Hide Idle (>14 d.) Chans
← 2021-05-21 | 2021-05-23 → ↓
asciilifeform: $ticker btc usd
btcinfobot: Current BTC price in USD: $37882.59
asciilifeform: !w poll
watchglass: Polling 17 nodes...
watchglass: 24.28.108.235:8333 : Could not connect! (Operator: trinque)
watchglass: 185.85.38.54:8333 : Could not connect!
watchglass: 84.16.46.130:8333 : Could not connect!
watchglass: 185.163.46.29:8333 : Could not connect!
watchglass: 108.31.170.100:8333 : (pool-108-31-170-100.washdc.fios.verizon.net) Alive: (0.097s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Blocks=684552 (Operator: asciilifeform)
watchglass: 205.134.172.26:8333 : Alive: (0.088s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Return Addr=0.0.0.0:8333 Blocks=684552
watchglass: 54.39.156.171:8333 : (ns562940.ip-54-39-156.net) Alive: (0.108s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Blocks=684552
watchglass: 205.134.172.6:8333 : (172-6.core.ai.net) Alive: (0.102s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Return Addr=0.0.0.0:8333 Blocks=684552
watchglass: 205.134.172.4:8333 : (172-4.core.ai.net) Alive: (0.152s) V=70001 (/therealbitcoin.org:0.7.0.1/) Jumpers=0x1 (TRB-Compat.) Blocks=684552
watchglass: 143.202.160.10:8333 : Alive: (0.244s) V=70001 (/therealbitcoin.org:0.7.0.1/) Jumpers=0x1 (TRB-Compat.) Blocks=684552
watchglass: 205.134.172.28:8333 : Alive: (0.138s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Return Addr=0.0.0.0:8333 Blocks=684552 (Operator: whaack)
watchglass: 208.94.240.42:8333 : Alive: (0.176s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Blocks=684552
watchglass: 213.109.238.156:8333 : Alive: (0.274s) V=99999 (/therealbitcoin.org:0.9.99.99/) Jumpers=0x1 (TRB-Compat.) Blocks=684552
billymg: http://logs.nosuchlabs.com/log/asciilifeform/2021-05-18#1036533 << is this the line you're talking about? inside the loop in gen_chanlist?
snsabot: Logged on 2021-05-18 13:36:23 asciilifeform: billymg: the most egregious sharp edge i've not had time to sand off, is the channel menu, it could be generated with 1 sql query rather than N, in principle
billymg: if so, it's a tricky query. only way i could see to get it into one SQL query is with a 'union' statement between, essentially, num_chans queries
billymg: and performance is worse using union vs just doing individual queries
billymg: my results on my rockchip: http://paste.deedbot.org/?id=v6kO
whaack: asciilifeform: When planning what to do for my pushtx command for my blockexplorer, I dug up some old log thread about this whole high/low s ordeal. I did not fully grok everything, I believe. For one, was there actually a softfork regarding this - i.e. do miners actually reject blocks with a 'high s' signature? Or are the transactions
snsabot: (trilema) 2016-01-13 mod6: so currently, the main thing im working on/looking at is a possible high/low-S vpatch
whaack: just not relayed?
billymg: asciilifeform: actually, perhaps i spoke too soon. more complete results here: http://paste.deedbot.org/?id=mT4_
snsabot: Logged on 2021-05-22 12:37:22 billymg: my results on my rockchip: http://paste.deedbot.org/?id=v6kO
billymg: very minor savings (when db/proggy on the same box), but probably worth refactoring just for neatness and for users who may want to set up their db on a separate box
whaack: For the record, if it's not deemed stoopid, I want to create a 'private' (i.e. only accepts txns pushed through the irc interface) mempool with the most trb-compliant and sane relay policy. So, following the conclusions I gathered from reading the thread, it should accept high or low S txns, and relay them, whether or not miners today accept the high s transactions is another story.
whaack: Practically speaking, it may be worth issueing some warning to the user that currently miners 'dun like transactions congruent to 3 mod 17'
trinque: whaack: not trying to be a dick here, but what makes you think IRC is private?
whaack: I don't think irc is private, that's why I put private in quotes and added the parenthetical clarification, I meant to say that the explorer has a separate mempool that can only be accessed via irc commands
trinque: got ya
trinque: whaack: the thing would eat a paste of a transaction? seems like tx hex would be longer than IRC maxlen sometimes.
whaack: trinque: Yes, to deal with this problem I added a feature where arguments can be passed directly into an irc line or via a link to the paste of the arguments
whaack: This feature afaik is only useful for pushing raw txns, every other command has arguments that fit in an irc line, but I made the feature available for every command.
bitdashbot: (whaacked) 2021-05-22 whaack: !e view-raw-block 728
Apocalyptic: Regarding the libera.chat drama, Freenode ops abuse
shinohai: i adapted tcinfobot to sun on lisp ircd ok
verisimilitude: Hello, Apocalyptic.
Apocalyptic: Hi verisimilitude
verisimilitude: That link is amusing.
verisimilitude: ``In my view this risk is sufficiently high that the only reasonable course of action at this point is to assume out of caution that all Freenode PMs are tapped.''
verisimilitude: Only idiots weren't already doing this.
shinohai: verisimilitude: lmfao
shinohai: tyvm deedbot
Apocalyptic: verisimilitude, I agree, expecting PMs in cleartext travelling on a network you have no control over to be private is lulzy
shinohai: for rears of wartime service
shinohai: *years
verisimilitude: Since that link mentions Hacker News, I'm almost tempted to try pointing to my work there again, but it would likely be ignored. I should find more places to discuss my work, but so many are just spam holes nowadays.
verisimilitude: I'd upload my work to Reddit, but that would require me to use Reddit; nearly every time I glance over there, more posts have been silently removed from discussions.
verisimilitude: On that note, have a website, Apocalyptic?
Apocalyptic: verisimilitude, sadly I do not, nothing worth publishing, this may be relevant to your question
snsabot: Logged on 2019-12-26 23:35:24 asciilifeform: really, nuffin at all ?!
verisimilitude: Oh, alright.
verisimilitude: I feel the pain of never doing enough, by my thinking.
Apocalyptic: I find myself spending an increasing amount of time in meatspace, and when in front of computer I mainly do math/prog puzzles a la Project Euler
verisimilitude: Care for a similar suggestion?
Apocalyptic: depends what you're into really
verisimilitude: Consider writing CHIP-8 games.
Apocalyptic: I have nfi what CHIP-8 is, will look it up
Apocalyptic: I did my fair share of game writing when I was younger, can't say it's still a topic of interest these days
verisimilitude: I'll provide a good link.
verisimilitude: My website also has a few of my games, which I heavily document.
verisimilitude: Writing CHIP-8 games could be fun, like puzzles.
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2021-05-22#1037233 << what it oughta do , is to only look at lines which hit the db since the last max index.
snsabot: Logged on 2021-05-22 12:36:59 billymg: and performance is worse using union vs just doing individual queries
asciilifeform: rather than at entire db.
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2021-05-22#1037235 << 'soft fork' was the enemy term -- i.e. prb won't relay such blox, and prb-based cartel miners won't mine'em or mine on top of'em
snsabot: Logged on 2021-05-22 12:38:11 whaack: asciilifeform: When planning what to do for my pushtx command for my blockexplorer, I dug up some old log thread about this whole high/low s ordeal. I did not fully grok everything, I believe. For one, was there actually a softfork regarding this - i.e. do miners actually reject blocks with a 'high s' signature? Or are the transactions
asciilifeform: whaack: the things that happens when you actually broadcast a high-S tx has yet to be adequately 100% explained, afaik.
snsabot: (trilema) 2017-07-25 asciilifeform: interestingly asciilifeform recently learned that the 'mutate high to low S and broadcast malleated tx but ONLY if a 'doublespend attempt' ( you retransmitting, say, with patched trb ) is detected ' thing is STILL running
asciilifeform: (if not obv. from above - initially nothing happens. until you try to spend that coin again, at whatever point, even month later -- suddenly the old tx is dug up, immediately, (by who? why? nfi) malleated into low-S, and... mined into a block.
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2021-05-22#1037241 << outta curiosity, what's the practical use of a manual 'stuff tx here' gadget? i never understood the appeal
snsabot: Logged on 2021-05-22 12:42:47 whaack: For the record, if it's not deemed stoopid, I want to create a 'private' (i.e. only accepts txns pushed through the irc interface) mempool with the most trb-compliant and sane relay policy. So, following the conclusions I gathered from reading the thread, it should accept high or low S txns, and relay them, whether or not miners today accept the high s transactions is another story.
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2021-05-22#1037244 << a thing where can ~search~ mempool , display stats, etc. imho would be nifty
snsabot: Logged on 2021-05-22 13:04:33 whaack: I don't think irc is private, that's why I put private in quotes and added the parenthetical clarification, I meant to say that the explorer has a separate mempool that can only be accessed via irc commands
snsabot: Logged on 2021-05-22 15:47:57 Apocalyptic: Regarding the libera.chat drama, Freenode ops abuse
snsabot: Logged on 2021-05-19 17:24:59 asciilifeform: btw, for anyone who doesn't know, this is what the 'volunteers' of fleanode historically volunteered .
whaack: asciilifeform: Dafuq, that is quite strange re the delayed malleation, seems like some sort of random fuck-you move being done by someone with a better position in the graph of nodes, I don't see why it would be done by a miner.
asciilifeform: whaack: i suspect it is a side effect of some algo used by miners to fuck competitors, but nfi specifically how
asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2021-05-22#1037265 << ''is full of people who are fulla shit, and the vermin of the world inhabit it..'
snsabot: Logged on 2021-05-22 16:09:00 verisimilitude: I'd upload my work to Reddit, but that would require me to use Reddit; nearly every time I glance over there, more posts have been silently removed from discussions.
whaack: asciilifeform: There was a mention of a fork related to the high/low s problem, is there a block number associated with it and a publication of the blocks that were orphaned somewhere?
asciilifeform: whaack: there was never a fork in the 'hard' sense, i.e. with 2-sided war. earliest afaik mention of the problem in the logs.
snsabot: (trilema) 2015-10-15 BingoBoingo: The thing is does 0.5.X produce low-s transactions?
whaack: asciilifeform: Right, I was wondering if there is a documentation of 10+ blocks or so that were lost in the transition, or if those blocks even exist
asciilifeform: whaack: they may actually exist on your hdd. (that is, if you used e.g. asciilifeform's torrent to light up your noad; it descended from an old enuff noad to possibly contain this orphaned chainlet)
asciilifeform: trb's db in fact retains orphaned chains on disk, 4evah
whaack: asciilifeform: Nope, I did not use your torrent
whaack: but that is good to know and if i find the time maybe i'll pick them out and try to document them for historical purposes
asciilifeform has not made any efforts to catalogue these historical dead chains. they do exist on the disk, if you need'em for sumthing
asciilifeform: whaack: if you write a detector for'em, i'ma happily run it on my db and give you the output.
whaack: thanks, although in the case I do I'll probably download your torrent anyways
asciilifeform: it's still up fwiw.
asciilifeform: BingoBoingo: lemme know if you have any objections to this, but i'm thinking of auto-transforming qntra links to point to my mirror the way i do with btcbase linx.
asciilifeform: iirc the dns is not controlled by BingoBoingo and will at some pt expire or goatse.
asciilifeform: !q seen-anywhere BingoBoingo
snsabot: BingoBoingo last seen in #alethepedia on 2021-05-09 17:32:29: in wp-config.php
bitdashbot: BingoBoingo last seen in #alethepedia on 2021-05-09 17:32:29: in wp-config.php
asciilifeform: billymg: plz change the knob on yours to sumthing other than !q.
asciilifeform: !q uptime
snsabot: asciilifeform: time since my last reconnect : 0d 14h 41m
bitdashbot: asciilifeform: time since my last reconnect : 2d 0h 6m
billymg: asciilifeform: will do
whaack: asciilifeform: So I am a bit confuzzled as to what happened in this fork. My understanding is that the powerrrangers said "here's a new rule regarding signatures, all in favor raise your hands" then ~everyone raised their hands but only, say 30%, actually followed through with implementing the rule. So most miners continued as
whaack: normal, allowing high and low s transactions, and then some other miners mined a minority chain 'hey we have a new rule why r u guys not following the new rule!!!!' . Now eventually this split in mining resolved, but my question is did the miners who stuck with the permitive high/low s succcumb and jump onto the minority chain, or did the minority chain give up and jump onto the high/low s chain,
whaack: and then on that chain the new "low-s-only-rule" started to become enforced at a later block number?
asciilifeform: iirc this was the july 4 '15 oops-fork, aha
asciilifeform: afaik the minority-chain got out-POW'd, resolving the fork simply by 'mass'.
whaack: Looks like pete documented it at block 363,731 -> 363,736
billymg: http://logs.nosuchlabs.com/log/asciilifeform/2021-05-22#1037281 << by "what it oughta do", do you mean what postgres should be doing in its query plan, but isn't? or that query should be rewritten to coax postgres into using a more efficient plan?
snsabot: Logged on 2021-05-22 18:01:32 asciilifeform: http://logs.nosuchlabs.com/log/asciilifeform/2021-05-22#1037233 << what it oughta do , is to only look at lines which hit the db since the last max index.
asciilifeform: billymg: think about it: there's no reason to walk the entire DB, simply to find 'what was the most recent line marked $chan' etc
whaack: asciilifeform: I don't see how saying 'the minority chain got out-POW'd' helps distinguish which chain succeeded, by definition the minority-chain is the one out-POW'd, no?
asciilifeform: billymg: what it oughta do, is to only examine the lines that weren't in the db during the last query.
billymg: asciilifeform: yes, that makes sense
asciilifeform: whaack: the non-low-S-enforcing chain, to be concrete, was out-POW'd.
asciilifeform: whaack: after this, via means unknown to asciilifeform , the miner cartel was established, and at a certain pt (when?) high-S became unmineable (or rather, will be mined after malleation by known party , as i described. can do the experiment yerself, afaik it still happens today)
billymg: asciilifeform: here is the query plan for one chan, and the time it takes to execute: http://paste.deedbot.org/?id=5ZXr
billymg: asciilifeform: i just don't have a frame of reference to know if 37ms is a bad time for that query or not
asciilifeform: billymg: well, it's ~37ms * how many chans, probably
asciilifeform: before long, 'adds up to real money'
billymg: a dozen give/take
asciilifeform: also will depend on vagaries of disk cache etc
billymg: sure
whaack: asciilifeform: alright, thanks, I may have some more questions on this but for now i'm going to go setup a slackline on the beach
asciilifeform: i suspect the bulk of time, billymg , is spent on long-dead chans (assuming linear walk)
asciilifeform: whaack: no prob.
billymg: asciilifeform: i'm just trying to get to root of the sharp edge you mentioned earlier
snsabot: Logged on 2021-05-22 12:33:11 snsabot: Logged on 2021-05-18 13:36:23 asciilifeform: billymg: the most egregious sharp edge i've not had time to sand off, is the channel menu, it could be generated with 1 sql query rather than N, in principle
asciilifeform: billymg: it's a simple fix, i simply have not gotten to it. (if 0 new lines in db since last page load -- should not even run the query.)
billymg: i can rewrite the N queries into a single query with N unions, and takes about the same amount of time, maybe coupla ms less. but if db is on a separate box this would be a decent gain considering the network latency cost of making N separate requests
asciilifeform: if 1 new line -- that's the only line that should impact the channel list generation.
asciilifeform: billymg: the channel list generator should only look at the lines entered into db since it last executed.
asciilifeform: does this make sense ?
billymg: asciilifeform: ahhh, so you were talking more about saving state somewhere in the logger, to prevent unnecessary queries, rather than simply doing a more efficient query?
billymg: if that's the case then i misunderstood earlier
asciilifeform: billymg: both, but forgot to mention the essential detail, lol
asciilifeform has not looked under the hood of the logger in quite some time, before the recent threads with billymg
verisimilitude: I'm having a hard time imagining something in sufficient detail, asciilifeform, loper-os. How would it be a high-level architecture, non-von Neumann, and yet also have characteristics such as orthogonal persistance; backing up some state to disk is easy to think, but requires being able to treat memory as inert.
asciilifeform: verisimilitude: how much do you know re how cpu cache worx ?
verisimilitude: I know a good bit.
asciilifeform: verisimilitude: observe that it works without you ever explicitly concerning yourself with it.
verisimilitude: I've already realized that using transactions enables grouping together the important memory to persist, but this is still von Neumann.
asciilifeform: verisimilitude: these are 100% separate problems.
verisimilitude: Gee, I've read a good bit on when the caching doesn't work.
asciilifeform: verisimilitude: i dun think i've ever owned a box where 'doesn't work' in the sense of returning random values
asciilifeform: verisimilitude: as for 'how persistence on non-vm machine?' -- the simplest method is that whatever computing elements used, have a non-volatile storage inside, large enough to hold 100% of whatever volatile state they have; and access to sufficient power (stored in supercap) to guarantee that the save takes place on interruption of power.
verisimilitude: Transactions enable disregarding minor state, such as intermediate results, when persisting to disk, so it was fun to consider how almost all of the RAM can be safely ignored for this.
verisimilitude: Alright, so very specialized hardware was the answer.
asciilifeform: verisimilitude: the current vn state of the art is imho analogous to a bus with separate steering wheel and pedals for each wheel.
asciilifeform: yes it is possible to drive such a bus.
asciilifeform: but it would look quite ridiculous compared with a proper bus.
verisimilitude: I've a better analogy.
verisimilitude: I want to draw a little comic depicting how languages travel; the C language ignores traffic lights, as an example; APL is the only one not driving, but instead riding a train.
verisimilitude: Well, not an analogy, but a metaphor.
asciilifeform must bbl
Apocalyptic: asciilifeform, any ETA for phuctor's return ?
gregorynyssa: verisimilitude: is that your own website? you are also Russian American like asciilifeform?
gregorynyssa: http://logs.nosuchlabs.com/log/asciilifeform/2021-05-22#1037381 << I am quite interested in this as well. groundbreaking and also cinematic if I might say.
snsabot: Logged on 2021-05-22 20:07:13 Apocalyptic: asciilifeform, any ETA for phuctor's return ?
gregorynyssa: "we will find out who diddled the keys and why"
gregorynyssa: http://logs.nosuchlabs.com/log/asciilifeform/2021-05-22#1037378 << APl is beautiful but it is arguably not a general-purpose language
snsabot: Logged on 2021-05-22 19:00:58 verisimilitude: I want to draw a little comic depicting how languages travel; the C language ignores traffic lights, as an example; APL is the only one not driving, but instead riding a train.
gregorynyssa: I say arguably because if you have studied it deeply enough, you can make it behave more like a general-purpose language, but few people ever reach that stage
gregorynyssa: if anyone has any recommended books or papers concerning APL, please let me know. I still intend to learn more of the language down the road.
gregorynyssa: I am not as good as I would like to be.
verisimilitude: Is which website mine?
verisimilitude: APL may not be general-purpose, but then few languages actually are.
snsabot: Logged on 2021-05-22 20:39:23 verisimilitude: Is which website mine?
verisimilitude: APL in particular isn't good for linked structures.
snsabot: Logged on 2021-05-22 20:40:50 verisimilitude: http://verisimilitudes.net
verisimilitude: I find general-purpose usually means equally bad at everything.
gregorynyssa: the term general-purpose in my experience usually means two things: (1) strong support for I/O; (2) strong support for string-processing.\
gregorynyssa: Tcl took #2 to the extreme; not without reason IMO.
verisimilitude: I find that absurd.
gregorynyssa: Tk is pretty bad, I will admit.
gregorynyssa: the cleverness of Tcl did not extend to Tk.
verisimilitude: I just don't see I/O as related in any way to this.
verisimilitude: It could be argued most I/O today is just due to poor abstractions.
trinque: sounds instead like the requirements for webshitlangs, meaning no offense to gregorynyssa
trinque is up to eyeballs in such languages himself most days
trinque: gregorynyssa: what do you find clever in tcl?
verisimilitude: I'd claim I'd rather kill myself than use such languages, but the truth is I'd rather kill someone else.
trinque: verisimilitude: you fixate too much on your feelings; no one cares
verisimilitude: It's a joke.
verisimilitude: I suppose I should joke less here.
trinque: watch, I'll do the low-effort thing too. holy shit gregorynyssa, uplevel and upvar mean I can't understand *any* procedure without reading the entire call-stack.
trinque: now I suppose I should proceed to a non sequitur about some other language I Like (TM)
verisimilitude: I don't consider my jokes low-effort, although suppose some are.
verisimilitude: Anyway, I appreciate the criticism.
trinque: back on subj, "general purpose lang" is indeed often a fig leaf for "had no particular design principles, thus accreted popular hacks"
← 2021-05-21 | 2021-05-23 → ↑
Random(asciilifeform) | Download hourly DB snapshot | Get Source Code