billymg: there are patches available for m4 to get around the >=glibc 2.28 problem, one was even included in my stock system's m4 ebuild
billymg: asciilifeform: yeah, i could build on my old laptop, but i'm trying to be able to build it here as well (since i plan to format and rebuild what's on the laptop after i'm done with this)
billymg: asciilifeform: yes, i think the rotor method makes sense for me
billymg: i read whaack's and jfw's posts on the non-rotor build method. i haven't tried it yet but given all the other "upgrades" my 2021 gentoo probably has i think it might be best to stick with the rotor method. i just need a way to get around the m4 bootstrapping
billymg: now i'm trying to get trb working and am running into this constraint
billymg: nodes_host_key DO UPDATE SET'
billymg: caught a bug in my crawler's genesis though, where two of the sql queries use a different index name than the one that gets defined when initializing from bitdash_schema.sql. i'll post a regrind of the genesis soon but if anyone runs into it the fix is to change the two instances of 'ON CONFLICT ON CONSTRAINT unique_host DO UPDATE SET' to 'ON CONFLICT ON CONSTRAINT
billymg: http://logs.nosuchlabs.com/log/asciilifeform/2021-09-07#1056800 << this method works, was able to get the crawler running. i installed all the python libs i needed by specifying exact versions, e.g. `pip install -Iv psycopg2==2.8.6`, and at least with my small list of required deps (flask, psycopg2, and requests) all were available
billymg: bingoboingo: does jingling keys in front of the camera like parents do for babies also work? if no one's tried it yet you could start the trend
billymg: asciilifeform: i plan to write up a complete guide for this build, including the source files and tarballs where not the default, once this is all done (so far have working mp-wp, just need the crawler and logotron now)
billymg: for anyone curious, running `python2 get-pip.py "pip < 21.0"` with this get-pip.py worked. now to see if it'll install the py2 packages i'm looking for...
billymg: when war crime tribunals
billymg: but i assumed it could still be used to install python2 packages, for legacy support and whatnot
billymg: no, it is
billymg: despite it being an ebuild for python2.7
billymg: and it worked in that it installed a pip, but it installed a python3 pip
billymg: asciilifeform: not yet
billymg: if there's a way to manually build from source i haven't found it yet, closest thing i found was build python from source with '-ensurepip' flag
billymg: this pip thing is harder than i thought it would be too
billymg: oops
billymg: \quit
billymg: heh
billymg: asciilifeform: what's the downside of keeping fossilzed python2 around? like gnat 4.9.4 for example?
billymg: rather than try and keep a working python2 alive
billymg: ah, as in, these must be ported to ada or something?
billymg: asciilifeform: what do you mean by depythonization?
billymg: so now i'm just going to build pip 20 from source rather than trying to go through portage
billymg: so i'm thinking python-exec and eselect-python might have been a red herring and actual breakage is in portage
billymg: required*
billymg: and... if i added -python3_9 to USE emerge would throw an error saying python3_9 is requred
billymg: asciilifeform: despite the ebuild file clearly listing 2_7 as a compatible version
billymg: asciilifeform: interesting note there, even with that pip ebuild in hand, and PYTHON_TARGETS and PYTHON_SINGLE_TARGET set to python2_7, emerge wouldn't include the flag
billymg: asciilifeform: i've given up at the moment. the effort was mostly so i could emerge pip-20.2.3, the last python2 compat build in portage
billymg: asciilifeform: yeah, i was hoping i could make some progress there, but still stuck
billymg: dev-lang/python-exec, downloaded and built from ebuilds prior to obvious python2 sabotage, and still no luck
billymg: asciilifeform: unrelated, i spent several hours fighting with my build of "modern" gentoo trying to get python2 to be the default. i went through the change logs for app-eselect/eselect-python and
billymg: asciilifeform: but why sell for e-usd when you could spend directly for literally everything you would otherwise buy with the e-usd? which seems to be the case now in el salvador, i'm assuming since it's legal tender you can pay your rent in it too
billymg: asciilifeform: right but lets you choose where the tax goes. not for billymg of course, for me i would just have to pay double tax (capital gains to us and import tax to orc gov). but i imagine there are some braver than i am who can be more flexible
billymg: asciilifeform: i'm not quite as cynical as you on this, eventually there will be (if there isn't already) one of those amazon mail forwarding importers, and they'll be taking btc too
billymg: lol at dump on the day el salvador enacts 'legal tender' law
billymg: $ticker btc usd
billymg: http://logs.nosuchlabs.com/log/asciilifeform/2021-09-05#1056499 << updated. thanks for the feedback
billymg: cgra: ttyl
billymg: cgra: ^ essentially that list, though asciilifeform's watchglass has a configurable knob for 'peershots', my crawler has that set to 5, not sure what alf's watchglass is set to
billymg: !w peers 205.134.172.26
billymg: cgra: i'm personally not familiar with the trb code at all, so quite possible it was a mistake in labeling on my part
billymg: or at least recently connected (though this would also mean it's a misnomer to refer to them as 'connected peers' as that implies currently connected)
billymg: cgra: hrm, technically you might be right, i had always assumed before that meant the node was also connected to those peers
billymg: signpost: overall really cool, and i'm happy now that i finally have a vtron on this box
billymg: not because i didn't want to but just because i forgot it was there until reading your notes on it again now
billymg: signpost: btw, did not use this, but rather mounted my primary drive inside the chroot, navigated to the files, and built with the standard e.g. `make vdiff`
billymg: however, when not chroot'd in, but only with the path to pentacle's gnat added to my PATH (so that e.g. `which gnatmake` works in the host env), i received some errors when trying to build the vtools binaries
billymg: so now i have a working vtron on the host box
billymg: signpost: i tried out your pentacle thing, with it i was able to chroot in and build the vtools binaries. this time all of them built as static binaries, without me even needing to modify the orginal gpr files to add static flags
billymg: signpost: grabbed your files, going to try it out in a bit
billymg: asciilifeform: gotcha, ok
billymg: asciilifeform: unavoidable? or you think it's a bug in my gpr file?
billymg: asciilifeform: ldd ksum, ldd vpatch -> "not a dynamic executable", ldd vdiff -> http://paste.deedbot.org/?id=hXdc
billymg: grabbing it now
billymg: i haven't given up yet on this box, but may be my last attempt at trying to contort upstream into something usable
billymg: oh neat
billymg: signpost: what is pentacle again?
billymg: they built, but still don't work on my non-gnat box
billymg: the sections i added to add the -static flags
billymg: btw, asciilifeform: do these gpr files look right for vtools: http://paste.deedbot.org/?id=PYvK
billymg: signpost: sure, still stuck
billymg: my struggle is getting gnat on this box
billymg: mats: if you have a box that you can just put asciilifeform's dulap on the trb build is easy
billymg: so gotta be gnat-gcc?
billymg: ah
billymg: asciilifeform: i'm using that one though, i'm building through portage, so it's whatever they pull when the bootstrap flag is set, which from my distfiles appears to be gnat-gpl-2014
billymg: i could try: "In this case, globally adding -fPIC to C[XX]FLAGS resolves the issue, although this practice is discouraged because the executable end up being PIC-enabled as well." -- and if it works, try to find/patch the individual files that need the flag by writing a patch for the ebuild
billymg: specifically Case 3
billymg: asciilifeform: re the gnat build errors, i'm looking at this: https://wiki.gentoo.org/wiki/Project:AMD64/Fixing_-fPIC_Errors_Guide
billymg: alright, will give that a shot, thanks
billymg: asciilifeform: these use("-static") lines?
billymg: asciilifeform: are there some flags in the makefile i could tweak to force it to build with static linking?
billymg: i.e. those binaries should work on a machine that doesn't have gnat?
billymg: so it should be working??
billymg: it is
billymg: and i do have the standalone ksum that comes with vtools in path too (otherwise v.pl fails the commands check in the beginning)
billymg: asciilifeform: e.g.: http://paste.deedbot.org/?id=VV-2
billymg: i can press genesis patches, but any patch beyond the genesis fails with hash mismatch
billymg: asciilifeform: one sec
billymg: yeah, vpatch even kinda works
billymg: does that seem right?
billymg: it's just when i copy the compiled binaries over to this box via usb stick and attempt to use they don't work correctly
billymg: my vtools work fine on my other box
billymg: which leads me to believe something in gnat is also a runtime dependency for those binaries
billymg: asciilifeform: also, through experimentation, i was not able to build the vtools binaries on my gnat box and then run them successfully on my current box
billymg: asciilifeform: if i have a working gnat on my other machine, how would i manually move that over to this one in order to bootstrap? i don't know the full list of binaries that would have to be moved over
billymg: asciilifeform: here are the build logs
billymg: one sec, lemme get the build logs
billymg: asciilifeform: gnat-gpl-2016 and with gnat-gpl-2014 as the bootstrap
billymg: was there ever a definitive resolution of the 'recompile with -fPIC' error? i'm hitting it now when trying to build gnat
billymg: shinohai: thanks, will read through this and the linked items and see if i can get it going
billymg: ah, gotcha
billymg: asciilifeform: i was just looking at these steps: http://thebitcoin.foundation/trb-howto.html
billymg: ah, yes, it does
billymg: i plan to build trb on here as well though, does that also need a gnat?
billymg: maybe i can just copy over the binaries for now
billymg: asciilifeform: yeah it works, at the moment i only need it to build vtools
billymg: my laptop seems to have a gnat-2016 installed but i have no idea how it go there (from a few years ago when i was just entering random commands until something worked)
billymg: signpost, asciilifeform: i believe my system is glibc currently rather than musl
billymg: asciilifeform: so better to leave gcc-4.9.4 in the slot, and build a copy of gnat-2016 manually?
billymg: is there a way to have gnat-2016 (4.9.4) and standard gcc-4.9.4 in a gentoo system at the same time? i just tried to emerge gnat-2016 and got blocked by gcc-4.9.4
billymg: signpost: ah true, probably has a lot to do with the flavor too
billymg: asciilifeform: the taste of the backyard chickens is so much better too, like different type of meat altogether
billymg: signpost: makes sense re: NE, my US family is all in alfistan and they all love big brother so not the best allies in these apocalyptic times
billymg now has chicken coop with 20 some chickens, first dip of toes into self sufficiency
billymg: signpost: re: rural, what made you choose northeast as opposed to other US locations? also if you ever find the time to blog about the homesteading i'd really like to read
billymg: "Why chrome is still not dependency for openrc? - You still have time before election... " - posted nov 2, 2020
billymg: asciilifeform: you might find this thread amusing. the broken english and overall tone reminded me of the 4chan "why you want rail for kalashnikov" meme
billymg: if i add it to python-exec.conf `eselect python list` shows it, but as (uninstalled)
billymg: asciilifeform: some weirdness with python2.7, was able to install it just fine, and eix shows it installed along with the 3.x version, but eselect-python won't recognize it
billymg: http://logs.nosuchlabs.com/log/asciilifeform/2021-09-01#1055919 << works, at least can startup the server and connect with psql, not tested beyond that
billymg: !. uptime
billymg: hrm
billymg: !. uptime
billymg bbl
billymg: asciilifeform: my gcc4.9.4 install seemed to come with a few sub-variants (see paste)
billymg: i dunno what the different variations mean though
billymg: first try
billymg: lol it built with --nodeps
billymg: ah, right, i still keep forgetting about that
billymg: i don't currently have any tmpfiles packages installed, but postgres wants one
billymg: asciilifeform: yeah, i followed that guide and used the linked ebuilds to rebuild openrc and emerge xorg-server
billymg: asciilifeform: nice, and i've incorporated those into my local build now too
billymg: it's for postgres at the moment but other packages have asked for it too
billymg: asciilifeform: i'm getting blocked by the masked virtual/tmpfiles package, i noticed there's one on my rockchip but looks like it could be a shim version
billymg: mats: veggies and meat or veggiemeat?
billymg: i think it's funny how ridiculously obvious it is that the mask thing is a class signifier
billymg: flights*
billymg: chartered flight don't require masks either, perhaps they'll take btc when exchg rate is $1mn
billymg: lol
billymg: signpost: lol, i haven't tested the bandana look here yet
billymg: bingoboingo: yeah... also not personally interested in that either
billymg is still holding out for maskless travel
billymg: bingoboingo: i still want to visit but last i checked there was still mandatory quarantine or similar that i couldn't stomach
billymg: asciilifeform: what's the trade concession? pay upfront for x years of tax free import or something?
billymg: mining*
billymg: signpost: there are two companies working hard to decentralize minging now: http://upstreamdata.ca and http://gam.ai
billymg: signpost: just meant the extrovert/salesman type
billymg: http://logs.nosuchlabs.com/log/asciilifeform/2021-09-01#1055809 << this assumes kraken uses the same stack for its customers as it does for its cold storage
billymg: http://logs.nosuchlabs.com/log/asciilifeform/2021-09-01#1055806 << yeah, will always need a pete_dushenski or dorion out on the golf courses. i don't even think this is a bad thing, just gotta find one
billymg: i know i know, just joking
billymg: it's already a cult
billymg: how about a religion?
billymg: lol i wish i knew
billymg: signpost: in that case any contribution to wotco comes with tax bill as well
billymg: perhaps can share their secrets
billymg: but also seems like there are people in this chan that have this solved
billymg: i'm also not sure how *any* number circumvents the spendability problem
billymg: signpost: sure, also unclear whether this is dependent on current exch rate or not, perhaps in october btc at 100k lol
billymg: i understood the 100 number to mean, can safely say "fuck you" to fiat clients for now
billymg: signpost: just looking at the nicklist in here i estimated there might be 8 people who'd be willing/able to put up 12.5 (for me this is sort of my ceiling but there might be others here who could put down 2-3x that)
billymg: i figured that's what we were talking about re ascii's 100 number
billymg: former
billymg: signpost: re funds for wotware i'd be willing to commit ~10
billymg: i think i'm going to shelve this for now (since i have a working firefox) and move on to the rest of the stack
billymg: asciilifeform: hmm, my font issue seems to be a bit different. chromium appears to be unable to access system fonts, but renders fine when the website embeds fonts via css
billymg: while others render fine
billymg: i might need to install some extra fonts though, certain websites render with missing text
billymg: i glanced at the patches and assumed they were making it optional
billymg: asciilifeform, shinohai: this version of chromium finally built on my dbus/systemd-free gentoo. the solution was to comment out the musl related patches in the ebuild and set the -system-icu USE flag. i'm a bit embarrassed that i didn't try commenting out the musl patches sooner since i'm not on a musl build but
billymg: thimbronion: would be happy to help there as well
billymg: thimbronion: ah, what's the change? i was looking forward to it actually
billymg: this could be interesting, i'd also chip in to free asciilifeform
billymg bbl
billymg: mats: i think there are two camps, those who believe that the only tax on selling btc is the one printed (e.g. 20% longterm capital gains), and those who believe there is also a hidden tax (e.g. you're now on list of who to go after first when they 6102 it)
billymg: asciilifeform: so if 10 people here each put up 10 you'll be fulltime on asciiwarez??
billymg: currently i have one, my local /usr/local/portage one
billymg: how many overlays can one have?
billymg: shinohai: you rolled your own ebuild for that? i didn't see it in the tree
billymg: asciilifeform: i actually haven't even gotten to that part of this build! still on stock gcc 10 whatever
billymg: trying now with 512gb swap and the -system-icu flag
billymg: asciilifeform: no dice, even with 512gb of swap
billymg: shinohai: building again with 100gb swap and -j9...
billymg: might try that next actually...
billymg: 0 swap right now, but 1tb drive, so could add a large swap file
billymg: it does look like it's just running out of memory, but running with -j1 didn't work either (still runs out of memory)
billymg: shinohai: and with -system-icu, a slightly different error: http://paste.deedbot.org/?id=z6Re
billymg: shinohai: http://paste.deedbot.org/?id=C8Pa
billymg: shinohai: hmm, toggling that flag doesn't seem to have any effect on chromium's build, it's ignored in the USE list
billymg: as in possibly fixable with -ltinfow?
billymg: shinohai: but potentially this is the same problem i'm running into with chromium now?
billymg: s/built emerged/emerged
billymg: shinohai: ah, sorry, i think i missed that logline earlier somehow. i built emerged weechat from portage, no special steps/flags were necessary
billymg: i too would love to know, and would pitch in
billymg: seems they occur about once per year
billymg: i always love these "what's asciilifeform's magic number" threads
billymg: asciilifeform: same error with -j1, but at least since it was only one thread it made it easier to see the actual line it failed on
billymg: will now go back to using system-icu and try with -j1
billymg: asciilifeform: ok, will queue that variation
billymg: asciilifeform: this is the error with system-icu, running again to get the different error with -system-icu
billymg: and i'm building the version from this overlay, the upstream versions not fooled by atk-bridge shim, whereas this one makes atk optional provided you have working gtk+
billymg now has working browser so probably won't put too much more time into this, but is now building chromium again in order to reproduce the errors for the logs
billymg: and hey, look at that, firefox-91.0.2 built successfully, and works
billymg: seems, to me at least, equally 'spray and pray' approach as massaging css
billymg: lots of forum posts pointing to nodejs version, or maybe it's icu version, or clang version, etc. etc.
billymg: then it ran into another bug with ninja and again looked like it ran out of memory
billymg: some users reported similar issues, suggested building with -system-icu USE flag
billymg: chromium failed on some ninja mksnapshot thing, and looked like machine ran out of memory (despite having 32gb to work with)
billymg: i'll post relevant details here, currently churning through another attempt at firefox
billymg: i'm now running into other problems with both chromium and firefox
billymg: asciilifeform: it might be open source, but still require commercial license?
billymg: asciilifeform: not sure, but i know it wouldn't emerge prior to installing gtk+3
billymg: asciilifeform: sublime requires putting an accept line in package.license, yes
billymg: ignore last few characters, getting used to weechat
billymg: nnn
billymg: * We try not to diverge from upstream when there is no clear reason to. Avoiding dbus in the context of graphical applications was not, until now at least.Yy
billymg: * It is used in unittests frameworks such as dogtail.
billymg: * Upstream made the decision to make accessibility framework mandatory.
billymg: The reasons for why at-spi is not optional in gtk+:3 are:
billymg: luly bug report thread here: https://bugs.gentoo.org/show_bug.cgi?id=669234
billymg: someone also made an ebuild for it, which also worked and tricks portage into thinking it's installed
billymg: for anyone who's curious, this atk-bridge-2.0 shim worked for me. was able to emerge gtk+3 without pulling in any systemdisms, and from there was able to emerge my preferred text editor, 'sublime text'
billymg: punkman: exactly
billymg: imo "renounce" is legal pill and therefore only works if rule of law exists
billymg: mats: heh, yeah. and there are public examples of people who renounced, then later usg decided "he didn't do it right, plz extradite"
billymg: but i haven't looked in a while, not sure
billymg: mats: i believe that's what was being discussed re: the '30%'. iirc < 2mn and it's just a ~$2500 filing fee
billymg: i know, 'bonsai kitten' and all, but many seem to recognize 1) something's fucked up and 2) something must be done
billymg: asciilifeform: i think a lot of people realize this, and admitting there's a problem is the first step
billymg: asciilifeform: or apprentice, doesn't have to be own dna
billymg: asciilifeform: then have children, transfer keys to them
billymg: asciilifeform: yes, they've been 'firing on all cyclinders' lately, but i still don't think they can do this forever (like they do with gld)
billymg: heck, in last year
billymg: last i checked btc more than tripled in last 10y
billymg: asciilifeform: that's hyperbole, even by your own account btc has done much much better than inflation
billymg: asciilifeform: it does seem like they've done to btc the same as they did to gold via paper markets, but btc has properties that gold doesn't that make this harder to sustain
billymg: asciilifeform: doesn't have to be actual plot as described, but there are other ways that computing 'gives claws', e.g. we now have btc, which can be stored in such a way that can never be seized, short of (successful) torture
billymg: dominant weapon* tech
billymg: and yes nothing has happened yet, but the point in orwell's essay is that based on the dominant tech of the time is where society leads, given enough time
billymg: and if that's true then greatest weapon is the screen, not the bomb
billymg: point i'm trying to make is if you could MITM their devices you might be able to get something done, even if they can't "think". this afterall is exactly what enemy does
billymg: you don't need a million "thinkers" just one who can send messages to their screens
billymg: asciilifeform: but say their version of fb told them to go to davos
billymg: asciilifeform: that example seems like the problem is their version of fb (information tech) and not lack of resistance (willing to suicide seems like enough motivation)
billymg: asciilifeform: i've been thinking of that 'you and the atomic bomb' essay and it's central point. what i was wondering is whether the atomic bomb is really still the dominant weapon today, or if you could say the dominant weapon is now information technology, and what this would mean (i.e. does a sane os on sane silicon give people their 'claws'?)
billymg: asciilifeform: it's a shitty state the world is in now, won't disagree there