17400+ entries in 0.03s
mod6: agreed. let's try to keep it just down to the tabs for now.
mod6: Yeah, sounds good. Thanks, want to get all of this done in a sane order and timely fashion.
mod6: (a vpatch to tie them all together)
mod6: But it probably would have been the last thing to go into the tree before a release patch.
mod6: Ok. I can get that done tonight yet.
mod6: Ok sounds good mircea_popescu. I'll stop my current work on the high/low patch. I'll spend this week getting the changes [.wot in the pwd, and mechanical post-patch hash checking] in V [v99996] released. Then I'll immediately begin work on a re-alignment of all the patches we currently are distributing.
mod6: unsigned char pchSig[10000];
mod6: not sure what to do abut this tho:
mod6: i suppose if we continue to use ECDSA_sign, we can continue to use: memcpy(&vchSig[0], pchSig, nSize);
mod6: I'm not an expert on the openssl internals. If you think I should not trust ECDSA_do_sign and reimplment with ECDSA_sign, I can.
☟︎ mod6: there was a discussion around this, and decided to leave them in as a result. will dig up...
mod6: so the deal there is that was just some debugging info -- that may be useful to someone at some point.
mod6: what are you driving at here?
mod6: i believe my implementation to be correct.
☟︎ mod6: i think there seems to be questions around why i have two booleans.
mod6: which is why i added that part.
mod6: and asciilifeform, keep in mind that this is a *must* fix. i think that is vastly important that we discuss this in channel.
mod6: added complexity will result in bugs.
mod6: polarbeard: keep in mind that i considered all of this over a week ago -- and decided against it. so yeah, i have put a lot of thought into this.
mod6: from a user standpoint, i feel like its more conusing anyway.
mod6: because it causes confusion to have -forceS=0 (low) and -forceS=1
mod6: right, i'd prefer not to use an int.
mod6: remember, i want a default state to *not* require a flag at all.
mod6: how would one set this singular boolean?
mod6: am i not using two booleans, polarbeard?
mod6: If we simply get rid of 'fHighS' and 'fLowS' and replace with say 'fForceS', then how can we differentiate between entering into the high or low if block?
mod6: I believe we need two flags.
mod6: so what we want, is three different ways: no force one way or another, force high, or force low. remember, i originally didn't want to do a force high at all. but since asciilifeform requested this, i'm entertaining this option.
☟︎ mod6: Or, we can halt everything else, and create a new genesis and re-align, re-vpatch and re-sign everything.
mod6: And then we can re-align post release.
☟︎ mod6: I'm prefectly happy to reconstruct the patch with spaces to stay in alignment with the current (albiet unwated) spacing scheme.
mod6: So the difference in '
http://dpaste.com/2EZJVXA.txt' is that the block of code in Sign() is spaced with tabs. And to Mr. P.'s point it should probably be all in a file or nothing. Otherwise its mis-matched. Some with spaces, some with tabs.
mod6: mircea_popescu, asciilifeform, ben_vulpes et. al., any further thoughts if we should clean up the code alignment (currently with spaces) post release and do a one-time clean up with tabs?
☟︎ mod6: Also: From now on, if you are not in my WoT, and you are a n00b here, any pm's will be ignored, and you will feed my ignore list.
mod6: <+mircea_popescu> yeah cuz crazy argentines, they're like excited sheep in the airport. << haha, it was a big crowd. You picked me out pretty quick.
mod6: Being honest and forthcoming about matters will always go much father with me than anything else. ftr.
mod6: To me, this all felt like subversion. I didn
mod6: For posterity: Do not do this. If you get a neg rating and you care about your wot ident, do be an adult and do as I said. Lot less problems and drama this way.
mod6: And I understand that part too Mr. P. Typically, this is what seperates us from 'dog'.
mod6: <+ben_vulpes> mod6, polarbeard, asciilifeform, mircea_popescu: y'all follow? << this is my 0.00000002: If you don't trust a man, no matter what his keyid is today, how can i ignore that and trust a new ident? The correct thing to do as a man, admit your mistake, and or ask for forgiveness and move on. Simple as that. We're all adults here.
mod6: heh, there was such a crowd at the exit of the security area, me and Mr. P. walked right by eachother I think.
mod6: ah, thats right. forgot you blew away your keys. but still, i gvae you ample time to resolve this with ben.
mod6: i asked him to resolve this, but this has gone too far now.
mod6: the dispute? i dunno. his ident was 'punindented
mod6: mircea_popescu: he was talking to me in pm when he decided not to reconcile with ben and created a new ident. i can't sign stuff from a guy who i know that my co-chair has neg-rated.
mod6: so lets get this resolved and perhaps we can move forward.
mod6: polarbeard: perhaps not. but wot & signing patches is based on trust. you must have a positive rating from members for patches to be accepted.
mod6: polarbeard is punindented, i very much appreciate his offer of help to work on the R.I. but this is a non-starter until ben and polarbeard/punindented reconcile their differences.
mod6: I think you need to ack your previous identity and make ammends. Thats the first step.
mod6: polarbeard: I apprecate your offer to work on the Reference Implementation, but before you get too far along you need to reconcile that -1 from ben.
mod6: anyway, yah, its not like it called malloc or whatever, its on the stack. i am retard.
mod6: This is from: Design & Implementation of the FreeBSD Operating System by McKusick
mod6: Such segments usually must be phyiscally contiguous in main memory and must begin at fixed addresses. We shall be concenred with only those systems that do not visibly segment their virtual address space."
mod6: maybe i was just remembering this: "Most machine architectures provide a contiguous virutal address space for processes. Some machine architectures, however, choose to partition visibly a process's virtual address space into regious termed segments [intel, 1984].
mod6: memory allocation is done by the os at runtime. when you ask for some bytes, there are kern mechinisms that handle this request, are there not?
mod6: ya, i think i was just confusing some kern shit that i'd read.
mod6: thanks for the links tho
mod6: i gotta read up on paging again i guess.
mod6: the only place i can find that is in util.cpp:ShrinkDebugFile()
mod6: <+asciilifeform> mod6: the principal obstacle is the lack of a computer. << oh you're saying a target arch here? or?
☟︎ mod6: which btw for testers is kinda nice -- you can make a really large volume and attach to which ever instance at will and keep your OS small, or swap it out easily without ever having to clobber your build/sandbox area
mod6: no, i use /mnt/btc-dev/.bitcoin since mine is in an attached volume.
mod6: 12G /mnt/btc-dev/.bitcoin/debug.log
mod6: # du -sh /mnt/btc-dev/.bitcoin/debug.log
mod6: i could be wrong here.
mod6: no it shouldnt crash unless there is no more ram left available in the entire computer.
mod6: it should ask the os for 200k of mem, this should be allocated in pages (usually 4k ea. iirc) it may not be contiguos
mod6: not if someone builds something that works, and does not break anything.
mod6: <+asciilifeform> we don't have a computer. << well, this goes back to the same thing. currently we're married to some linux varient to do (future) building of flashable universe
mod6: (im just spitballing here)
mod6: you buy a flash rom (from S.NSA? or from local place you trust??) then get V, build a flashable universe, flash the card, drop it into some device and then you can talk to it over some wire from your work station
mod6: i think it could be pretty neat, in the end.
mod6: eh, sorry: does the idea of a flashable universe built from rotor still make sense.
mod6: does anyone else think that this is *not* a decent consideration at least? even taking into consideration the issues we ran into with pogo?
mod6: yah, i still think this might be a resonable path. but lot of thought still must go into this whole thing. tonight was a great discussion on it tho.
mod6: asciilifeform: so are we basically, with rotor, able to not only build a static bitcoind, but couldn't we also build a flashable rom that contains said static binary - a flashable universe so to speak?
mod6: sure, the whole: 1, 2, inf. problem.