1800+ entries in 0.001s
ossabot: Logged on 2020-02-06 19:19:05 mircea_popescu: diana_coman,
tbh,
the reflection looks pretty cool ;
there's prolly a hook
there
to specify "no reflection past heightmap" or somesuch
to get rid of
the ultramarine legs
diana_coman:
http://logs.ossasepia.com/log/trilema/2020-02-06#1957921 -
thanks; re
that hook, I haven't seen it really and I doubt it exists as such;
thinking of it, it's more likely
to be something of a side effect again of how
the water plane,
terrain and character sprite interact; for
the
time being I'll let it be as it doesn't seem a high priority really.
mircea_popescu: diana_coman,
tbh,
the reflection looks pretty cool ;
there's prolly a hook
there
to specify "no reflection past heightmap" or somesuch
to get rid of
the ultramarine legs
diana_coman: mircea_popescu: it has indeed everything needed as far as I know and certainly rsa, keccak, oaep,
the whole package.
diana_coman: mircea_popescu: I'd very much like
to finally move off koch-gpg and so ~anything in
that direction is quite of interest
to me; I'll gladly help re eucrypt
too if/when needed, if
that's a block for anyone.
mircea_popescu: which is perhaps a fine candidate of folding in (or at least,
that's
the general
thrust of
my comment -- "why do i have
to go outside of
the v
tool for v work ?")
mircea_popescu: jfw, re
the whole
musl & locales issue, it might be an idea
to signal
to
them, "look, we use musl, and we don't
think
this is a good idea". irrespective of whether it does anything, at least
that way
they can't say
they didn't realise "unanimity" is hallucinated etc.
mircea_popescu: meanwhile in 2010-vintage romanian free love,
http://archive.is/gybYM (16 yo gypsy alpha raped 10yo visiting girly in school
toilet. anal.
then zipped up, went back
to playing ball.)
hanbot_abroad: mircea_popescu i managed
to get a comment into your mod queue
mircea_popescu: ditch
the shitty hashes stick
to keccak (which is particularly well fit for
this job, as it happens), ditch
the bad parameters use something convenient (why are keys 4096 bits and
the hash 128 ?)
mircea_popescu: re-writing
the signature part starting with say diana_coman 's eucrypt could give us a chance
to ditch all
the warts of a very usgistani past, and even maybe implement PSS or somesuch.
mircea_popescu: there's a lenghty pile of disadvantages
to
the current mechanism we use, not least of
these being
that it actually imports koch-pgp. it also does suspect signature shenanigans of all sorts, which could potentially present security risks
mod6: mircea_popescu: re: rss feeds & feedbot; will check it out. sounds like a good way
to stay abreast.
mircea_popescu: pick some people you actually care about, and connect with
those. stop
trying
to shinohai yourself
the imaginary world of where you've wrung
the persons out of
their deeds, and now looky
there's just
this pile of hallucinated freedom for you
to wither atop of.
mircea_popescu: where "me" is just about a wildcard, it stands for "anyone" because
that's what
the fuck "no separation from politics" means.
mircea_popescu: but in general, if you can't be arsed
to read my blog, you're cordially invited
to get
the fuck lost and in no case pretend like you're using my patches.
mircea_popescu: ain't gonna ever happen,
the republic's
the republic for reasons.
mircea_popescu: this imaginary self-situation in
the seat of god where lo and behold you'll somehow see all patches... what
the fuck do you
think
this is,
the github flatlands ?
mircea_popescu: and secondly,
the rpublic isn't intended
to work
this way. you're supposed
to hear about
the patches you hear about. you aren't happy with
the patches you hear about,
make better friends.
ossabot: Logged on 2020-02-03 20:22:10 mod6: All I'm saying, is
that without said possible future solution, it may be somewhat difficult
to
track
through 39 different blogs watching for patch submissions all
the
time.
mod6: I appreciate
the poking re
things
TRB.
jfw: observe for instance how
the gardener nearly missed a couple patches anyway but I poked, not even being a signer of any as yet
mircea_popescu: so do not ye worry "the gardener may miss one". really,
the walflowers aren't even worth pissing on if
they were on fire.
mircea_popescu: i really don't
think
the known failures of
the "engineer" mindset need any encouragement.
this whole "i'll overwhelm my horrible idiocy with supposedly worthy deeds" nonsense has little space left
to frolic in.
mod6: I like
this better
than mirrors
too, as it seems
that
those always
tend
to disappear or end up way out of date without explanation.
mircea_popescu: mod6,
there's a bunch of code shelves you might end up on, if you do
the right and proper
things.
mod6: Yeah, not
that we would want stuff from nonpersons, just in
the case where, as we've seen recently (even with my own blog) where it was down for a few weeks, or months as people are between hosting or such.
mircea_popescu: if some shit someone vomited
that nobody could be arsed
to mirror disappears,
this is a great and wonderful gain for everyone.
mircea_popescu: i don't get
the logic whereby anything is ~wanted~ of someone who "disappeared completely"
jfw: as always, keeping information alive requires live people
to do it
mod6: I suppose
that's fair.
jfw: Or new maintainer starts new
thread, and someone puts up an archive of your blog assuming anyone cared enough
to keep one
mod6: (I appreciate
this discussion, fwiw.)
mod6: I suppose
the foundation should maybe have its own blog
then.
mod6: jfw:
This might be fine, but what if I'm not
the guy? Meaning, let's say
that
there are
threads on my blog, people publish stuff in
the comments, or wherever on
there, i dunno, pingbacks somehow. If someone else
takes over, I get hit by a bus, blog goes bye bye from non-payment, what happens
to
thread on new guys blog?
mod6: But again, if
there is going
to be a future solution for
this in
the form of some kind of existing bot extention,
then I'm all for
that. At least it would, presumably, notify
the gardener (somehow)
that
there is something
to inspect.
jfw: perhaps a comment
thread on your own blog where someone proposing a patch can "officially submit" it?
mod6: All I'm saying, is
that without said possible future solution, it may be somewhat difficult
to
track
through 39 different blogs watching for patch submissions all
the
time.
mod6: Anyway, sounds like
there might be a solution
to
this between what billymg and lobbes are working on + maybe another added step?
mod6: jfw: aside from a possible mp-wp (future) solution
to
the issue; several clicks notwithstanding, I mearly worry
that somehow,
the gardener may simply overlook or miss one.
jfw: "It puts it on me
to chase
these down" -
this seems
to be
the core of it; why not leave it
to
the sponsor(s) of a patch
to keep
their blogs online & organized enough
to find patches (category, code shelf or whatnot)? It's
then, what, couple clicks per patch or a wget
to import
to your own collection -
trivial compared
to
the effort of actually reviewing
the patch I'd imagine
mod6: Ah, alright. Maybe
that's it, just a few moving parts here.
lobbes: if I ever get
the
mp-wp logger complete.. well it interfaces with
the mp-wp database already so
theoretically possible
to hook
that up
to what billymg is doing down
the road
BingoBoingo: Maybe not all of it, but with
the lobbes bot
that logs into mp-wp,
there's space
mod6: Ok, I saw
this in January, I may need
to re-read it, but on first pass I didn't gather
that
this would do what I needed it
to do.
mod6: Now
the idea
to make a bot
that would somehow interact with blogs, is an idea I hadn't
thought of before.
mod6: I fully agree, mailman is a ancient artefact back from, probably literally,
the pre-september internet days. Yes, it's quite, well, frustrating
to say
the least.
BingoBoingo: The second biggest problem is
that mp-wp is growing, billmg is putting work into making mp-wp do code hosting more cleanly, and
the mailman software
the mailing list used is an awful, poorly documented
thing. If an IRC patchbot is
the way
to go, hooking it into mp-wp seems like
the better direction
to go.
mod6: yes. a bot
that would do
this over irc, is basically what I've been
thinking about.
BingoBoingo: The biggest problem with
the mailing list is
the mail part.
mod6: I dunno, anyway, I
thought it could be good for people
to go ahead and submit a vpatch, and seal
to
the bot, where it can be checked for L1/L2, and stored for
the future, all in one place, making less work for
the gardener.
mod6: Not
that I love having a centralized bot, either. It's just, I'm not 100% sure what
the best way
to go is on
this.
The mailing list was centralized, it worked fine.
mod6: TMSR Lords and others seem
to publish all
their code on
their blogs, which, I
think is fine. But my hang-up with allowing people
to post
TRB patches/seals on
their blogs instead of sending
them in is
two-fold: 1) It puts it on me
to chase
these down. 2)
Then I have
to place
them somewhere for long-term keeping anyway. As we've seen, people's blogs get rather large, hard
to find
things, or disappear complet
mod6: Now, I've
thought
that over a bit;
the first
thing
that came
to mind was deedbot. However, I
think
that deedbot probably isn't
the right place for developer doodles. Plus, each one costs actual money.
mod6: I find it fairly easy
to go and dig stuff up in
there if I need
to do so. I
tried
to stand it back up, several
times, in fact, since it's home on Pizarro went down. But I haven't been successful
there. Upon speaking
to jurov about it, he suggested instead
that we create a bot for
this purpose instead.
mod6: For me, I've always liked
the mailing list. It's cumbersome, and it's had it's pain points in
the past. Everyone, at one
time or another, has had a problem getting
things stuck in its queue, etc. However, when it works, it does work pretty well. I like how it checks
the WoT on submission, has an archive, and we can all go back and look at it years and years later.
mod6: Regarding your question re IRC bot for vpatch submission: I've
thought about
this quite a bit. And I struggle with
this.
mod6: mod6_phexdigit_fix <<
This one I'd like
to put in for sure, as it seems, at least according
to my own analysis
that
this is a legit bug (does anyone disagree with
this analysis?). Just would need a simple regrind at
this point. Can do
this month as well.
mod6: asciilifeform_whogaveblox << I'm more open
to comments on
this one
too, I don't see a big issue with it being in
the
tree. I'm also using
this one in a
test environment (for quite some
time now). Haven't seen any issues with it at
this
time. It would need a simple regrind. Can do it
this month as well.
jfw: ^ yep, complete non-starter
to debug "I'm having
trouble with block 00000000...1<snipped>"
mod6: mod6_excise_hash_truncation <<
This one in particular is nice, and I've been using it in my
test environments for a while. I'm fine at
this point with adding it in, it just needs a regrind (again). Can work on
that
this month.
mod6: (23:28) <+jfw> oh sorry, asciilifeform_aggressive_pushgetblocks is in
there, just
the other
three
then. << correct.
mod6: I've updated
the How-To document
to reflect
that, but otherwise, have had good luck on my own with building. Am interested
to hear how it goes for others,
though.
mod6: Let me work
through
these here. Glad you're going
to
try out a build. Please do let me know how it goes. I did just find
three 'gotchas' on CentOS (6.10); whereby 'bison', 'flex' and 'patch' all need
to be installed manually. Apparently
these are not a part of
the default system.
jfw: oh sorry, asciilifeform_aggressive_pushgetblocks is in
there, just
the other
three
then.
mod6: Hi
there jfw,
thanks for
the questions.
jfw: Re item 4, why is a bot needed for vpatch submission? Don't blogs already cover all
the aspects - publishing, commentary, discussion, referencing, notification? (Perhaps your article will clarify, in which case don't mind me.)
jfw: mod6: glad
to see progress on
the keccak
tree. I intend
to
take a look at
the patches and
try a build, might not be for a while
though. What I'm not seeing
though: what is
the status of asciilifeform_aggressive_pushgetblocks, mod6_excise_hash_truncation, asciilifeform_whogaveblox and mod6_phexdigit_fix ?
mircea_popescu: incidentally, anyone know
this eoghan hayes ( rootpaw / lleti) redditard ?
jfw: Not
that I seriously
think
they *should* use
those for all numbers, I can't see
that being anything short of hideous given
that it's C, manual memory management and all, "what do you mean you don't know how big
the struct will be"
ossabot: Logged on 2019-02-17 10:03:36 mircea_popescu: "/*
The count field we have in
the main struct object is somewhat limited, but should suffice for virtually all cases. If
the counted value doesn't fit, re-write a zero.
The worst
that happens is
that we re-count next
time -- admittedly non-trivial in
that
this implies some 2M fdes, but at least we function. */"
jfw: ah right, I forgot, in soviet russia if-statement sorts
the other way so
they had
to copy+paste+tweak it for
that locale.
jfw: I don't see
the 5mb of code fwiw; if you can have nested expressions anywhere,
then you get hidden
temporary variables, so in principle costs no more
to allow
them anywhere.
That gets
to
the essential vs accidental complexity
though,
totally wouldn't surprise me
to find 10k LoC in gcc dedicated
to compound expressions in if-statements specifically.
mircea_popescu: perversely,
this "superior" code's even more inscrutable
than ye olde asm. which civilised people could in fact read, unlike
the
current shit mircea_popescu: on
the contrary :
the "terseness" supposedly achieved by if (a(x) > b(y) ) z sorta constructions is not merely standing on its own ; but it creates personal investments. now you can't comment your code, either, because you're so clever you saved five bytes of
text at
the cost of five megabytes of object code. so if you
then write a sentence explaining wtf you did, well... you're a sucker now, aren't you!
mircea_popescu: this doesn't automatically mean EVERYTHING has now
to be unwound, and so branches only on registers for computers and women kept in
the gyneceum like ye olde greeks.
mircea_popescu: the average family's not any happier since women are permitted
to seek divorce
than
they were before ; and code gained nothing from
this particular "improvement"
that seemed but never delivered.
mircea_popescu: you know, not every
thing
that seems clever actually delivers what it promised.
jfw: ah, back
to assembly language
then, branches only operating on registers? heh
mircea_popescu: it doesn't even fucking save memory,
the five bytes
this'll save over a lifetime is dwarved by
the five megabytes in extra libc needed
to support
the insanity
mircea_popescu: jfw, i honestly don't like evaluating ifs.
there's really nothing gained, besides compiler weight.