22300+ entries in 0.041s
mod6: so with "based on" ... am i undstanding that correctly that column is a link (pointer) to it's previous version?
mod6: Anyway, np. I'll work on it.
mod6: for instance, 'auto' is in there a bunch of times, but only one version obv. made it in.
mod6: Ok, I can certainly manage coming up with a list of everything that went into -RELEASE. And probably a handfull of others that didn't.
mod6: <+kakobrekla> lol buttfuckex << haha
mod6: <+jurov> mod6 sorry, that disadvantage of having staging evironment, posting dev links right and left...shoulda abolish that << no worries :]
mod6: i used to smoke a pipe a long time ago.. not sure if I could get used to e-pipe. pipe tobacco has a nice smell, etc.
mod6: asciilifeform: hahhah i never knew!
mod6: didn't i see Ben McGinnes just post something to the thread about MP & the keys on the gnupg list or w/e?
mod6: mine is like: eleafworld.com/product/detail.php/eleaf-istick.html -- the one with the digital display i had with me in b-a.
mod6: asciilifeform: that's just the tank, there is a 30W battery that it screws into from the bottom.
mod6: the airflow on this thing is beatutiful
mod6: BingoBoingo: 15mg nic
mod6: mircea_popescu: hahah
mod6: it's "sub ohm" 0.5ohm / Range 15-30W kangeronline.com/products/kanger-subtank-plus-clearomizer
☟︎ mod6: it's so much better than old one. i'll find a link...
mod6: just thought i'd share that :]
mod6: this new vaporizer tank i got is blowing my mind.
mod6: is see what you're saying, i did it about ~5 mins ago. no worries.
mod6: i threw mine in there again, it's 2nd in line or something "(queued, 2 moduli remaining...)"
mod6: danielpbarron: might be a good time to quick retest your key before the next batch
mod6: jurov: ahh, ok, so let's talk about the best way to do this when we get a chance.
mod6: not sure what jurov was referring to.
mod6: <+mircea_popescu> mod6 baby what happened there ? << not sure. anyway, even if I add in ".com" or w/e to that link jurov posted, still nada.
mod6: <+jurov> mod6, ben_vulpes, everyone:
http://bluesky/ml/test/patches.html needs to manually fill data for "Released in/Based on" columns - just version strings for every patch, help appreciated << this is 404 for me.
mod6: <+ascii_field> 'Is it possible that a keyserver running the old, buggy PKS code ... << this is gotta be trolling, right??
mod6: mike_c: hey! glad you like it :]
mod6: maybe its this is_prime function
mod6: ahh, i see, you gotta pick the bases for a randomly.
mod6: oh it might actually do this somewhere anyway... there are some references in the docs... although i haven't found it in the code yet.
mod6: lol, or why not just use Miller-Rabin instead?!
mod6: by the end of the night i was digging into prime selection. gnupg does fast fermat checks in several places, but im starting to wonder if it wouldn't also be benificial to just check against a list of "Carmichael numbers"
mod6: <+mircea_popescu> maybe i wanna use e = 2686977, whose business is it. << i was looking at this yesterday too, had the same impression. mpi_set_ui( e, 65537);
mod6: <+mircea_popescu> any idea what caused the oom ? if anything identifiable at all. << nothing specific yet. although the charts are interesting. Process Switches hit 4k before going to zero.
mod6: and 19 broken moduli nao
mod6: <+asciilifeform> mod6: and am i correct in my reading that thermonuke ver. never died ? << correct.
mod6: <+mircea_popescu> mod6 nice work on the perf tests btw. << thx!
mod6: there's a link at the top of that paste to where I found it.
mod6: ok, so gnupg v2 relies on libgcrypt to generate the keys, i finally found the code: dpaste.com/3106AJC.txt Seems that it does the same thing as v1 when doing the Euler totient.
mod6: ftp://ftp.ring.gr.jp/pub/net/gnupg/gnupg/
mod6: i see one here as old as 1.2.0
mod6: asciilifeform: naw, im searching for it.
mod6: BingoBoingo: yeah, probably was ported 2x since then would be my guess; first from CVS to SVN then from SVN to GIT
mod6: heheh, there are like even random '^L's in this file.
mod6: yikes! well, that confirms it, not that i was ever going to use gnupg v2 anyway, but now that i've looked at the code, CERTAINLY NOT.
mod6: holy shit gnupg v2 is bizzare! X_X
☟︎ mod6: so above there they do a swap of 'p' & 'q' to make p the smaller of the two. then they take the nlimbs of 'p' twice to allocate the space for each t1 & t2. which i thought was kinda strange.
mod6: seems, like it also exists at least in 1.4.12 too
mod6: gnupg v1.4.13's rsa.c yup
mod6: they're both declared with type MPI, was just curious i guess.
mod6: mpi_sub_ui( t2, q, 1 );
mod6: mpi_sub_ui( t1, p, 1 );
mod6: i see below the actual maff is being carried out in here:
mod6: take a look at this:
http://dpaste.com/0SQPBKC.txt Is there any reason when allocating the space for p & q to do Eulers totient they would initialize the space with 'p' and 'p', instead of 'p' & 'q'?
☟︎☟︎ mod6: asciilifeform: you're familiar with mpi right?
mod6: we'll be correlating all the data and putting together a bit of a press-release and notes along with it as soon as the v0.5.3 perf test is complete
mod6: <+asciilifeform> mod6: 3 ? <+mod6> Ok, the v0_5_3_1-RELEASE performance test is complete. Charts are posted here if you'd like to view:
http://thebitcoin.foundation/v0_5_3_1-RELEASE-Charts/ <+mod6> You'll notice that there are two times that during the full blockchain sync, bitcoind got oomkill'd. <+mod6> Again, this is just the bitcoin-v0_5_3_1-RELEASE -- no patches added. <+mod6> And for reference, here's where you can view the performance tes
mod6: <+asciilifeform> why would we expect a meaningful difference in performance between 1 and 2 ? << i don't, but we need a benchmark anyway so we can make some assertions
mod6: ok, so v0.5.3 is compiled and running -- im doing a nmon perf collection on that as well. then we'll have all three to compare.
mod6: <+mircea_popescu> mod6 as an item in a list. << hehe
mod6: "lists are for nerds."
mod6: then straight down to 0.
mod6: Process_Switches hits >4000 on v0_5_3_1-RELEASE-Charts when it oomkill'd
mod6: Again, this is just the bitcoin-v0_5_3_1-RELEASE -- no patches added.
mod6: You'll notice that there are two times that during the full blockchain sync, bitcoind got oomkill'd.
mod6: all you need is one trip to the post office
mod6: asciilifeform: thx for posting the werker
mod6: asciilifeform: heh, now these 21 guys are saying they wanna make embeddable mining chips for smart phones. glwt. seeing as how they run out of power in like 9 minutes as it is.
mod6: cannot access #bitcoin-assets: No such file or directory
mod6: there's ~20 hours you'll never get back
mod6: anyway, i have high hopes that some of these steps could even be automated to aliviate the pain of building this by hand.
mod6: ah, yeah. that makes sense.
mod6: and trinque, thanks for all your help :]
mod6: mircea_popescu: ahh, yeah. having gone through the gauntlet already on the thing at least a dozen times in aws, and also now having re-read the Gentoo-Handbook, i feel like I'm very close on this. To complete the guide, i wanna test out all of the steps on real hardware so the steps are accurate. I believe it'll entail something like creating a USB boot image, then booting off of that, then creating a stage3 in a chroot, then deploying that to