log☇︎
900+ entries in 0.281s
asciilifeform: doc/ChangeLog: * gpg.sgml: Small typo fixes 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: i was referring to the gpg users.
asciilifeform: ;;later tell mircea_popescu http://www.loper-os.org/pub/20166313-pg1_2.png and http://www.loper-os.org/pub/20166313-pg2_2.png (depdfization of dorre and klebanov's report on gpg whitening idiocy)
asciilifeform: ;;later tell mircea_popescu https://threatpost.com/gpg-patches-18-year-old-libgcrypt-rng-bug/119984 << the hannobockization is ready!!!1111
asciilifeform: gpg 1.4.10
asciilifeform: i disentangled it from gpg last year, and posted result.
asciilifeform: mircea_popescu: all bignums in gpg are passed around as pointers to 'mpi' data structure.
asciilifeform: i hesitate to use word 'correct' in relation to something gpg does...
mircea_popescu: that gpg reads the correct number
mircea_popescu: ok. and this does not actually correspond to the numeric format as used by gpg's bignum because mpi reasons.
asciilifeform: this may be enough to blow away a good fraction of extant gpg pubkeys.
asciilifeform: i built it, refuses to run without 'gpg-agent'
mircea_popescu: ftr thee test is : if test "x$GPG_ERROR_LIBS" = "x"; then
asciilifeform: gpg-2.0.22 and libgcrypt 1.5.1
mircea_popescu: ~/gpg-2.0.30/libgpg-error-1.24/src$ ./gpg-error-config --prefix
asciilifeform: find it and run gpg-error-config --prefix.
asciilifeform: when you built gpg-error, it shat out a binary, gpg-error-config
mircea_popescu: asciilifeform :checking for GPG Error - version >= 1.13... no
mircea_popescu: from autogen : "--with-gpg-error-prefix=@SYSROOT@" << what format does that take if not fucking path
asciilifeform: you want this only when building gpg.
mircea_popescu: fucking nightmare. so the lib-error shit compiled, but obviously ./configure --with-libgpg-error-prefix="/home/mircea/gpg-2.0.30/libgpg-error-1.24/" does nothing, with or without quotes
mircea_popescu: ~/gpg-2.0.30/libgcrypt-1.7.3$ make
asciilifeform: you gotta build libgcrypt and then tell gpg's ./configure where to find it
mircea_popescu: asciilifeform does building gpg build libgcrypt auto ?
asciilifeform: again it isn't in gpg !
asciilifeform: just like in gpg 1.4.10.
asciilifeform: gpg 2.x uses gcrypt lib.
mircea_popescu: so basically... the best key produced by stock gpg is... wait for it... about 700 or so bits strong.
asciilifeform: if somebody wants to replicate on gpg 2.x, plox.
asciilifeform: aaah did i ever mention that gcc 5.x won't build gpg 1.4.x ??
mircea_popescu: situation : you go to make key with stock gpg, set it to 4096, ie 512 bytes. it makes you the sign key with 512 entropy bytes, then makes you the encrypt key wirth the remainder 68, and that's it.
asciilifeform: ;;later tell mircea_popescu http://qntra.net/2016/08/rng-whitening-bug-weakened-all-versions-of-gpg/#comment-67543
asciilifeform: 'exec summary' for mircea_popescu et al: all gpg keys ever generated have at most 2048 bits of effective entropy. ☟︎
asciilifeform: Framedragger: observe the 'neverhappening' of the ssh key pops, which have 0 to do with gpg policy.
asciilifeform: sorta how we're still using gpg.
asciilifeform: not with classical gpg format
asciilifeform: in gpg it is rsa sig of all the bits in the turd other than itself.
mircea_popescu: 3) as it's theoretically nonsensical, it feeds a bunch of idiotic yet unconscious expectations in the mind of the user. i am well convinced that the general "i'll pull myself by own breeches" attitude of "creative" people is both due to and resulting in the gpg selfsig
asciilifeform: well, the way it is done in gpg (rsa sig of sha1) is indeed retarded
asciilifeform: (no version of, e.g., gpg, appears to)
asciilifeform: iirc it was something like gpg --enarmor foo.gpg
asciilifeform: mircea_popescu: but current idea is, an instrumented gpg.
asciilifeform: incidentally gmp (what ~everybody, at least gpg users, use) is braindamaged in a certain important way
asciilifeform: quite like what, e.g., gpg does, but in ~batches~ ahead of time.
mircea_popescu: hm apparently that doesn;t exist in stock gpg. my bad.
mircea_popescu: and the pile itself could just be put through gpg --armor, afaik you don't HAVE TO encrypt. i think ?
mircea_popescu: i don't get it, why not just use gpg armor ? wtf uu
asciilifeform: GPG Key fingerprint = 58CA 54C7 6D53 942B 1756 01D3 44D5 214B 8276 4ED5
asciilifeform: that'd be more or less standard gpg mechanism.
asciilifeform wonders what the boecks will move on to now that 'snore, gpg won't import the keyz' is addressed
asciilifeform: 'though the pgp keys in question do not import into gpg, they can be imported in a couple javascript pgp-trons: https://www.igolder.com/PGP/encryption/ and http://ianpurton.com/online-pgp/ ...' ☟︎
asciilifeform: fromphuctor: http://btcbase.org/log/2016-08-03#1513678 << example of rsa key generation, as seen in gpg ☝︎
asciilifeform: mircea_popescu: my best hypothesis is a) khadeer generated key with, e.g., 'jihadcrypt' b) winblowz gpg with the memcpy from rng nopped out by ???
mircea_popescu: asciilifeform incidentally, "owned by whitening" is not altogether a bad theory wrt the null-entropy keys. ie, "they replaced rng with null-outputting one, never noticed because whitening". this, of course, doesn't explain why gpg would end up with null-generated keys, but whatevs.
mircea_popescu: and speaking of gpg deplorable state asciilifeform can you think of any possible reason the damned thing doesn't come a) bundled with ent and b) with ready implemented tests of local entropy while c) key generation is a subset of entropy testing in all cases ? ☟︎
mircea_popescu: mod6 ok but none of this resulted in "current gpg shitfest is acceptable"
asciilifeform: even that snipped from gpg
asciilifeform: mircea_popescu: http://wotpaste.cascadianhacker.com/pastes/b7ce7e35-60d3-4e97-85e2-1f8c87607d36 << relevant excerpt. gpg 1.4.10. ☟︎
mircea_popescu: one obviously hopes that gpg did. but then again the brutal teacher that is experience shows it's wiser to suspect everyone merely hoped someone else did it.
mircea_popescu: first, i will crack all your fingers in the piano. THEN you MAY be allowed to buy ancient laptop and learn to gpg, MAYBE. ☟︎
mircea_popescu: thanks fucking god gpg doesn't clearsign such nonsense.
mircea_popescu: similarly, apparently there's a difference between "owning" a "bitcoin gem" that exists chiefly as a succession of forum posts ; and "owning" a "physical bitcoin" that exists chiefly as a succession of gpg-signed promises.
mircea_popescu: and you've got A LOT of things built in, not just "antispam". you got admin hardening ; gpg comments ; auto-spam tagging, ddos protection, anti-trackback flood and etc. ☟︎
mircea_popescu: where "works" is defined as i can select and paste a gpg armored segment and gpg can process it.
mircea_popescu: shinohai http://trilema.com/wp-content/uploads/2016/07/wp-mptargz-shinohai1.gpg
mircea_popescu: shinohai http://trilema.com/wp-content/uploads/2016/07/wp-mptargz-shinohai.gpg
mircea_popescu: asciilifeform http://trilema.com/wp-content/uploads/2016/07/wp-mp-alftargz.gpg
mircea_popescu: phf http://trilema.com/wp-content/uploads/2016/07/wp-mptargz.gpg << it's all there ; run a grep over the files for "fill your own" to see what you need to add ; you'll also need a theme (goes in /wp-content ) and obviously a mysql db. ☟︎
asciilifeform: well no, roughly 95% of the code is related specifically to displaying gpg-format keys.
mircea_popescu: if gpg was intended as a sort of otr, "user creates subkeys forever", it's shockingly poorly implemented.
mircea_popescu: Framedragger i don't see much merit in the whole scheme. gpg does something stupid and then maybe salvages some edge of it. mmkay.
mircea_popescu: there's 10x to 100x more ssh-rsa than gpg-rsa. it's a moot point.
asciilifeform: phuctor is, among other things, specifically machine for illustrating the defects in gpg format
mircea_popescu: different packagings of THIS SAME OBJECT are not to be entertained. we control ther fucking horizontal, not "gpg" or "opensshit".
asciilifeform: understand, if i were to switch phuctor to storing IN this format, NO key it spits out will ever be eatable by gpg. ☟︎
asciilifeform: will 'burn down their farms' like gpg burned down microshit crypto.
mircea_popescu: alternatively you could explain to your boss that a) he's not terribly informed, and if he's interested in this sort of thing b) really should make gpg id and show up here with his q's.
mircea_popescu: mod6 aww, you didn't choke on the submit part did you ? just gpg --encrypt --armor -r bingo and put the text into dpaste or wotpaste.cascadianhacker.com
mircea_popescu: no gpg, no wot, no personal security like that nisman idiot, no independent voice -> jwz, you know it.
mircea_popescu: http://btcbase.org/log/2016-06-06#1477612 <<< gpg-shotgun marriage! :D ☝︎
asciilifeform: ^ extraction of gpg privkey via ~hour of microphone.
mircea_popescu: "my gpg works" does not excuse gpg.
mircea_popescu: (also not particularly useful in theory, but then again gpg is a pos.)
mircea_popescu: asciilifeform ftr, i have nfi what you're using, but most people here are using either aes256 or twofish, i would suspect. and if they do not - they should edit ~/.gnupg/gpg.conf and add personal-cipher-preferences AES256 TWOFISH right above the line that says personal-digest-preferences SHA512 ☟︎
mircea_popescu: Framedragger for rsa yes ; for gpg ... mno.
asciilifeform: gpg is 'healthiest horse at glue factory'
asciilifeform: Framedragger: read the logz, gpg and compatibles are retarded in 101 ways
asciilifeform: http://btcbase.org/log/2016-06-03#1476739 << the aes key used by everybody here when they gpg is precisely 32 bytes long. (not that this is something to be proud of.) ☝︎☟︎
asciilifeform: interestingly, not 1 of the last 3 phuctor pops, imports in any version of gpg i have.
asciilifeform: a modern gpg key
mircea_popescu: http://btcbase.org/log/2016-05-31#1474274 << deeply don't, gpg with sym key is liek... mmmkay. ☝︎
asciilifeform: gpg --recv-key cbe2d0bd
asciilifeform: has nfi that openssl was never in gpg ?
asciilifeform: yeah, just so HAPPENS that BUG makes fingeprint on old gpg of randint modulus key appear sAME.
asciilifeform: namely, 'openssl WAS NEVER IN GPG'
asciilifeform: that src tarball (latest) is ~6MB, and DOES NOT INCLUDE GPG SOURCE
ascii_butugychag: and gpg creates dupes when folks sign one another's keys, yes
mircea_popescu: jurov i see "gpg: armor header: Version: PGPy v0.4.0 gpg: packet(6) too short"
asciilifeform: iceblox: yes, but what version of gpg ?