log
500+ entries in 0.042s
a111: Logged on 2017-07-13 15:16 asciilifeform: the other thing to remember, is that the win from writing in ada - but not in ada in general, but the style demonstrated in ffa in particular -- remains even if YOU HAVE NO ACCESS TO GNAT and gotta compile by hand into asm. because it forces the style of algo that CAN be safely so expressed - i.e. without presumption of pointerolade arithmetic, gc, or other cost-externalizing electrosocialisms
asciilifeform: ( the thing weighs far too much for anyone to ever audit in the sense contemplated with e.g. ffa )
asciilifeform: e.g., ldd ./bin/ffa_calc returns
asciilifeform: i built a ch10 ffa with it, and the result runs.
asciilifeform: AAAAAAnd we have a winner! ave1's musltronic gnat builds static, musltronic ffa !!!!
mircea_popescu: fromdeedbot, you can also compile alf's ffa and do it all by hand :D
asciilifeform: ( and ffa correspondingly delay by a new month ... )
asciilifeform: thing is, i also have massively overdue item that i'm in the process of rolling out ( reignition of ffa series ! )
mod6: One other thing about ada-vtron, at the time, I was using system commands to execute `gnupg'. Where as now, perhaps we can use ffa/peh.
mod6: see, I did think that alf had produced something of this kind that was connected to ffa.
asciilifeform: and it worx for building e.g. ffa.
asciilifeform: ( of the compiler that you produced, that is, when it builds e.g. ffa )
asciilifeform: i'll point out, this is not equivalent to 'build ffa', the gpr file is part of ffa, and contains essential functionality ( mainly , incorporates the restriction flags, mandates staticity )
ave1: asciilifeform, that work has been done, it builds a native compiler with gcc/g++ and gnatmake but not with gprbuild yet (this last one can be build on the target machine). I can build the FFA code on an aarch64 machine with the resulting compiler.
asciilifeform: i'ma read comments/barfology laters; off to ffa room nao!
trinque: have you built your own trb node? used V? understood it? do you have a working gnat? built asciilifeform's ffa? built diana_coman's eucrypt? stood up a gentoo from scratch? fertile ground all over.
mircea_popescu: so, if your desert march results in a jewel of code, a la ffa, sure. if your desert march results in ample "lulz" as we call them, ie, intricate, unforgiving documentation of orcs' idiocy, sure.
spyked: http://btcbase.org/log/2018-04-12#1797815 <-- must confess that I am eager to read FFA spark.☝︎
asciilifeform: ( or see the ffa article series, http://www.loper-os.org/?cat=49 , currently on sabbatical but due to resume after i come back from upcoming biznistrip )
asciilifeform: no-dynamic-allocation is also a Good Thing, for instance in my FFA crypto lib ( http://www.loper-os.org/?cat=49 ) this property exists
a111: Logged on 2018-04-04 00:46 asciilifeform: phf: i was vaguely hoping he might grasp this by playing with pehbot / reading ffa ; but loox like no dice so far
asciilifeform: phf: i was vaguely hoping he might grasp this by playing with pehbot / reading ffa ; but loox like no dice so far
asciilifeform: btw, ftr : ffa comes back from sabbatical next wk !
mod6: which is the only version i've been able to stand up that seems to work 100% with ffa.
asciilifeform: hanbot, mircea_popescu , et al : observe in phf's http://btcbase.org/patches/vtools-vpatch/tree/vtools/vdiff.gpr , he did not forbid the linking of libc into the library ( as i did in http://btcbase.org/patches/ffa_ch1_genesis/tree/ffa/libffa/ffa.gpr#L46 ) . apparently this worx on some gnat, but on hanbot's -- barfs
mircea_popescu: we will evidently have a ffa-based, canonical gpg replacement. EVENTUALLY. until such an eventually, i don't feel so great recommending anyone gpg (or, heavens help us, http://btcbase.org/log/2018-01-22#1774477 -- just as i had to do, and recently). so a drop-in, eucrypt-based, "good enough" item is more than useful.☝︎
a111: Logged on 2018-03-08 14:38 ave1: I've started on http://btcbase.org/log/2018-02-28#1786498, in Ada code and now I'm debating if I shouldn't just move the code to C as the eucrypt interface is C. But then maybe eucrypt will move to FFA someday in the future and Ada will be a benifit. (Working on code with lot's of C calls in Ada is not nice...)
mircea_popescu: http://btcbase.org/log/2018-03-08#1787224 << i don't expect eucrypt will ever move to ffa. this is in no sense a disidence, or any negative comment on ffa whatsoever. they are intended and designed as very different usecase solutions -- note the speed differential incumbent. eucrypt works as a "good enough" item, it principally intends to support a game, and same-level crypto needs. it's consequently to be light, fast, and ~r☝︎
ave1: I've started on http://btcbase.org/log/2018-02-28#1786498, in Ada code and now I'm debating if I shouldn't just move the code to C as the eucrypt interface is C. But then maybe eucrypt will move to FFA someday in the future and Ada will be a benifit. (Working on code with lot's of C calls in Ada is not nice...)☝︎
deedbot: http://ave1.org/2018/on-ffa-chapter-2/ << ave1 - On FFA Chapter 2
deedbot: http://ave1.org/2018/on-ffa-chapter-1/ << ave1 - On FFA Chapter 1
asciilifeform: ada folx: re making ada strings out of the c variety : strlen(char *) is a potentially lethal op ( suppose the nullterminator is missing ) so it will never be called implicitly by ada. you gotta either call strlen deliberately on c side, and then form ( can be on stack , declare ... Foo : String(1 .. Length) ... , say, a la http://btcbase.org/patches/ffa_ch4_ffacalc#L53 ) a proper ada string and copy the cstring into it.
deedbot: http://thetarpit.org/posts/y04/06c-ffa-egypt-proof.html << The Tar Pit - Egyptian div and mul "work": a correctness proof for the arithmetic operations in FFA Chapter 5
a111: Logged on 2018-02-23 15:45 mod6: I was thinking lsat night about my version of V in Ada, and am using shellouts there for the gpg related things. even with an integrated FFA, still need to add in an integrated Keccac from s.mg - those two I can work around. Others might be harder than it sounds.
mod6: I was thinking lsat night about my version of V in Ada, and am using shellouts there for the gpg related things. even with an integrated FFA, still need to add in an integrated Keccac from s.mg - those two I can work around. Others might be harder than it sounds.
spyked: hey mod6, I just checked v99993 using ffa ch1-5 and also used it to test a press of phf's vtools and it works fine. NB: Debian systems are now pretty much broken and won't allow setting the default gpg to v1, so I had to manually replace the gpg calls in v.pl to use gpg1.
mod6: mircea_popescu: well, never the less, I'm still very much behind the utility of ffa
asciilifeform: mod6 tried his hand at 'tester of whether ffa really does constanttime modexp'; asciilifeform comments on 1 set of output. is all.
asciilifeform: after that you can e.g. time echo "??\`\`*[print 0x]##[ == 0x]#[ * 0x]#" | ./bin/ffa_calc 1048576 32 lotsaones.bin | tr -d '\n' | python
asciilifeform: echo ".0~#" | ./bin/ffa_calc 1048576 32 | xxd -r -p > lotsaones.bin
asciilifeform: but know that 'iron mul' has been re-introduced (will be made selectable soon enuff) and you can use ch10 ffa to diagnose 'evil' cpu
mod6: hmmmod6@localhost ~/ada/alf/ffa/ch10/ffa/ffacalc $ command -v tr
asciilifeform: time echo "??\`\`*[print 0x]##[ == 0x]#[ * 0x]#" | ./bin/ffa_calc 1048576 32 /dev/urandom | tr -d '\n' | python
asciilifeform: >> http://btcbase.org/patches/ffa_ch10_karatsuba << ty phf !
mircea_popescu: never you mind, you've got the ffa to do.
asciilifeform: hanbot: and indeed i'm still grunting out prev item (ffa ch10 ) in conveyor.
asciilifeform only nao finished the proggy for ffa ch10, a backbreaker. nao for the text...
gabriel_laddel: asciilifeform: unfortunately not. Am here because I finally have a nice setup where I won't get kicked regularly. While I have been reading some FFA and DC I do not have anything running at the moment
asciilifeform: hey gabriel_laddel -- other than substances... do anything noteworthy in recent times ? ever read ffa series ? interested in s.bisp ? read diana_coman's www ?
mircea_popescu: http://btcbase.org/log/2018-02-02#1779409 << this should prolly be standard preface of ada behaviour discussions ; ffa or non-ffa related. asciilifeform maybe you should add it as a comment you know ? at the top of the pile, "if anything's strange say @me in #trilema, please preface the saying with and-here's-my-grep-v"☝︎
asciilifeform: in other experiments, http://btcbase.org/log/2018-01-09#1767440 appears to hold true across all ffa chapters.☝︎
asciilifeform: and http://btcbase.org/patches/ffa_ch4_ffacalc/tree/ffa/ffacalc/os.ads
asciilifeform: the 'main is in ada; some imports from c' variant is also illustrated in ffa, in http://btcbase.org/patches/ffa_ch4_ffacalc/tree/ffa/ffacalc/cmdline.ads
asciilifeform: ( it is imho unlikely that anyone will demonstrate an rsatron with fewer moving parts than ffa-ch5 )
asciilifeform: diana_coman: in some applications, speed doesn't really matter almost at all. and in those, it is even now possible to, e.g., take ch5 ffa for rsatron.
diana_coman admits that is one of the things she really, really likes about ffa
asciilifeform: diana_coman: the observation re 'take as much speed as you actually need, but not more, if costs ugliness' is correct and is in fact the basis for ffa
asciilifeform: in re shifts/rotates ( as seen in http://btcbase.org/patches/ffa_ch1_genesis/tree/ffa/libffa/w_shifts.ads ) one nitpick : Shift_Left, etc. are not per se gnat-specific; they are in the standard. but for some perverse reason their ~invocation~ is implementation-specific, and requires this gnaticism.
a111: Logged on 2017-11-16 17:36 asciilifeform: the only way to detect overflow on risc-v is the algo i used in ffa.
a111: Logged on 2018-01-30 16:45 asciilifeform: !~later tell ave1 http://ave1.org/2018/on-ffa-chapter-1/#comment-15
asciilifeform: !~later tell ave1 http://ave1.org/2018/on-ffa-chapter-1/#comment-15
trinque: http://btcbase.org/log/2018-01-29#1777911 << not at the moment. when my stack's popped to ffa will try to do my homework on one of the g5s☝︎
asciilifeform: and when i say 'complete' i mean zero-libraries, a la ffa. kernel calls for i/o, and that's it.
asciilifeform: ( not to mention WBool, present in ch1, is constrained to arithmetize on 1bit. but indeed Word is smallest physical unit used, there is no byte-addressing in ffa )
asciilifeform: ave1: 'Word is the smallest unit used for the arithmetic in FFA' is not strictly true, there is a HalfWord ( see ch9 )
a111: Logged on 2018-01-29 15:46 ave1: I see you are already at ch9, I only just finished with ch1: http://ave1.org/2018/on-ffa-chapter-1/.
ave1: I see you are already at ch9, I only just finished with ch1: http://ave1.org/2018/on-ffa-chapter-1/.
asciilifeform: phf: why didja truncate to the ffa bitness ?
asciilifeform: and oh forfuckssake, apparently asciilifeform cannot change any of the '2017' to '2018' in ffa , because :
asciilifeform: fwiw ffa is endian-insensitive .
asciilifeform: ffa for instance is result of 'why not see what constant-time costs'
ben_vulpes: http://btcbase.org/log/2018-01-27#1777455 << great exercise, actually. type in each chapter's version of ffa☝︎
mod6: lol, i've not "formally" started it. I have however, looked through most of 1, 2, 3, and 4. Did the homework for ffa_calc. But again, not enough to say "I've done it. Lemme sign."
asciilifeform: but asciilifeform for one will read his vtron. and ffa homework answers.
asciilifeform: douchebag: alternatively, ( or best of all, in conjunction ) read the ffa series and do the exercises.
asciilifeform: just like if a d00d managed to sneak in and steal it, it ain't 'exploit for ffa' . etc
asciilifeform: BingoBoingo: let's develop your picture. because it still isn't a complete one . because e.g. knowing the magicpacket for that particular box's nic, is not 'exploit for ffa', it is for $nic.
mod6: Honestly, I loved the homework for ffa_calc. That was awesome.
asciilifeform: the other fortunate bit, at least re ffa, is that it is not complicated from the ada pov
mod6: <+asciilifeform> mod6: i dun particularly disagree with any of this, but the pov that 'vetting ffa' is a 1time thing, that can be done and then 'is done', imho is mistake : each user must read it ~himself~. << I basically just mean for me & ben.
asciilifeform: mod6: i dun particularly disagree with any of this, but the pov that 'vetting ffa' is a 1time thing, that can be done and then 'is done', imho is mistake : each user must read it ~himself~.
mod6: I'd like to see the Republic continue to expand the number of trb nodes we have available this year. There are activities currently afoot that immensly support that. Getting FFA vetted and used as a base lib really will help get the ball rolling for any proposed trbi.
mod6: Some of this is my fault, I've been trying to keep up here. Getting kinda swampped with a bunch of things at once. But! These are all good things. FFA, eucrypt, ada, vtron stuff, et. al.
asciilifeform: hey NoSatoshisHear -- you say you read the ffa series ? in what order of complexity does modular exponentiation run ?
NoSatoshisHear: ffa, some cool stuff, making a braindead c one just cause I can learn better. Sure wish I could pour some fast multiply hardware, but don't know enough prolly.
asciilifeform: what you can do, is either a) operate bitwise, ffa-style. b) mandate an endianism for the data format. and since you know the endianism of the machine you're building on, in particular build, you know whether the machine's is opposite or same as the data's.
mircea_popescu: but in practice, it turns out i get ffa and other things, because my friends love me even more than i love myself.
asciilifeform in the midst of fairly hefty conversion of '5000 log frags' to proper blogotronic text -- in re ffa . but ffa is in this respect substantially simpler item; it is not a mega-mystery how to do arithmetic.
mircea_popescu: we got fna, ffa and faff!
spyked: http://btcbase.org/log/2018-01-24#1774969 <-- nah, finished just this december and enrolled into ffa class just after. :D thus my earlier http://btcbase.org/log/2017-11-12#1736817 derping☝︎☝︎
mircea_popescu: and ffs is more fundamental math than << ffa, obviously.
asciilifeform: ( ffa , yes, is made to work ~ on people ~ primarily )
mircea_popescu: you are a terrible classifier, you know that ? seriously, ffa, 3rd grade ? what next, telekinesis, preschool ?
asciilifeform: in that sense yea. ffa however is not frege, it is moar al-kwarizmus
asciilifeform: not in ffa either, lol
asciilifeform: i found it interesting to consider, why winblowz '0days' exciting to 19yo n00b, but e.g. ffa -- not
a111: Logged on 2018-01-23 17:53 lobbes: http://btcbase.org/log/2018-01-23#1774737 << btw, ty ben_vulpes for writing this article. After I get archive .zip delivery up and running I will be taking time to get my castle in order. This will include: 1) learning 'v' 2) get a working trb 3) testing my damn fgs already 4) ffa chapter 1 (at least)
lobbes: http://btcbase.org/log/2018-01-23#1774737 << btw, ty ben_vulpes for writing this article. After I get archive .zip delivery up and running I will be taking time to get my castle in order. This will include: 1) learning 'v' 2) get a working trb 3) testing my damn fgs already 4) ffa chapter 1 (at least)☝︎
asciilifeform: observe, e.g., in ffa.
asciilifeform: ffa makes it for the first time both doable and appealing imho.