186000+ entries in 0.103s
erlehmann: asciilifeform i have probably done more work on automatically capturing non-existence dependencies in real-world usecases (which seems also
trivial, in retrospect) – so is it now my idea? why should i care what
the guy says if he does not followup with code?
erlehmann: also non-existence dependencies. yeah, i read it at djb's website, but if his implementation does not exist (not published = not existing in practical
terms), i don't care about his opinions regarding
that.
erlehmann: however, having a single command in a build script
that says “rebuild
the
target,
then rebuild
the current
target” seems
trivial in comparison.
erlehmann: often not, yes.
to understand why URbit is designed in such an idiosyncratic way, i always suggest
to read unqualified reservations first.
erlehmann: i have no interest in arguing about djb's merits
tbh
erlehmann: but
then again i was sceptical of weev's aryan awakening at first as well
a111: Logged on 2017-05-31 18:37 asciilifeform: 'Generating large amounts of
truly random data is expensive. Fortunately,
truly random data can be simulated by pseudorandom data produced by a stream cipher from a much smaller key. (Even better, slight deficiencies in
the randomness of
the cipher key do not compromise security.)
The literature contains several scalable ciphers....' -- djb et al
erlehmann: i believed
the suggestion
to let NIST generate your private key was sarcasm
erlehmann: i
think it is like fefe said: a)
tenure b) does not care
erlehmann: phf i believe redo is an unpolished gem. sadly, many people implement it ALMOST correctly and
then do something stupid
erlehmann: and by ”massively shitty” i mean apenwarr used sqlite because “filesystem is slow of course”.
turns out 300 lines of shell script are faster
than more
than double
that amount of python if you actually benchmark and not
talk out of your own ass.
erlehmann: second: “i have written a redo implementation, does it work correctly?”, i again lost him. he suggested
to meet at a place at a
time during
the conference and i
thought i knew
the place but apparently i erred because
there was no place with
the name i
thought he had said.
erlehmann: first: “where is your redo implementation?”, i believe his answer was like “has
to be cleaned up before release” or something before i stupidly decided
to come back later bcause he was answering crypto questions
a111: Logged on 2017-05-31 14:17 asciilifeform: in recent sads, 'Our batch prime-generation algorithm suggests
that,
to help reduce energy consumption and protect
the environment, all users of RSA—including users of
traditional pre-quantum RSA—should delegate
their key-generation computa-
tions
to NIST or another
trusted
third party.
This speed improvement would also allow users
to generate new RSA keys and erase old RSA keys more frequently, limiting
the damage of key
theft.'
erlehmann: i had
two interactions with djb regarding
that
erlehmann: i
think
the main reason for me implementing and using redo is
that
the scripts work fine without all
the logic. it is just an optimization.
erlehmann: as
the saying goes in german: wenn du einen hammer hast, kannst du die ganze welt nageln!
erlehmann: regarding redo: needing dependencies is not limited
to compiling programs. datasets are also something. i build my website with redo. i have managed converted media files with redo. i would not want
to wait hours for
the re-encoding of each file every
time i rebuild a web site.
erlehmann: rc shell grammar is listed on
the man page. very small. nice. monkey like.
erlehmann: whatever, i use rc shell for day
to day work. but sh is portable and bash is not.
erlehmann: i also consider building stuff not very interesting. but redo is so simple
that it can be implemented in <250 lines of shell script.
phf: erlehmann:
try it now
a111: Logged on 2017-06-03 00:36 phf: also my own phf-shiva-swank is broken somehow, probably because i was pressing with not "real" v
tooling. actually i need
to fix
that..
a111: Logged on 2017-05-31 14:12 mircea_popescu: i suspect
the idea is
that systems which require something like make are broken anyway.
a111: Logged on 2017-06-03 00:35 erlehmann: asciilifeform do you have an opinion on DJB redo? or maybe even a simpler design? i would only
trust my own implementation and MAYBE
the one from jonathan de boyne pollard (although he requires a C++ compiler as a dependency, lol)
erlehmann: phf i hope
this change flies without a vpatch, should be evident it is not malicious
erlehmann: phf add following CSS
to stylesheet: pre { margin-bottom: 0; }
phf: (fwiw
this
table based approach also looks shitty in lynx)
phf: but
that makes copy pasting harder (you can't just select a chunk, because it now includes line numbers and blame)
phf: so
the way it should really be solved is
that i need
to better understand colorizing code (which i lifted from elsewhere) so
that i can wrap
the whole
thing in pre and just format it without
tables by doing <blame ...> <lineno ...> <line>
erlehmann: phf
the reasons seems
to be
that every single line on
the right side is wrapped in <pre> and <span>
phf: hmm, k. i haven't paid much attention
to
that styling, because i don't
think anyone's using it, including myself.
the press part definitely needs more ux work
phf: well
that was a bad example because
the file is all genesis
phf: it does an equivalent of patch, but without calling out
to c programs and without
the result (or intermediate steps)
touching
the file system at any point
erlehmann: well, vdiff would never be possible if diff and patch were definite about
their inputs
phf: for sure, btcbase is probably
the most aggressive, because
there's a very restricted state machine for parsing, but
there's still some ambiguity in recognition
that's an artifact of diff/patch being dodgy
erlehmann: if i ever get around
to making my own v
toolchain (probably in bourne shell, for portability and clarity), i'll address
that of course.
erlehmann: v looks like a nice approach as far as i can see (and is
the reason i chose
to join #trilema), but as i commented, i suspect
there exist parser differentials
phf: also my own phf-shiva-swank is broken somehow, probably because i was pressing with not "real" v
tooling. actually i need
to fix
that..
☟︎ phf: there's a handful of
those, for example in
the experimental
there's
three patches from polarbeard, a guy who joined as we were regriding a bunch of shit, so he had
to update his patches a bunch of
times and
then gave up. nobody cared
to reintegrate his patches properly yet
erlehmann: asciilifeform do you have an opinion on DJB redo? or maybe even a simpler design? i would only
trust my own implementation and MAYBE
the one from jonathan de boyne pollard (although he requires a C++ compiler as a dependency, lol)
☟︎ erlehmann: phf
that makes …
too much sense.
thanks.
phf: erlehmann: so
the arrow from nowhere basically means
that
the patch is broken, because it requires an antecedent
that's missing
phf: i actually patched it in
the offline, but haven't had a chance
to update
the deployment
erlehmann: apparently it does not come from anywhere in
this graphic
erlehmann: ah funny, it does not work with a slash at
the end
erlehmann: as a free-software enthusiast myself, i have managed
to experience many invitations
to “netflix and chill” entirely without netflix. freedoms preserved!
erlehmann: > A friend once asked me
to watch a video with her
that she was going
to display on her computer using Netflix. I declined, saying
that Netflix was such a
threat
to freedom
that I felt uncomfortable with promoting its use in
this way.
erlehmann: asciilifeform if you are interested, i suggest
to email rms yourself. he answers.
erlehmann: one could probably extend david wheeler's diverse double compiling
to any
type of
tool if sufficiently paranoid
erlehmann: thus
the remark with
the star
trek replicators, which he likened
to compilers. just do it yourself.
erlehmann: as far as i can remember, we did not
talk about fpga. his example was some company creating your board and not you.
a111: Logged on 2017-01-16 23:36 asciilifeform: leaving entirely aside
the question of whether ice40 can in fact be made
to do anything useful with
the 'open'
toolchain discussed earlier, or whether a
toolchain
that required clang, llvm, and
ten other poetteringesque abortions is 'open'
erlehmann: i once met rms at a conference and asked him about viability of free hardware. his answer was along
the lines of 1. apparently non-free hardware has worked for decades 2. whoever makes your board can still subvert your
trustworthy design, you can't check
that 3. maybe if we have star-trek-style replicators one day, hahaha
a111: Logged on 2014-12-11 01:52 asciilifeform: decimation: notice
that all known fpga manufacturers (xilinx, altera, lattice, a few others) have
the same business model
a111: Logged on 2015-06-17 13:17 asciilifeform: you can pick up a
textbook and write a dram controller for fpga from first principles - and it won't work. because, for starters, only a small number of output cells in
the chip can function on both rising and falling edge of clock cycle (what 'ddr' means) and only xilinx's closed
turd knows where
they are in
the routing fabric;
erlehmann: shit
toolchains are endemic in lots of fields it seems
erlehmann: for some reason i
thought of asciilifeform when i learned about
the MNT VA2000 amiga graphics card
phf: ascii's approach was
to put custom firmware into an old keyboard, so
that's half way
there
erlehmann: i
think it is unlikely
that a non-self-built keyboard can do most
things relevant
to daily usage in hardware. especially layers, mod keys and compose.
erlehmann: i suspect
the effect is not as notable for other people as for me, but qwertz hurts, neo2 does not.
erlehmann: chatting is something where i prefer
to prevent my hands hurting
erlehmann: well, in my experience it
takes max.
three weeks
to learn a layout.