: right, that it could've been typed out by hand
mircea_popescu: well could've. was it or wasn't it ?
phf: you're assuming that i'm trying to say something useful!
mod6: Looks like this UTF-8 problem with the dashes goes all the way back - and looks like we've all signed it with that in there:
mod6: I pulled the original posts from the ML, and then started looking for "e2 80 94" which I ~think~ is the bad char.
mircea_popescu: chiefly because we didn't actually understand at the time what we were doing.
mod6: Yeah, but furthermore, I don't think the "no UTF8" rule had been established at that point.
mod6: I'd have to dig for that to prove tho.
mircea_popescu: im pretty sure it was not.
mircea_popescu: i'm guessing the only way out of this predicament is to have an actual regrind, into a new genesis, of all the patches we actually want.
mod6: ^ indeed.
mod6: And rather than do a fast job to complete this ; I think the right approach is to comb through the entire thing and ensure that no UTF-8 is hiding.
mircea_popescu: well... think it through for sure.
mod6: *nod*
mod6: I'm on it, The Foundation is going to close out the month with the SoBA. This work will be high priority post month-end.
a111: Logged on 2018-02-25 20:29 mod6: And rather than do a fast job to complete this ; I think the right approach is to comb through the entire thing and ensure that no UTF-8 is hiding.
mod6: Thanks phf, I think I'm gonna conjure something up over here too just to see.
phf: yeah, of course
mod6: (doesn't hurt if others do this as well!)
BingoBoingo: <phf> easiest way to get a rise out of my iranian colleagues is to suggest that samovar and olivier salad were not "invented" by iranians << I enjoy telling the Argentines their mate has too many sticks and twigs
ben_vulpes: !!up drunk_foxx_
deedbot: drunk_foxx_ voiced for 30 minutes.
drunk_foxx_: What does it mean to be voiced? :thinking:
ben_vulpes: drunk_foxx_: that you can talk; dig
ben_vulpes: !!down drunk_foxx_
ben_vulpes: see, now nothing
ben_vulpes: !!up drunk_foxx_
deedbot: drunk_foxx_ voiced for 30 minutes.
ben_vulpes: anyways, who is your daddy and what does he do
ben_vulpes: trinque: possible to get's list of deeds to link to my fingerprint instead of the signing key fingerprint?
ben_vulpes: it is of the lowest priority
trinque: I am really opposed to putting any work whatsoever into managing gpg's unnecessary complexity
trinque: somebody gimme a better rsatron and I'll use that instead.
deedbot: << PizzaroISP - Pizarro, incorporated!
mod6: <@deedbot> << PizzaroISP - Pizarro, incorporated! << cheers!
mircea_popescu: ya srsly, can't have arbitrary relinkings
a111: Logged on 2018-02-25 19:55 mircea_popescu: who put that abomination in there and why!
mircea_popescu: unloadable non-page.
mircea_popescu: im sure all https github is the same.
phf: oh it's the https
mircea_popescu: i mean this literally, they went offline at some point.
phf: From 82a10c81707dcff5ee24dec7ef7ebf8eccfded03 Mon Sep 17 00:00:00 2001
phf: From: Gregory Maxwell <>
phf: Date: Sat, 10 Mar 2012 16:05:28 -0500
phf: Subject: [PATCH] Resolves issue #922 - "wallet passphrase timeout of several
phf: years doesn't work"
phf: 2^31 milliseconds is only about 25 days. Also clamps Sleep() to 10 years,
phf: because it currently sleeps for 0 seconds when the sleep time would cross
phf: 2^31 seconds since the epoch. Hopefully boost will be fixed by 2028.
phf: ---
phf: src/bitcoinrpc.cpp | 6 +++---
phf: src/util.h | 4 +++-
phf: 2 files changed, 6 insertions(+), 4 deletions(-)
phf: diff --git a/src/bitcoinrpc.cpp b/src/bitcoinrpc.cpp
phf: index 68cc17b518dc..5571c344a322 100644
phf: --- a/src/bitcoinrpc.cpp
phf: +++ b/src/bitcoinrpc.cpp
phf: @@ -1548,7 +1548,7 @@ void ThreadTopUpKeyPool(void* parg)
phf: void ThreadCleanWalletPassphrase(void* parg)
phf: {
phf: - int64 nMyWakeTime = GetTimeMillis() + *((int*)parg) * 1000;
phf: + int64 nMyWakeTime = GetTimeMillis() + *((int64*)parg) * 1000;
phf: ENTER_CRITICAL_SECTION(cs_nWalletUnlockTime);
phf: @@ -1584,7 +1584,7 @@ void ThreadCleanWalletPassphrase(void* parg)
phf: LEAVE_CRITICAL_SECTION(cs_nWalletUnlockTime);
phf: - delete (int*)parg;
phf: + delete (int64*)parg;
phf: }
phf: Value walletpassphrase(const Array& params, bool fHelp)
phf: @@ -1619,7 +1619,7 @@ Value walletpassphrase(const Array& params, bool fHelp)
phf: "Stores the wallet decryption key in memory for <timeout> seconds.");
mircea_popescu: dude never heard of pasting
phf: CreateThread(ThreadTopUpKeyPool, NULL);
phf: - int* pnSleepTime = new int(params[1].get_int());
phf: + int64* pnSleepTime = new int64(params[1].get_int64());
phf: CreateThread(ThreadCleanWalletPassphrase, pnSleepTime);
phf: return Value::null;
phf: diff --git a/src/util.h b/src/util.h
phf: index 31d3275290e4..d5e8a71cd0f6 100644
mircea_popescu: !~later tell phf sorry about that ; but srsly either bvulpes link or simply From: Gregory Maxwell <> would have sufficed.
jhvh1: mircea_popescu: The operation succeeded.
phf: now this could be misconstrued as a passive aggressive "how about now! can you read it now!", but my middle click copy paster failed, and then it's znc that's throttle sending. i figured i could kill znc, but mp was faster on the draw
mircea_popescu: shit occurs.
phf: anyway, yes, it's maxwell
mod6: Thanks for digging that up phf! seems to have trouble pulling the patch directly from shithub. I've grabbed it and posted it myself for the ages:
ben_vulpes: !~ticker
jhvh1: ben_vulpes: Bitfinex BTCUSD ticker | Best bid: 9648.3, Best ask: 9655.5, Bid-ask spread: 7.20000, Last trade: 9653.5, 24 hour volume: 30808.8059532, 24 hour low: 9280.4, 24 hour high: 9840.0, 24 hour vwap: None
ben_vulpes: !~ticker --market all
jhvh1: ben_vulpes: Bitstamp BTCUSD last: 9689.21, vol: 10584.23934686 | Bitfinex BTCUSD last: 9655.1, vol: 30808.46580311 | Kraken BTCUSD last: 9692.2, vol: 7413.36383628 | Volume-weighted last average: 9668.13248173
mod6: ty
mod6: bah, sorry. ignore this ^
trinque: mircea_popescu et al, new deedbot command is live:
trinque: !!ledger
mircea_popescu: !!ledger
trinque: well son of a bitch
trinque: all all the commands are limited to pm, one sec
mircea_popescu: a, that's good.
mod6: heh, does !!ledger encode the whole block chain for a guy to get?
mircea_popescu: i expect it's the story of !!balance
trinque: no, it's the wallet ledger for you
mod6: nice trinque ; so all movements per wot account 'eh?
mircea_popescu: lel at fractionals.
trinque: fuggin lisp.
trinque: !!balance
trinque: huh, got some wires crossed then. bbl
trinque: !!ledger
trinque: alrighty, got some timestamps to make more human, but there it is.
ben_vulpes: !!ledger
trinque: spent a good amount of time today getting the logbot-command class working with ben_vulpes' multiple channel support
ben_vulpes: sweet trinque does that mean i can get deedbot in #pizarro then?
trinque: I don't see why not, after I make sure other parts (voicing etc) work
ben_vulpes: no need for voice over there yet, but sure no rush
trinque: next item on the list for the wallet is subscription processing.
trinque: don't know if that's interesting to pizarro, but it's coming.
ben_vulpes: neato
trinque: there'll be some fee involved for those, would like to hear what interested users are willing to pay.
mircea_popescu: trinque what functions would work out of chan though ? seems the ones that don't work in pm (!!deed say) shouldn't ; and the ones that work in pm... welll...
trinque: he's asking, ben_vulpes what functions did you want
mircea_popescu: i suppose if you actually make the item then that'd prolly want its own #pizarro list ; but also by now it's such a complex piece of gnarl might even want to separate a rss-bot out of deedbot
a111: Logged on 2018-02-19 14:56 mircea_popescu: trinque should prolly make the public blog rss thing a l1 blackballable item. ie anyone deedbot trusts can ask (in channel) for any item to be discontinued -- and it is.
ben_vulpes: having wallet commands would be handy, unless mircea_popescu wants to see pizarro invoices/payments in #trilema
trinque: already is separate
trinque: the whole thing is modular.
mircea_popescu: im not sure i want to have you maintain rss-shenanigans as part of deedbot proper, they're widely disjunct seriousness.
ben_vulpes: does invoice work in pm?
mircea_popescu: ben_vulpes all those pay withdraw etc work in pm afaik.
ben_vulpes: yeah, tis what i thought as well
trinque: all wallet ops work in pm
trinque: anyhow, this is definitely not some piece of gnarl.
mircea_popescu: the rss stuff ?
mircea_popescu: how can it not be, it has to read random websites.
trinque: I don't understand the question.
trinque: "deedbot" is something like 10 separate services.
mircea_popescu: running on the same box ?
mircea_popescu: ah. so then it's a name-only thing, might as well have a rss bot name on the rss bot item
mircea_popescu: otherwise you know, if the process running on the box where you do payments ALSO curls every 5 minutes... not so great.
trinque: it already does not do this.
mircea_popescu: very nice and good.
trinque: there's a publisher/subscriber model in use, services being subscribers which are notified of new messages they might be interested in
trinque: they then in turn can publish responses that exit the thing connected to IRC
mircea_popescu: so there's three classes of things : 1. deedbot "proper" let's say, doing !!deed + the wot + the whole !!up system ; 2. deedbot payments ; 3. deedbot fluff, doing things like rss.
mircea_popescu: now as to 1, i don't think it should be out of chan, on one hand because "secret" deeds plainly don't make sense ; on the other because the voice thing is ~probably~ premature in channel that is both new and customer facing. but in principle certainly could in the future have local l1 with voice model for diff channel, i think.
mircea_popescu: 2 already works in pm, so there's no absolute reason shouldn't work in any other channel, if you're willing to put the work into that.
mircea_popescu: and 3 really should be a separate bot i think.
mircea_popescu: conceivably 2 should be a separate bot also ; but a) as long as tyou're maintaining both and b) they really need very much the same level of security and assurance, it doesn't seem worth the hassle.
trinque: kinda overcomplicates my way of handling IRC to separate into 3 connections to IRC.
mircea_popescu: well, i was more probing if you actually want to implement the whole rss thing yourself or pass it along to some noob.
trinque: ah. if a worthy noob is interested in handling rss, he may.
mircea_popescu: course as it happens... pete_dushenski just decided he's not really interested... maybe was looking for a way in ?
a111: Logged on 2018-02-18 02:41 TomServo: Wish I could help with any current woes, if that's possible lemme know.
mircea_popescu: joecool ? you doing anything these days ?
mircea_popescu: douchebag ? spyked ? esthlos ? ave1 ? mats ?
mircea_popescu: jesus christ look at that, plenty of list
trinque: upstack a bit, I can see the argument for separating functionality into meaningful (nick)names
mircea_popescu: well for one thing one can't /ignore the rss cuz of deedbot. it's a little much.
trinque: under the hood, the wallet's called "mensarius"
mircea_popescu: lol. srsly, what's wrong with paybot
trinque: ah other than it being in that language you keep in your torture room, nothin!
mircea_popescu: so apparently $$$ can't be a nick
trinque: untaken even
mircea_popescu: of course. that's why i say, why not own the central word.
trinque: works for me. I'll snag it
mircea_popescu: freenode is about as sparse as a camho's brain, no need to get clever.
mircea_popescu: deedbot wasn't fucking "mercurius" after all...
mod6: ben_vulpes: *thumbsup*
trinque: douchebag: << what came of this?
a111: Logged on 2018-01-26 20:20 ben_vulpes: douchebag: do yr marching orders make sense?
lobbes: !!ledger
lobbes: neato trinque. very useful
lobbes: mircea_popescu: while I'm still working through my self-directed remedial mathematics studies to understand merkle-trees at a functional level, what I have learned so far got me thinking:
lobbes: so, here's what I understand thus far re: my goal (plox to correct me): essentially construct a 'tree' of the set of all webpages archived. the 'root' will be the hash of all branches. where I'm still fuzzy is how the remaining branhces/nodes are to be structured (though perhaps this may become clear to me upon further study)
lobbes: I am guessing it will be something like: 1) root 2) nodes for 'distinct url', each who has child nodes representing 3) historical instances of that 'distinct url', each who has child notes representing 4) the collection of files that make up that webpage
lobbes: and if 4) is correct, then should that be further broken down into a webpage hierarchy? Because if so, we may be limited using as a middle-man as the zips it spits out do not have the file structure intact; always 'flat' (e.g. this archived orlol:
mircea_popescu: lobbes you understand what the AST is for a [webpage
mircea_popescu: lobbes you understand what the AST is for a [web]page [of code] ?
ave1:, I'm working on custom runtime / zero footprint ada code
a111: Logged on 2018-02-26 00:39 mircea_popescu: douchebag ? spyked ? esthlos ? ave1 ? mats ?
lobbesbot: ave1: Sent 1 week, 3 days, 5 hours, and 53 minutes ago: <asciilifeform> so i finally tested, exhaustively, and turns out you were completely correct re the inline pragma, gcc ~sometimes~ ( when it wants... ) inlines when it is not given in the .ads . i'ma fix this in ch12, this find deserves own chapter, with re-done timings for errything
ave1: Also, as it happens on some bot code in ADA, but this is slow going
← 2018-02-24 | 2018-02-26 →