Hide Idle (>14 d.) Chans
← 2019-02-17 | 2019-02-19 → ↓
ave1: asciilifeform, the september version is not only static but also some download links fixing, http://btcbase.org/log/2019-02-18#1897851
a111: Logged on 2019-02-18 03:18 asciilifeform set up a rebuild of ave1 gnat (june) on dulap, with http://bvt-trace.net/2019/02/gnat-zero-cost-exceptions-and-asynchronous-task-aborting-part-2/comment-page-1/#selection-163.2-167.52 . tomorrow will see what came out of this.
ave1: also, I was working on the september version, but I'm currenlty stuck on the aarch64 port, something with the jmps, files being undefined etc.
a111: Logged on 2019-02-17 14:27 feedbot: http://thetarpit.org/posts/y05/084-gutenberg-ii.html << The Tar Pit -- gutenberg.org part zwei, a dissection
diana_coman: http://btcbase.org/log/2019-02-18#1897858 -> ah, so broken links were only in the older versions? anyway: ave1 please do me a favour and point any fetch/download scripts to a mirror of the stuff on your own website, there is no way around this. For one thing I'd much rather download from a republican site and for the other as experience has already amply shown, any external site WILL move them, change them, drop them and it will at best br
a111: Logged on 2019-02-18 06:13 ave1: asciilifeform, the september version is not only static but also some download links fixing, http://btcbase.org/log/2019-02-18#1897851
diana_coman: eak your scripts, at worst make it download something else.
asciilifeform: ave1: i changed mine to work from local copy of the tars, back when first tested. is there anyffin else new in the sept. ver.?
mircea_popescu: this is a good idea. put the hashes somewhere and there you go.
asciilifeform: i'ma put whole thing in a signed manifest
asciilifeform: sadly had to restart the experimental build ~2h ago, botched the config last night
asciilifeform: strangely , the -j32 thing aint working for good chunk of the build, possibly dun apply to the bootstrap gnat..?
asciilifeform: really this whole thing oughta take ~30m on dulap ( 32 cpu )
mircea_popescu: so why doesn't it ?
asciilifeform: looking at the kilometre of bash that comes with the thing, for the answr
asciilifeform: ( the fact that diana_coman's built in ~1h, suggests that i broke my local config there )
mircea_popescu: possibru
asciilifeform: mircea_popescu: this run is a test of bvt's 'kill weaksymbol-ism' pill
mircea_popescu: it's such a wonder to me that thing links.
asciilifeform: realized when i woke up that really oughta have first tried plain-old sjljistic ave1-gnat , and see if this cure is even necessary there
asciilifeform: ( will do after this one )
asciilifeform: mircea_popescu: it aint built yet, so cannot yet say 'it links'
asciilifeform wonders if he's the only 1 who is disgusted at released opensores 'flagship' soft that builds with 9000km of warningolade
mircea_popescu: the warning list prolly changes by machine.
asciilifeform: they aint machinistic warnings. mostly 'uninited variable', 'missing sentinel', etc
asciilifeform: aand it built. nao to see ~what~ built..
feedbot: http://trilema.com/2019/the-secret-of-santa-vittoria/ << Trilema -- The Secret of Santa Vittoria
asciilifeform: result : builds with bvt's patch with 0 eggogs, his tarball test functions 100x w/out segfault
asciilifeform: but this is with zcx. then went to build with sjlj, found, grr, that it :
asciilifeform: gnat1: RTS path not valid: missing adainclude and adalib directories
asciilifeform: despite both paths being set correctly
asciilifeform: looking in the log, iirc diana_coman had same issue, & cured
mircea_popescu: wunderbar
asciilifeform: notyetwunderbar, still gotta 1) sjlj 2) try test w/ naked one and see if pill even necessary
asciilifeform: not to mention 'test may reveal presence of bug, but never absence'(tm)(r)
asciilifeform: would also like to see wtf i did to the -j32 thing
asciilifeform: ( currently it's a 4hr build, lol )
asciilifeform: and hrm, loox like we still dun have a genesis for this thing
asciilifeform: ave1 are you gonna bake one ? someone's gotta, and i will with own hands if no one has it
mircea_popescu: he should, really.
asciilifeform: imho yes
asciilifeform currently wonders why the fuck an aborted thread needs to unwind its stack frame by frame, it has no further biz in this world after it stops, so wtf even for, new thread gets entirely new stack
asciilifeform: i can see that it might have to let go of any locks it holds, but what if holds none ? this is easily shown at compile time
asciilifeform: meanwhile, in world of ancient warez finds : http://nosuchlabs.com/pub/warez/artek.zip << ada-83 compiler for msdos, circa 1987.
asciilifeform: fits on floppy...
bvt: hello. quick report: i bootstrapped avegnat using asciilifeform's tar (thanks!), had to change the the paths in linker scripts and la-files, will make a post on that.
bvt: ran the tests for exceptions race, libgcc is fine in gcc4.9, locks are in place, so it seems that it is indeed another gcc5ism
asciilifeform: bvt: built with sjlj mode ?
bvt: nope, zcx. sjlj can the next step
asciilifeform: fwiw mine also worked ( i.e. didn't bomb in bvt't test ) on zcx. but not tried sjlj yet.
asciilifeform: ( as i understand, battlefield proggies will use sjlj moar or less exclusively )
mircea_popescu: bvt there's reasons we picked gcc 4.8/9
mircea_popescu: incidentally re http://pizarroisp.net/2019/02/17/pizarro-isp-february-17th-update/#selection-45.0-45.26 -- you can't practically wait with this whole thing until either aarm64 is fixed or the aliens invade. what are you gonna do for capacity ?
mircea_popescu: and i somehow still don't see the only item of any import in there, which kinda makes the whole list derisory.
asciilifeform: BingoBoingo mircea_popescu has point : are you gonna produce a list of spammees, or just what is it you have hands full with, and asciilifeform has to dredge the net with own hands and halt all other work ?
BingoBoingo: asciilifeform: I'll have a list of targets tonight
asciilifeform: BingoBoingo: ty, would really like to see motion on this front asap
asciilifeform: and would rather not have to do whole thing from scratch
asciilifeform: re capacity : also needs discussion. currently considering rk array (1u), two x64 boxen of dulap type, and additionally a set of apu1d4 (see logs, or https://archive.is/8JBbH ) .
asciilifeform: re rk -- not errybody is running ada proggy, and errything else appears to work quite well there.
asciilifeform: in part the exact mix will depend on what BingoBoingo yields in terms of userbase prelim. hunt
asciilifeform: and on what existing userbase says ( asciilifeform still ! waiting ! for any passengers ! to stand up, re http://btcbase.org/log/2019-01-15#1887214 . if none by end of this wk, i will assume there are none ! )
a111: Logged on 2019-01-15 15:32 asciilifeform: it is also time to speak of the next crate. and the customary four cargo slots. 1 is to contain a replacement for uy1 ( and ideally will run cuntoo . ) 1 will contain a 1u that holds rk's. this gives 2 1u slots remaining, they can be occupied by colo passengers, if these stand up and wish to ride , or pizarro irons, at BingoBoingo's option.
asciilifeform: BingoBoingo: if no passengers stand up, crate will contain 100% piz buildout materiel . and would like your input on precisely what, also asap.
asciilifeform: BingoBoingo: it is also my understanding that the new rk is presently idle. if this is still so, i'ma commandeer it for propaganda ops. lemme know.
asciilifeform: BingoBoingo: plox to make use of mircea_popescu's suggested, in log, leads for the hunt, but do not limit to it
asciilifeform bbl,meat
mircea_popescu: asciilifeform your problem's open tho, what the heck are you gonna do, put up more arms we can't compile for ? put up something else ?
mircea_popescu: if bitcoin were moar entrepreneurial we'd see stock price movement in pizarro whenever such events happened lol.
BingoBoingo: asciilifeform: new rk is indeed idle
BingoBoingo: http://p.bvulpes.com/pastes/Opg8n/?raw=true << asciilifeform Here's what I've sorted so far
feedbot: http://trilema.com/2019/wild-is-the-wind/ << Trilema -- Wild is the Wind
BingoBoingo: asciilifeform: What sort apu1d4 offers what sort of density to the Rack Unit not that we are revisiting x64
BingoBoingo: 6 or 8 to the 1U chasis?
feedbot: http://bvt-trace.net/2019/02/gnat-zero-cost-exceptions-and-asynchronous-task-aborting-part-3/ << bvt's backtrace -- GNAT Zero Cost Exceptions and Asynchronous Task Aborting. Part 3.
BingoBoingo: What a coon
asciilifeform: BingoBoingo: ty, will take a look shortly, 1st would like to macroexpand mircea_popescu's point
mircea_popescu: i was looking for that pic a whole ago, you know. "where's the one with the bunch of them huddled together like so manby peasantsa"
mircea_popescu: possibly at the time of trump's washington dinner "humiliation". by whom, by this crop of the barn-born ?
asciilifeform: what we have this wk, is a 'sjlj-gnat dun go on arm64'. i dun expect this is a permanent disease, but i do expect it'll take time to fix, esp. if asciilifeform has to do with own 2hands. until then, rk cluster is stuck with the obsolete compiler ( which runs e.g. 'lamp stack' and even zcx-gnat , but obv. won't run the emerging standard sjljistic one until fix. )
asciilifeform: so asciilifeform leans to installing a qty of x64 boxen of roughly same cost profile as rk ( but with added bonus that they eat sata drives ) : apu1. these also have bonus of published electrical schematic and custom asciilifeform bios from src.
mircea_popescu: how many cores in the rk cpu ?
asciilifeform: and in apu -- 2. but apu still cleans its clock for most processes, oddly enuff (tho not by large margin)
asciilifeform: ( this is somewhat of a black art, it is probably possible to come up with a benchmark where the rk wins )
mircea_popescu: well, doesn't seem like the end of the world, then, to say "this is the rk, has so and so list of advantages as before discussed (truly independend box, etc) and the one drawback meanwhile discovered that you can't really do tmsr-like threading in ada on it, which may be fixed later"
asciilifeform: correct.
mircea_popescu: doesn't seem like the end of the world to "can't thread properly" on a 4 core box.
mircea_popescu: so nothing really lost here.
asciilifeform: the apu loses on physical space ( it takes the sq. metrage of 6 rk , just by itself ) and cost ( +30% over rk ) aand naturally x86itude.
mircea_popescu is satisfied.
asciilifeform: mircea_popescu: i'd still like to see the arm64 wtf resolved
asciilifeform: cuz srsly
mircea_popescu: but it's not actually seated on anyone's neck.
asciilifeform: presently correct.
asciilifeform: i.e. i won't be holding up the expedition for its fix.
asciilifeform: mircea_popescu: apu wins substantially on 1 score -- it can eat ~3~ msata ssd's
asciilifeform: and equipped with 3 GB nics.
asciilifeform: ( on rk, 0 and 1 respectively )
mircea_popescu: yeah, but... if you'rte gonna beef-server, might as well get a 32 core monster.
asciilifeform: i also have an old and well-oiled toolchain for it, where can e.g. throw customer payload directly into rom, if desired
mircea_popescu: basically there's not much room for middle here. either rk or else monster -- even things such as shared hosting are better off on large x86 server than a bunch of mid ones
asciilifeform: mircea_popescu: apu is something rather like a x64 rk.
mircea_popescu: kinda my point.
asciilifeform: ( btw there was to be a demo unit of same in the 1st crate, but did not fit masswise )
BingoBoingo: APU can also be kitted to Gutenburg mirror with appropriate drives and ECC RAM
asciilifeform: BingoBoingo: iirc no ecc on either
asciilifeform: ( i'd pay in gold for small box with ecc, dun exist afaik )
asciilifeform: it remains a serious wtf to me : think, bolix had ecc.
asciilifeform: in 1984.
BingoBoingo: Ah, apu2 is ECC ram, but newer stinkier AMD "Jaguar"
asciilifeform: we won't be using apu2 at piz
asciilifeform: it is infested
BingoBoingo: bobcat's plenty fast
asciilifeform: ( not unless somebody pays to colo one, at any rate )
asciilifeform: see, asciilifeform has an actual blobless bios for apu1.
BingoBoingo: Ain't nobody doing blobless for "Jaguar"
asciilifeform: BingoBoingo: http://btcbase.org/log/2019-02-18#1897935 << to allow for power and sufficient thermal , 4 per 1u.
a111: Logged on 2019-02-18 23:15 BingoBoingo: 6 or 8 to the 1U chasis?
BingoBoingo: But Bobcat like the apu1 has is sitting on my desk
asciilifeform: ( it may be possible to pack'em tighter, but currently imho reliability trumps max utilization of cubic metrage )
BingoBoingo: Has only gotten faster since I first put it on the desk in 2010 or 2011
feedbot: http://trilema.com/2019/the-rose-tattoo/ << Trilema -- The Rose Tattoo
asciilifeform: BingoBoingo: apu1 will be exactly same as when i 1st laid hands on it in '14
asciilifeform: afaik they haven't changed the chipset.
asciilifeform: that's half the appeal, all of my recipe for it still worx.
asciilifeform: trb-capable box, btw, i have one going .
asciilifeform indeed expected that it's 100% gcc5ism
asciilifeform: in entirely other noose, the xray worx...
mircea_popescu: cool deal.\]
asciilifeform: 1st photo subj is those dead samsung ssd stix BingoBoingo sent in.
asciilifeform will post a scan once he gets a decent exposure, this may not be today, aint a priority item
asciilifeform: will also be interesting at some pt to roast a ~working~ ssd; in theory soft xray eats ssd
asciilifeform: nobody seems to publish the test data on subj tho, aside from 'golden toilet' ic houses
asciilifeform: re ^ -- the steel shell is a bitch, will need tuning for kv. but for demo purposes, the FG analogue board photo ~did~ come out pretty well, will scan it when it dries.
trinque: http://btcbase.org/log/2019-02-14#1896360 << far better than I could've said it, but yes. your "from cause" is foundational for sanity, let alone republic. I gather some have taken "sanity" as a term of art there, not the literal meaning.
a111: Logged on 2019-02-14 07:45 mircea_popescu: trinque let me know how close i got.
trinque: what happens when some uninspected and unconscious part of mod6 wells up from the undermeat and "just wants to" be happy/please wife/stop kid screaming
feedbot: http://www.loper-os.org/?p=3029 << Loper OS -- PCB Radiography Kindergarten.
trinque: http://btcbase.org/log/2019-02-14#1896530 << certainly! does it match your last run?
a111: Logged on 2019-02-14 19:06 diana_coman: http://btcbase.org/log/2019-02-14#1896356 -> this finished: the signature still does not verify; trinque let me know if you want to see the result of this run too
trinque: http://btcbase.org/log/2019-02-17#1897392 << cool, snagged and will dissect. did the bootstrapped system boot for ya?
a111: Logged on 2019-02-17 14:14 bvt: trinque: genesis (does not verify) from successful cuntoo deployment: http://bvt-trace.net/src/genesis-14.02.2019.vpatch
BingoBoingo: In other news, taking my walk tonight I got the celebrity treatment for ~3 minutes. A group of local youth in the city for back to school asked me for directions and were thrilled to discover in Gringo in the wild who could indeed offer directions.
BingoBoingo: I did not join the group as it was composed of 4 dorks and 5 morbidly obese fem-presenting mayos.
trinque: http://btcbase.org/log/2019-02-17#1897638 << glibc has been banned from my own machines for some time. haven't missed it on desktop workstation, server, embedded.
a111: Logged on 2019-02-17 15:50 mircea_popescu: some time to look at things and consider matters will be needed ; but i specifically want to hear something from asciilifeform ave1 bvt diana_coman phf spyked trinque
trinque: http://btcbase.org/log/2019-02-17#1897703 << nope, emacs patches exist, I've run
a111: Logged on 2019-02-17 16:09 asciilifeform: for instance, emacs has yet to be cured, as i understand
asciilifeform: trinque: oh hey
asciilifeform: are these posted ?
BingoBoingo: It has been a while. Turns out he got a job of sorts at a Confiteria. Sleeps in an actual bed most nights now.
trinque: asciilifeform: iirc last thing I built was in portage proper, built atop musl. this was probably a late and fungal version number
trinque: https://gitweb.gentoo.org/proj/musl.git/log/app-editors/emacs?qt=grep&q=emacs << looks like youngest they had was 24.5, have sinced removed patches from the overlay claiming mainline works
mircea_popescu: trinque wb!
trinque: howdy mircea_popescu
ave1: asciilifeform, re http://btcbase.org/log/2019-02-18#1897899, a genesis of the build scripts or gcc?, with the build scripts we'll get diffs of diffs (which should work now)
a111: Logged on 2019-02-18 17:43 asciilifeform: ave1 are you gonna bake one ? someone's gotta, and i will with own hands if no one has it
mircea_popescu: asciilifeform nice! pic is quite crisp huh.
BingoBoingo: Downright pornographic. All it's missing is the shaved snatch!
← 2019-02-17 | 2019-02-19 → ↑
Random(trilema) | Download hourly DB snapshot | Get Source Code