Hide Idle (>14 d.) Chans


← 2019-09-11 | 2019-09-13 →
lobbes: ok, well confirmed that both the znc2tmsr and irssi2tmsr converters output to spec re: action lines. Looks like it is just the raw log knob. I will dig into it, fix, and patch
lobbes: wow, this might end up being the smallest patch ever. forget one line, this is one character in one line! fix tested and worx. patch incoming..
lobbes: well, I was wrong; still small patch, but ended up being > one char. Good thing I retested and reread
feedbot: http://blog.lobbesblog.com/2019/09/raw_line_fixkvvpatch/ << lobbesblog -- raw_line_fix.kv.vpatch
lobbes: asciilifeform: ^ for your testing and approval
lobbes off to bed
mircea_popescu: in other greats outside the walls : https://www.jassa.org/?p=10723 << this dude ~actually sourced~ an extremely rare (online, it is in the basic toolkit of any serious medievalist, excellent late 1800s edition bei Boretius/Krause (BK)) source (Inquisitio de theloneis Raffelstettensis, BK 253) and then translated it.
mircea_popescu: the translation's far from terrible ; the item's important because a) it plainly discusses the slave trade (which, yes, underpinned most of the imperial economy eight years ago) and references a coin that's a.1) unatested anywhere else in the (voluminous, and pompous) imperial paperwork and a.2) isn't the official coin.
mircea_popescu: not to put too fine a point on it : yes slave trade ; no dubaloos.
diana_coman: interesting
mircea_popescu: eight yeaes lol. i mean eight centuries.
diana_coman: I can't find anything about the author(s?) there though, no contact no nothing.
diana_coman: of the blog I mean, ofc.
mircea_popescu: last article from the 9th, though, so he's active.
mircea_popescu: i guess ima leave him a comment.
diana_coman: aha, looks active at least
diana_coman: meanwhile, after ~2 hours, I *finally* finished the log-reading of over-night logs, go me; it just strikes me that with #o I made myself even *more* to read, lol.
mircea_popescu: but of course.
mircea_popescu: spyked, you know what, re img pastebin ? i think im just going to revive ye ancient dtng service
mircea_popescu: asciilifeform, this acceptabru on ye pizarro ?
mircea_popescu: (and since we're doing history : anyone know of a study describing anne hutchinson's self-evident influence upon the concept of a "witch" in the colonies, and retelling the "witch trials" as her necessary epiphenomenal epigonic downstream ?)
mircea_popescu: (long story short, eugenie grandet of the colonies really really wants to impress le petit monsieur vane-charles, if not with a bit of cold pheasant for breakfast then at least with some gossip circle fireworks. needless to say the local bureaucrats weren't impressed with any of it.)
mircea_popescu: trinque, pastes 504 ?
mircea_popescu: also re http://logs.nosuchlabs.com/log/trilema/2019-09-07#1934962 : evidently the correct equation is (27*3+10)^4 = 68574961, because both small and capital letters. 68mn should be plenty.
snsabot: Logged on 2019-09-07 16:45:11 mircea_popescu: 37^4 = 1874161, no way anyone's doing 2mn pastes/week
trinque: looks to be working from here, sure you're using paste.deedbot.org (same IP as deedbot.org)
trinque afk 'til later
asciilifeform: http://logs.nosuchlabs.com/log/trilema/2019-09-12#1935740 << ty lobbes ! i will test & emplace this today
snsabot: Logged on 2019-09-12 02:42:06 lobbes: asciilifeform: ^ for your testing and approval
asciilifeform: http://logs.nosuchlabs.com/log/trilema/2019-09-12#1935742 << pretty great link. surprised to see archaeologist like this writing in engl.
snsabot: Logged on 2019-09-12 04:14:57 mircea_popescu: in other greats outside the walls : https://www.jassa.org/?p=10723 << this dude ~actually sourced~ an extremely rare (online, it is in the basic toolkit of any serious medievalist, excellent late 1800s edition bei Boretius/Krause (BK)) source (Inquisitio de theloneis Raffelstettensis, BK 253) and then translated it.
asciilifeform: http://logs.nosuchlabs.com/log/trilema/2019-09-12#1935756 << it's a great idea imho. but prolly oughta be operated via commands in #t or similar , otherwise no amt of pipe or disk will ever suffice
snsabot: Logged on 2019-09-12 06:21:36 mircea_popescu: asciilifeform, this acceptabru on ye pizarro ?
asciilifeform: ha, pretty light log today.
dorion: http://logs/nosuchlabs.com/log/trilema/2019-09-11#1935621 << thanks, but the arm docs say ARMv8-A has ARMv7-A (AArch32) compatibility
snsabot: Logged on 2019-09-11 14:34:47 asciilifeform: http://logs.nosuchlabs.com/log/trilema/2019-09-11#1935522 << afaik 'b' won't work (unlike on x86)
asciilifeform: dorion: interesting. i was for some reason certain that it didn't have a 32ism layer
asciilifeform: dorion: have you tried this ?
dorion: asciilifeform no Sir. hence my question: is it better to patch boost for arm64 support or build arm 32 bit bitcoind to run on rockchip ?
asciilifeform: dorion: at some pt we'll definitely want a properly 64bit trb on arm, so can use >4GB. thing is, currently there appears to be no arm64 box on the market that actually has >4GB.
dorion: it seems to me the latter is more straightforward, but that doesn't means it's better.
asciilifeform: so in short term imho useful to build, supposing it'll run, the 32bit one
asciilifeform: already in the 'pogo' experiment we had a working 32bit arm trb. so i expect it is doable with existing toolchain for 32-under-64.
asciilifeform: you'll still need a box with at least 2GB ram, to operate usefully.
snsabot: Logged on 2019-09-12 10:53:39 asciilifeform: http://logs.nosuchlabs.com/log/trilema/2019-09-12#1935742 << pretty great link. surprised to see archaeologist like this writing in engl.
mircea_popescu: was fully expecting russian to pop in at any moment. didn't see it ; but the topic is very central to rus' nationality research, a mirror of the novgorod-constsantiniople trade, this slavic over-the-danube thing. so not all hope is lost.
mircea_popescu: besides, the coin's russian.
asciilifeform: mircea_popescu: he has some entirely good translations of ru texts. i ran into 1 piece where he has a 17th c. snippet, properly translated.
dorion: asciilifeform: ok, thank you. I've looked into the pogo experiment and will adapt was needed. I have box with >2GB ram: the rockchip.
snsabot: Logged on 2019-09-12 10:54:38 asciilifeform: http://logs.nosuchlabs.com/log/trilema/2019-09-12#1935756 << it's a great idea imho. but prolly oughta be operated via commands in #t or similar , otherwise no amt of pipe or disk will ever suffice
mircea_popescu: http://logs.nosuchlabs.com/log/trilema/2019-09-12#1935775 << not that you asked me, but atm the better is to do both of these and report findings.
snsabot: Logged on 2019-09-12 12:20:41 dorion: asciilifeform no Sir. hence my question: is it better to patch boost for arm64 support or build arm 32 bit bitcoind to run on rockchip ?
asciilifeform: mircea_popescu: atm we still have 'kolhoz' bandwidth, i.e. 'all you can eat', but i expect if we put up an image paster it'll saturate the pipe 100% . ( naturally if mircea_popescu wants to make arrangements to get monster pipe specially for this experiment, i'ma work with him to make it go. but promises to be costly. )
mircea_popescu: well, more of a "we cross that when we actually get to it whainot."
mircea_popescu: iirc original item was bad, but not nearly trilema-bad
asciilifeform: hey BingoBoingo , were you able to get data re this q ?
snsabot: Logged on 2019-08-24 12:59:29 asciilifeform: BingoBoingo: i also would like to find what it'd cost to widen the pipe. AND how much currently is eaten, and by whom. i have dark suspicion currently that we're dangerously low on pipe.
mircea_popescu: (then again trilema bad is a thing in its own class - for instance anyone recal lthe time of the deal wiht copypaste, when trilema served a few imgs on each page of the 8chan ever loaded ? came to astonishing amt)
asciilifeform: mircea_popescu: i expect you'll want mechanical disks for this item. lemme know the types , will procure'em ( when the crate is flown, mircea_popescu will have 2 own boxen, the current smg , and its spare ; piz will have the 2 vacated boxen, which mircea_popescu can either purchase or lease , lemme know which prefer . ea. takes up to 4 disks . )
mircea_popescu: can you put fast raid in there ?
mircea_popescu: ie no parity
asciilifeform: i.e. striped ? sure
mircea_popescu: hardware ?
asciilifeform: all piz x86 boxen have iron raid , w/ 4 hoses.
asciilifeform: '3ware' type .
mircea_popescu: very nice. so get then 4 whatever plate disks you folk recommend on the basis of your extensive experience with computer hardware, seagate ?
mircea_popescu: and make them a strip raid. and this;ll be the trilema.org / dtng box then
asciilifeform: got it.
asciilifeform adds to purchase list .
mircea_popescu: and if bw comes to grief, we can either further tune the software via same process trilema.org was invented anyway ; or else spand the pipe, we see theb
mircea_popescu will confess all this is pretty fucking exciting.
asciilifeform: at worst we'll simply use apache's throttle, aha. oughta work.
mircea_popescu: reminds me of the time trinque stood me up the famous (har har) http://logs.nosuchlabs.com/log/trilema/2019-08-25#1930911 box, in excitingness.
snsabot: Logged on 2019-08-25 07:37:18 mircea_popescu: (yes, believe it or not, i ran a "all films reviewed on trilema" server for like... a year. to see if anything interesting happens. major pipe and everything, came to high four figures all told, mostly hardware, soem sweat to get it up.
mircea_popescu: asciilifeform, yeah, isn't light log a pleasure on tired eyes!
mircea_popescu also never thought he'll live to have a use for pure stripped raid arrays.
dorion: http://logs/nosuchlabs.com/log/trilema/2019-09-12#1935789 << yes Sir, thank you. I will proceed with the 32 bit build first, blog it, then circle back to see wh
snsabot: Logged on 2019-09-12 12:32:09 mircea_popescu: http://logs.nosuchlabs.com/log/trilema/2019-09-12#1935775 << not that you asked me, but atm the better is to do both of these and report findings.
mircea_popescu: but this is the ideal application : trielma's backed up anyways, and losing dtng chanmaterial... well.
dorion: *what has to be done with boost for aarch64 support.
asciilifeform: mircea_popescu: if you want 100% striped (vs. e.g. 2x mirrored 2x striped ) i still recommend to think about how will back up such box. with the current pipe, backing up remotely e.g. 4TB, will take ~week (and that's if entire population of piz is idling)
asciilifeform: a 100%-striped 4disk array will perish if even 1 disk dies.
mircea_popescu: i have no intention of backing it up.
mircea_popescu: if it croaks, it'll get reinstalled.
asciilifeform: aite, then.
mircea_popescu: attempting to put in production alf-formulated republican theoretical theosophy earlier, "spin up instance on spot" or however it went
mircea_popescu: i confess as a merchant i have a literal and approximative grasp on theological arcana.
asciilifeform: mircea_popescu: we'll snapshot the os a la dulap/smg so can be done reasonably quickly, when disk goes.
mircea_popescu: or rahter, if. incidentally, there's some way to read smart data ? i get oyu don't use the buttholes, but locally there still is a way yes ?
asciilifeform: mircea_popescu: on 3w : smartctl -a -d 3ware,x /dev/twa0 (where x is the hose #)
asciilifeform: i do this regularly on own boxen, to determine which disk next to throw out & prophylactically swap
asciilifeform: the 1 twist is, not all failures follow the expected rate (esp. on mech hdd)
BingoBoingo: http://logs.nosuchlabs.com/log/trilema/2019-09-12#1935794 << Is the dtng box going to want its own line coming in?
snsabot: Logged on 2019-09-12 12:33:45 asciilifeform: hey BingoBoingo , were you able to get data re this q ?
asciilifeform: BingoBoingo: not initially. see whole thrd.
BingoBoingo eating thread
mircea_popescu: asciilifeform, what's a line ?
asciilifeform: mircea_popescu: net pipe
mircea_popescu: ah, absolutely not
mircea_popescu: in fact i expect very little traffic tbg
mircea_popescu: will be trilema.org/dtng meaning it'll still block bw leeching, meaning -0 interest from all the usual makers of something from nothing.
BingoBoingo: I suspect the dtng resurrection is the way to do image pasting, seeing how everyone else constantly resurrects the broken less functional takes on the chan format
BingoBoingo: In other news, Art Deco Today: http://archive.is/qMTzn
feedbot: http://trilema.com/2019/59-particulars-laid-down-for-the-regulating-of-things-a-selection/ << Trilema -- 59 Particulars laid down for the Regulating of things. A selection.
asciilifeform: ACHTUNG, panzers! lobbes , diana_coman , et al : reader.py on dulap updated to lobbes's raw-export-fix. vtree page updated w/ lobbes patch and lobbes , asciilifeform , sigs.
asciilifeform: diana_coman i recommend to read & emplace this one, so lobbes & asciilifeform can resync from your /log-raw : the old export mechanism was apparently broken in the case of 'action' lines.
diana_coman: asciilifeform: ok, added to the ever-growing list.
asciilifeform: ty diana_coman
asciilifeform not yet had to resync at all, hence never stepped on the mine.
asciilifeform: ( and apparently when diana_coman synced, got lucky, there were no 'actions' )
diana_coman: asciilifeform: hm, kind of weird, in all #trilema logs no actions??
asciilifeform: diana_coman: bug was specifically in /log-raw , rather than db (which iirc you ate wholesale, and then manually synced coupla lines during the interval remained )
asciilifeform: diana_coman: compare w/fix vs. your current to see the diff.
lobbes needs to emplace own fix on production logs. Will do this now..
diana_coman: nm, works now.
lobbes: kk
lobbes: ty asciilifeform for testin' and signin'. (and ty diana_coman for helping me test earlier)
lobbes: I still need to resync all of the borked log lines on my own logger, but that'll have to wait until a bit later.
diana_coman: http://logs.ossasepia.com/log/trilema/2019-09-12#1935856 - fix deployed; and thank you lobbes for the fix!
ossabot: Logged on 2019-09-12 15:43:45 asciilifeform: diana_coman: compare w/fix vs. your current to see the diff.
asciilifeform: diana_coman: tested, a++ worx
asciilifeform: ACHTUNG, panzers : testing a small cosmetic fix to reader.py . if anyone notices oddities on whatever browser, plox to write in.
← 2019-09-11 | 2019-09-13 →