log☇︎
348 entries in 0.663s
mod6: I should have examined/tested your mpi vpatches, alf. I'll continue to try to be a second pair of eyes, reading them. There's no substitute for reading. For those following along, take note.
mircea_popescu admires how the "finished" mpi managed to take a whole day of 3 people's time and shakes his head displeasedly.
diana_coman: 2 out of 5 hunks FAILED -- saving rejects to file mpi/README.rej and contents of README.rej are here: http://p.bvulpes.com/pastes/J8ESM/?raw=true
mod6: now, on the other hand, yeah, i saw the second_cut vpatch link removed from loper... but I went ahead and updated my sandbox to have alf's latest & greatest mpi-genesis.
mod6: alright, im about to check your new ones here. i can confirm that the original 'mpi-genesis.vpatch' (f254bedf1e3241eb9de17232b630a0614f1cc54ff9c5407d87d79174e211833bcfc0135c89b4abcab2446acd93137a8e1b0798704ad7e4d498cc52c836c82c2b) gets dropped on the floor because of the addtional timestamps.
asciilifeform: http://www.loper-os.org/pub/mpi/mpi-genesis.vpatch http://www.loper-os.org/pub/mpi/mpi-genesis.vpatch.asciilifeform.sig for the l0gz. ☟︎
asciilifeform: grep "+++" mpi-genesis.vpatch | cut -f3 -d' ' | sort > happy.txt
asciilifeform: grep "+++" mpi-genesis.vdiff | cut -f5 -d' ' | sort > sad.txt
mircea_popescu: atm there's bot, lam-par (terrible name), fg, mpi and ffa.
diana_coman: basically this "sane-mpi" tar.gz: www.loper-os.org/pub/mpi/mpi-genesis.tar.gz
mod6: 'sane-mpi'
mod6: are we trying to get this one going then? http://www.loper-os.org/pub/mpi/sane-mpi.tar.gz
mod6: where is the de-facto mpi tarball?
diana_coman: and I ran it : v99 --wot .wot --seals .seals patches a mpi_second_cut.vpatch
asciilifeform: mpi-genesis.vpatch (asciilifeform)
asciilifeform: mpi_second_cut.vpatch (asciilifeform)
asciilifeform: ./v.pl a mpi_second_cut.vpatch
asciilifeform: ./v.pl d mpi-genesis.vpatch returns nothing
asciilifeform: still fails to recognize that mpi_second_cut.vpatch exists
asciilifeform: mv patches/mpi-genesis.vdiff patches/mpi-genesis.vpatch ; mv .seals/mpi-genesis.asciilifeform.sig .seals/mpi-genesis.vpatch.asciilifeform.sig
asciilifeform: my mpi genesis had 'vdiff'
asciilifeform: sha512(mpi_second_cut.vpatch.asciilifeform.sig) : 7b14150fd5100dc90f7130a491214fceda5984fdad20491487d45727c4be88885dc9d9245e7f1bee30fa236e1e774e03f6cba8f34ccc4508eaed0df53af329a0
asciilifeform: fwiw sha512(mpi_second_cut.vpatch) : 594052a750c3ab2ad16bbd73c578df9d99a98cc9811e6537e452dc2386b58c24555918e8d02e9e8d35c25808ac31a10babf0614e2d647afdf4f58b38059af118
asciilifeform: pretty strange, mine and phf's vtrons ate it up without complaint, e.g. http://btcbase.org/patches?patchset=mpi
diana_coman: but it seems to think there are no descendants on it (so v d mpi-genesis.vpatch returns empty)
diana_coman: tch: Error! Could not find vpatch "mpi_second_cut.vpatch" ; when I run v a mpi_second_cut.vpatch I get: Error! Could not find vpatch "mpi_second_cut.vpatch"
diana_coman: mod6 or asciilifeform, can you help? I'm trying to press asciilifeform's second patch for sane-mpi and V complains that it can't find the vpatch file although it is there (and I checked the sig too and it's all fine); I'm running your V, version thebitcoin.foundation/v/V-20170317.tar.gz.mod6 ; I have the following folder structure: http://p.bvulpes.com/pastes/1ZwgW/?raw=true; Here's the error when I run v press verbose . mpi_second_cut.vpa
asciilifeform: i dun have any notion that ' mircea_popescu oughta pay for the mpi ' , it was made without any such agreement. ditto the serpent archaeology etc.
asciilifeform: in the process of writing one on top of my mpi, i realized that the approach is unsound, and will eventually get people killed. so went straight to what became ffa.
asciilifeform: ( clearly marked in the src per se, likewise, e.g. http://btcbase.org/patches/mpi_second_cut#L5350 )
asciilifeform: incidentally the work was reported in s.nsa broadcasts. sanitized-mpi is a s.nsa product.
asciilifeform: i will stand and say, asciilifeform did exactly the Right Thing, prepared , with proper care wartime ersatz ( mpi ) for if ( as turned out to be the case ) proper item ( ffa ) takes years.
mircea_popescu: leaving the mpi_second_cut as the basic root for further/other work also.
mircea_popescu: http://btcbase.org/patches?patchset=mpi&search= << it's in there, that's what it is.
asciilifeform: will argue that for mpi asciilifeform did 'no work' ?
mircea_popescu: i expect it'll be patched into mpi. which she's going to also sign to genesis. sorta like how i signed mp-wp, and unlike how you didn't genesis it, or did you ?
ben_vulpes: perhaps shoulda said "needs a stan/mpi style gutting"
asciilifeform: mod6: trb ml was really not imho the proper place for it: mpi is not used in trb
asciilifeform: also on phf's http://btcbase.org/patches?patchset=mpi&search=
mod6: asciilifeform: if there are updates (that are complete & published) for MPI, you may want to update the BTC-Dev ML with the new changes, and reference the original: http://therealbitcoin.org/ml/btc-dev/2015-October/000175.html
asciilifeform: also helps that my mpi builds a standalone static lib
diana_coman: it uses the mpi part and quite substantially so works
mircea_popescu: diana_coman so what's your call, rather write as mpi branch or rather stand alone ?
diana_coman: mircea_popescu, we can do it yes; I guess the question is where to start i.e. no point in starting from koch that I can see; starting from asciilifeform 's sane-mpi would be one; adds and deletes stuff
mircea_popescu: http://btcbase.org/patches/mpi_second_cut#L5394 maximal lulz
phf: i added readme to mpi patchset also
mircea_popescu: http://btcbase.org/patches/mpi_second_cut << dat red...
mircea_popescu: anyway. my conclusion is ima do the eu-crypto as a new genesis, because really most of the koch crap in mpi (esp the prng crap) got dirtched
phf: http://btcbase.org/patches?patchset=mpi http://btcbase.org/patches/mpi_second_cut http://btcbase.org/patches/mpi_second_cut/tree/mpi/mpi-mul.c#L108 etc.
asciilifeform: prior to realizing that ffa is the troo path, asciilifeform actually planned to entirely re-do the mpi item
mircea_popescu: i meant, rather than make a genesis for eu-crypto, just make a branch of your mpi
mircea_popescu: you proposing better idea to branch it off mpi ?
asciilifeform: and it's a perfectly legit ( manually ground, from mpi, just like trb genesis was from 0.5.3 ) genesis.
phf: mpi-genesis.tar.gz is not a vpatch though
asciilifeform: your copy of mpi-genesis.tar.gz was lost, or wat
phf: but oftentimes when i post a patch something comes up anyway. like the recent mpi release by asciilifeform is a vpatch, but it lacks a genesis, which breaks all kinds of assumptions (e.g. the tree visualizer wouldn't work at all)
diana_coman: I looked through the relevant mpi part but I can't say I get it to that degree to be able to figure this out
asciilifeform: ( gpg-mpi's mod exp doesnt do anything special with composite exponents, that would automatically speed the op )
diana_coman: so, changed those 2 lines to pkey.n = mpi_copy(skey->n) and pkey.e = mpi_copy(skey->e) ; correct?
asciilifeform: whereas declared as void public_rsa(MPI output, MPI input, RSA_public_key *pkey ) and void secret_rsa(MPI output, MPI input, RSA_secret_key *skey )
diana_coman: hm, public rsa uses mpi_powm; secret rsa uses crc; still though 100 fold
a111: Logged on 2017-11-02 15:20 asciilifeform: also 1.1s seems like a pretty long time for a 4096b modexp on traditional mpi.
asciilifeform: now perhaps diana_coman replaced the thing with mpi_powm( output, input, skey->d, skey->n ) or equiv. -- but then speed of encrypt and decrypt ought to be ~equal~
asciilifeform: also 1.1s seems like a pretty long time for a 4096b modexp on traditional mpi. ☟︎
diana_coman: :) thanks asciilifeform ; and once again a big thank you for carving out that mpi part - it helped a LOT
asciilifeform: same goes for diana_coman & mpi, keccak, etc.
diana_coman: chugging along, collecting some data on the mpi-sane-based rsa stuff
phf: asciilifeform: i wanted to add your mpi to btcbase patches, but realized that it's not a valid vpatch, since it doesn't have a genesis. i'll add an exception to mechanisms (right now it just 404's and all kinds of broken, basically), but!
asciilifeform: or perl; or even mpi(c)
asciilifeform: makes a static, dependencyless 'mpi.a' .
asciilifeform: diana_coman: my mpi builds on all platforms, but on crapple requires a 1ln fix in the make
a111: 74 results for "mpi", http://btcbase.org/log-search?q=mpi
asciilifeform: !#s mpi
a111: Logged on 2017-04-04 15:46 asciilifeform: hey i deautoconfed gpg's mpi lib 100%
asciilifeform: hey i deautoconfed gpg's mpi lib 100% ☟︎
ben_vulpes: i guess i can --verbose and get the mpi values?
asciilifeform: we won't be welding gnu mpi onto anything permanent.
asciilifeform: re 'open sores', even a ~very~ small gadget, and in fact one that started life as a stand-alone library: mpi (bignum) piece of gpg 1.4, was quite astonishingly painful to properly saw off the kochball
asciilifeform: http://btcbase.org/log/2016-11-26#1573465 << there is inly a decrufted mpi (bignum) from it, it is on my www ☝︎
asciilifeform: note that i no longer intend to use mpi for anything. it was posted as a deautomakeification demo.
asciilifeform: http://www.loper-os.org/pub/mpi/sane-mpi.txt
asciilifeform: http://www.loper-os.org/pub/mpi/sane-mpi.tar.gz
asciilifeform: http://www.loper-os.org/pub/mpi/mpi_second_cut.vpatch.asciilifeform.sig
asciilifeform: http://www.loper-os.org/pub/mpi/mpi_second_cut.vpatch
asciilifeform: re mpi referred to in the automake thread, updates:
asciilifeform: as seen in my mpi item earlier.
a111: Logged on 2016-08-22 01:02 asciilifeform: automake is a work of evil, first thing i did when sawing off the mpi lib is to nuke it.
asciilifeform: automake is a work of evil, first thing i did when sawing off the mpi lib is to nuke it. ☟︎
asciilifeform: mpi/ChangeLog: not stored in w. Suggested by Florian Weimer.
asciilifeform: (does anyone find it interesting that modern gpg uses ~same mpi implementation as dark age pgp, but with massaged function names ('mp_' vs 'mpi_') ?
asciilifeform: because i consider mpi bugs to be a thing of interest.
asciilifeform: 'Someone please prove me wrong, but I think there is a bug in the function mp_modexp_crt (RSA decryption and signing) in PGP23a's MPI library. Attached to this message is a program which demonstrates the bug. While testing Magic Money for lingering bugs, the client gave the error "Coin from server has bad signature!" I tried again with different coins, and the program worked. The proto.dat file had been cleared as the coins were rea
mircea_popescu: there's a pile of incidental nonsense (such as min length for name but not password hurr ; such as 2.0.30 current doesn't even fucking compile, such as etc) that shouldn't disappear under "oh, mpi"
mircea_popescu: yeah, apparently mpi ints are messy in ram.
asciilifeform: mircea_popescu: all bignums in gpg are passed around as pointers to 'mpi' data structure.
mircea_popescu: so when feeding the prime generator, it doesn't just feed ap ointer, but unwraps the thing through mpi-something
mircea_popescu: ok. and this does not actually correspond to the numeric format as used by gpg's bignum because mpi reasons.
asciilifeform: mircea_popescu: mpi was written by braindamaged folk who believed that bigint ought to be able to 'grow infinitely'
asciilifeform: mircea_popescu: not stack. the raw mpi turd.
asciilifeform: mircea_popescu this is still a raw mpi hexdump