Show Idle (>14 d.) Chans


← 2023-07-01 | 2023-07-03 →
11:54 phf 􏿽http://logs.nosuchlabs.com/log/pest/2023-07-01#1027975 << canonical place for vtools is http://btcbase.org/patches?patchset=vtools, but it has two problems at the moment a) it strongly follows tmsr model, which even survivining people don't really follow anymore b) it's not been up
11:54 phf 􏿽dated with the patches that were produced by third parties after i quit tmsr, and that i haven't had a chance to integrate. cursory look those patches are of dubious quality, but they are included with e.g. pentacle
11:55 dulapbot Logged on 2023-07-01 10:33:16 awt_akris[asciilifeform]: Where is the best place to get ahold of phf's V these days? I have it, but I'm attempting to document the installation process for the new station and would like to link to a V.
11:59 phf 􏿽a and b here are tied together. so the tmsr approach and the position was that you don't integrate patches without having read and understood them. something like a tarball of source as some people started doing on their blogs, or a semi-automatic approach where patch press is invi
11:59 phf 􏿽sible to the user like in pentacle were both originally seen as invalid approaches
11:59 phf because both kind of defeat the purpose of vpatches as being standalone literate artifacts
11:59 signpost don't know that I agree at all with "invisible" here.
12:01 phf so the original ideological position was that you kind of need to understand the press, the order in the press, and having read each individual patch, you choose conciously to apply said patches
12:02 phf so that was how btcbase/patches presentation was modeled. it doesn't make it trivial to grab the whole spitoon, instead you look at the press tree, patch after patch, download them as you understanding of the tree becomes refined
12:03 phf but of course that's not convenient when you "just want to" run vtools, so people started packing vtools into tarballs on their blogs. "you need vtools to compile my widget!! and here's a tarball of vtools, don't think about it"
12:05 signpost this is all pretty rude, and rooted in MP's sadism fetishes. the src tree is meant to be individually curated, but obviously needs to at least contain with what to self-build.
12:06 phf signpost, ok.
12:08 phf fwiw awt asked for "phf's vtools", but what's in pentacle is not phf's vtools, because it has a bunch of patches that were added later by people that i don't know, and that i don't think anybody has actually read. i'm giving a background on this sad state of things.
12:11 phf 􏿽and by read, i mean that later patches for vtools were made by bvt, and in order to press vtools, people just throw bvt's signature into sigs folder and call it a day. presumably to make the result "phf's vtools" one should harrass me to review the code, and sign it myself, which w
12:11 phf 􏿽ould be the V way of doing things. or sign the code with own sigs and then the answer is "well, there's phf's vtools, but it's missing some later additions. if you want to use trinque's vtools, i've read phf's and bvt's patches, and made my own signatures and you can use that inste
12:11 phf 􏿽ad"
12:11 phf etc.
12:15 signpost this cuts to the heart of the failed religion imho, for obvious reasons.
12:16 signpost one, you haven't bothered to look at the thing, or you'd have seen it's just a folder with a bunch of v-trees in it, and some scripts to denote the dependencies between them.
12:17 phf http://logs.nosuchlabs.com/log/pest/2023-07-02#1028040 << maybe try and understand my point instead of seeing is as some kind of slight and affront to your work
12:17 signpost two, if we're going to pull out "just want to", let's pull in "and anyway" from the anal child, because even if you had, it'd violate the religion in some other way.
12:17 signpost the inability to make incremental gains in that period is worth calling out.
12:18 signpost I've been trying to beat people out of this "well it builds on *my* pile of driftwood and cum socks" for years.
12:18 shinohai https://www.youtube.com/watch?v=Or8Ba20Q_OY
12:19 dulapbot Logged on 2023-07-02 12:14:35 signpost[asciilifeform]: one, you haven't bothered to look at the thing, or you'd have seen it's just a folder with a bunch of v-trees in it, and some scripts to denote the dependencies between them.
12:22 signpost to put a finer point on it, if awt wants to package his thing for normies, I'd build a static bin of it, and sign that. nobody's going to be enrolled in the religion that doesn't already have the inclination. and probably no one that happens to use it will check the signature.
12:23 signpost oughta make the thing into distributed telegram.
12:29 signpost http://paste.deedbot.org/?id=GC5y << explain to me what is invisible here.
12:30 signpost that tree *is* the interface the operator is meant to curate. take out all the keys and patches you like.
12:39 signpost guy will spray piss in all directions declaring himself the one true adherent to a thing he quit, and then when *gently* challenged disappears.
12:44 asciilifeform signpost: he did have a point tho. ( per mp's original ex cathedra iirc this knot was supposed to be cut by 'i'm signing with my shit-i-havent-read-cuz-itd-take-over9000-years key' , but no one, mp incl., actually did this at any pt )
12:46 * asciilifeform suspects that the correct pill here would be to resurrect the cl vtools, as shinohai tried to do, but 'over9000 years', asciilifeform still using the 'dirty' one discussed above
12:47 signpost it's too bad I didn't find time in the multiple years I worked on that item to read any patches.
12:47 shinohai cl vtools is very reliable, even made a Makefile for it.
12:48 asciilifeform shinohai: thrd wasn't , as asciilifeform understands, about 'reliable' tho
12:50 * signpost had about enough of the hypocrisy of folks doing most of their work in a clang built by apple, then coming along to say the bootstrap process for a vintage gcc was done sinfully.
12:50 asciilifeform http://logs.bitdash.io/pest/2023-07-02#1027621 << per asciilifeform , the correct pill for this entire class of problem is this , but we still aint got it ( asciilifeform's -- 'hands too short' , other folx did make attempts, most recently: [http://zzz.st/p
12:50 bitbot Logged on 2023-07-02 12:03:58 phf[4]: but of course that's not convenient when you "just want to" run vtools, so people started packing vtools into tarballs on their blogs. "you need vtools to compile my widget!! and here's a tarball of vtools, don't think about it"
12:50 asciilifeform rogz/#tunalisp][jonsykkel] , but it's a long ways off )
12:50 dulapbot (asciilifeform) 2022-03-17 asciilifeform: mangol: what's imho needed is a fully 'autarkik' scripting lang, which sucks in 0 shitlibs and runs, like 'M', on any extant pc iron under all linuxen from past ~15y
12:51 asciilifeform ... most recently: jonsykkel , but it's a long ways off )
12:51 shinohai "reliable" for me meaning I don't hafta scratch head wondering what patches I need and don't from 7`248 different ex-tmsr blogs.
12:52 asciilifeform shinohai: rright, but you can prolly see how this kinda thing threatens to throw out what little progress we made in re civilized ecosystem
12:53 asciilifeform 'grab this tarball, build, run' is precisely what the monkeys do. notion was to distinguish selves from the apes
12:53 * asciilifeform as guilty as the rest, also had 'vtools tarballs' linx in e.g. ffa
12:54 shinohai I can see that, yes.
12:54 asciilifeform problem is eminently solvable tho! and asciilifeform appreciates signpost's, jonsykkel's, et al. efforts to solve
12:55 asciilifeform it realy oughta be possible to bootstrap a readable vtron ( and w/out in the process having to bootstrap an entire linux, conceivably )
12:55 * asciilifeform must bbl
12:57 signpost http://logs.bitdash.io/pest/2023-07-02#1027642 << yes, in his hallucinated version of my distro he had a point.
12:57 bitbot Logged on 2023-07-02 12:44:32 asciilifeform[4]: signpost: he did have a point tho. ( per mp's original ex cathedra iirc this knot was supposed to be cut by 'i'm signing with my shit-i-havent-read-cuz-itd-take-over9000-years key' , but no one, mp incl., actually did this at any pt )
12:59 signpost there was no specific criticism other than "automated".
12:59 signpost okay, by the same token `v press` is sinful, and shouldn't be automated.
13:05 signpost let's address the egotistical "baw my work" barb too. I do not want to waste my time with yet another "oh shit, nobody actually knows how anyone else's pile of garage-wars shit builds". I will not dedicate another minute of my life to that.
13:05 signpost I do not want to see *others* wasting their time with that!
13:06 signpost nor wearing hair shirts and prostrating up and down for miles because it pleases some other idiot who thinks himself his better.
13:06 signpost fuck. that. religion.
13:10 signpost and one final point. this vulgar habit of filling in all the blank space in the conversation with the lowest possible explanation ("he does not read patches") is a solvent. and it works.
~ 2 hours 47 minutes ~
15:57 * asciilifeform ftr not read the patches in question. so 'not throws 1st stone'
15:58 asciilifeform signpost: there's a serious risk of 'conflating the warm and the soft': one wants proggy to 1) work 2) easy to build 3) make sense ('fit in head' ideally) . these reqs are in eternal tension; for so long as we're stuck on 40+yrs of idjitos/crashlang -- more or less incurably
15:59 signpost I'm not going to tolerate phf popping off about me contributing to some kind of moral decline.
16:00 asciilifeform asciilifeform introduced v to try to squeeze in some (3) in (1+2) which was starting to seem like a physical impossibility
16:00 signpost everyone was humping along a wad of shitty third-hand scripts to bootstrap a musl gcc before I fixed the problem.
16:00 asciilifeform signpost: indeed
16:00 asciilifeform gnat in particular
16:00 signpost nothing stops one from *removing* patches from what I put together, and I'd applaud it.
16:00 signpost that was the whole fucking point. put it together where it's clear what's needed today, and have where to ratchet it smaller.
16:01 asciilifeform ^
16:01 signpost I'm fucking tired of the phf-flavor narcissism where the guy accuses me of e.g. googling postgres internals on the fly, or whatever "I know what you're really up to" horse shit.
16:01 asciilifeform the solution to 'grr buncha patches nobody read' is inescapably 'read'em & sign', issue is that there is a megatonne of these and 2-3 people with any cycles at all atm
16:01 signpost note *I* signed what I read.
16:01 asciilifeform aha
16:02 signpost pentacle is the state of the post-tmsr world, as was.
16:02 * asciilifeform read earlier thrd as 'phf ran outta cycles and laments that various things will prolly never get fixed' rather than 'phf points finger at signpost an' says he's a lamer'
16:03 signpost http://logs.bitdash.io/pest/2023-07-02#1027621 << this is his description of what I did, and fuck him.
16:03 bitbot Logged on 2023-07-02 12:03:58 phf[4]: but of course that's not convenient when you "just want to" run vtools, so people started packing vtools into tarballs on their blogs. "you need vtools to compile my widget!! and here's a tarball of vtools, don't think about it"
16:04 asciilifeform atm we're down to ~5 folx with working hands. asciilifeform buried in liquishit and emerges 2-3x/yr. phf afaik ditto. signpost, awt, jonsykkel still kicking.
16:05 asciilifeform ( 6? billymg also emerges , but iirc is buried similarly to asciilifeform in commercial shite )
16:05 * asciilifeform apologizes if fughot sumbody in this roundup
16:07 * asciilifeform suspects fwiw that phf not has a box with signpost's pentacle yet, and so still on state of art circa '19. but not knows for certain
16:07 signpost http://logs.bitdash.io/pest/2023-07-02#1027615 << concretely without the "blockwise read" one I couldn't shove gcc through, and we can either sit and feel superior that "there should never be such a large patch" or put gcc in the trb situation where at least shit can be hacked off, and *those* can be readable.
16:07 bitbot Logged on 2023-07-02 11:59:00 phf[jonsykkel|deedbot|awt]: 􏿽a and b here are tied together. so the tmsr approach and the position was that you don't integrate patches without having read and understood them. something like a tarball of source as some people started doing on their blogs, or a semi-automatic approach where patch press is invi
16:08 * asciilifeform recalls
16:15 asciilifeform afaik atm erryone remaining is carrying 'naggum's cross' to some degree. so asciilifeform aint aboutta throw mp-style 'wtf wai ditntcha fix/finish/invent $item! lamer, traitor' at anybody
16:17 asciilifeform per asciilifeform's pov, awt's pestrons, or signpost's pentacle, or billymg's logger, are heroic feats, rather than some kinda 'fulfilled obligation'
16:18 signpost hey man, not even looking for applause, but thanks.
16:19 signpost http://logs.bitdash.io/pest/2023-07-01#1027580 << it's this.
16:19 bitbot Logged on 2023-07-01 10:51:15 signpost: an aggressively censorship resistant stack is possible with pest, something like I described above, occ (aka squirt/slurp), etc
16:19 * asciilifeform was not able to build much of anyffin, aside from the rather junkyard spec, ever since started ~80h/w consulting ugh, so impressed with folx who were
16:19 signpost the darkening is going to get worse.
16:20 asciilifeform ^ 'can take this to the bank'
~ 5 hours 11 minutes ~
21:31 awt_akris well shit. I *was* incorrect in referring to the desired v as phf's, but at the same time wanted to acknowledge that it was based on his work, as when I started using it (yes from a tgz) I thought it was dianacoman's. I was not alone in doing so and can see how irksome that could be.
21:36 awt_akris Also literally impossible to press a v without a v. So either you download a v or you write your own.
21:37 awt_akris The difficulty with Diana's version is that ADA keccac is a pita to produce.
21:37 awt_akris I myself built it on my one remaining gentoo, long ago, and have been using that binary on different systems for years.
21:37 asciilifeform awt_akris: mno, it is certainly possible (if arduous) to manually verify the pgp sigs and then patch -p0 < ...
21:38 awt_akris ^ tru
21:38 asciilifeform recall that asciilifeform originally advocated hand-sewn equivalent of this process before wrote vtron
21:38 asciilifeform ( it understandably pissed off folx like mp who 'wtf, over9000 patches/sigs and by hand?!' )
21:41 asciilifeform http://logs.bitdash.io/pest/2023-07-02#1027703 << he did most of the quite fiddly work in that vdiff
21:41 bitbot Logged on 2023-07-02 21:31:56 awt_akris[awt]: well shit. I *was* incorrect in referring to the desired v as phf's, but at the same time wanted to acknowledge that it was based on his work, as when I started using it (yes from a tgz) I thought it was dianacoman's. I was not alone in doing so and can see how irksome that could be.
21:41 dulapbot Logged on 2022-12-18 13:14:33 phf[asciilifeform]: i've used e.g. clion to refactor gnu diff for vtools, and i don't think i would've bothered and managed otherwise. highlights unused code, variable and function renames, jump to source that's not stuck in the 80s, "who calls", etc. etc. etc.
21:43 * asciilifeform has 'pressed by hand' many times, even in recent yrs
21:44 asciilifeform generally if you've already verified n patches, and nao you grabbed n+1st, it aint particularly more painful than using actual vtron
← 2023-07-01 | 2023-07-03 →