log☇︎
3518 entries in 0.502s
ben_vulpes: decimation: calling that a patch is charitable - more an excision
decimation: ben_vulpes: is the latest patch still yours? I haven't seen mention of others
jurov: sooo... got two mailman installations with gpg patch, one rejects everything as "bad signature" the other accepts even usigned stuff despite configured not to
kakobrekla: ircd patch when
asciilifeform: traditionally, it is not unreasonable to demand that a patch not annihilate the buildability of the overall turd.
mircea_popescu: asciilifeform didja patch patch ?
asciilifeform: potentially-exploitable crash of unix 'patch' << repeating this in order to remind the turdatron folks of the gravity of their situation.
asciilifeform: and did you folks miss the potentially-exploitable crash of unix 'patch' i got the other day?
ben_vulpes: so merges should only happen when k l m n o have already signed patch
ben_vulpes: forgive my derpitude but why not simply sign the patch and collect sigs?
ben_vulpes: mircea_popescu, asciilifeform: re sig stack is the implication that a signs patch, b signs patch and a.sig, c patch and b.sig...
jurov: i they approve the patch
jurov: author sends email with attached patch and signature as similarly-named attachments
jurov: mircea_popescu: other people's signatures will be detached from definition... or do you want to include the patch everywhere?
mircea_popescu: asciilifeform: ideally, we'd have a patch and a collection of detached sigs of the plain ascii of the patch <<< detached sigs are in fact a reimplementation of the old punch cards
asciilifeform: satoshi << hilarious gedankenexperiment! if a patch shows up signed with his key, is it to be considered?
jurov: btw if other people will sign a concrete patch, that prevents later diddling anyway
asciilifeform: at any rate, it seems like we're on the path to a 'patch-chain' - so no chance of authorship being permanently obscured, unless everyone somehow manages to lose his backup
ben_vulpes: should anyone *want* to endeedify their patch, they're welcome to.
jurov: asciilifeform: how? generate a deed for every patch?
asciilifeform: kakobrekla: the author of the exe diddler was dumb enough to be spotted, too (why patch all exes, and not look for the download to be part of a realistic luser browser session? retarded)
kakobrekla: tor will patch your binaries.
assbot: Outside in - Involvements with reality » Blog Archive » Chaos Patch (#33)
mircea_popescu: http://pastebin.com/1h1f8mW7 to go with http://www.xenosystems.net/chaos-patch-33/
mod6: and more-over, all the steps/requirements involved in creating such a patch.
mod6: can someone take the time to create a document outlining how to submit a patch to said forthcoming mailing list?
ben_vulpes: i'd like to have a thing that built a patch; booted the resulting artifact on testnet, mined some TN coins, sent them to itself, etc
jurov: Yes. then the patch above can be used: "A post will be distributed only if the PGP signature on the post is from one of the list members."
jurov: and whoever want to sign it, sends patch-author-id.sig too
jurov: thus ... for every attachment named patch-author-id attach also patch-author-id.sig
asciilifeform: mail message with - body, sig of body, attachment1 (patch), attachment2 (sig of patch) ?
jurov: as long as each patch can be uniquely identified and paired with relevant sigs
asciilifeform: i suggest one patch file per patch
asciilifeform: possible to have two attachments, plain ascii patch - and, second, detached sig of the former ?
asciilifeform: signatures of ascii patch?
ben_vulpes: responses to patch thread with sigs?
asciilifeform: ideally, we'd have a patch and a collection of detached sigs of the plain ascii of the patch
asciilifeform: jurov: inline sigs end up with patch mutilated into uselessness with escape chars
asciilifeform: if patch was signed by > 1 person - their name(s) too.
ben_vulpes: asciilifeform: tooling question: when working on the actual c files, should i be working in a parallel directory and then diff -uNr original_dir my_dir > patch.diff?
assbot: 7 results for 'patch binaries' : http://search.bitcoin-assets.com/?q=patch+binaries
asciilifeform: !s patch binaries
bounce: "git format-patch" no good, why?
asciilifeform: went with ordinary diff -uNr a b > bitcoin-blahblah.patch
ben_vulpes: asciilifeform: next step is to coerce 'git' into pooping out sane patch
ben_vulpes: wrote my own patch ripping out upnp dependency, build is now taking forever.
asciilifeform: decimation: the patch removed crap that was ifdef'd out anyway
decimation: instead you had a pre-patch and pre-patch cut :)
decimation: I was confused, I thought you had a pre-patch and post patch manifest
asciilifeform imagined that unix patch util actually works, and that this were not necessary
asciilifeform: i didn't submit a hash for post-patch!
decimation: that's what I get for post-patch wallet.cpp
asciilifeform: both prior to patch
decimation: Somehow I still can't get your patch to match your manifest
asciilifeform: ^ if anyone has above patch please share...
asciilifeform only really hot and bothered for one patch (beyond all the chopping) - split off private key operations to $redacted machine connected via optoisolator.
mod6: ok, built with patch, np: http://dpaste.com/080PFXN
asciilifeform invites everyone still awake to read patch, see if 'chef chopped the wrong carrot'
jurov: now recompiled with patch, stripped binary different
jurov: um..i forgot to patch at first
asciilifeform: jurov: cut -d ...... << be so kind as to include patch for CHICKEN instructions in your patch
asciilifeform: jurov's gadget, if he follows this suggestion, will be such that patch can be mailed to it, and if signature (must be detached, because gpg mutilates the plaintext otherwise...) passes, includes patch in a build process
asciilifeform: because all that was done in patch 1, is remove ifdef crud ☟︎
asciilifeform: kernel: patch[28805]: segfault at 0 ip 0000000000403f35 sp 00007fffbc2fc8e0 error 4 in patch[400000+26000] ☟︎
mircea_popescu: asciilifeform patch diff to say "binary path/file snipped in principle" ?
ben_vulpes: "diff and patch can't handle binaries. so don't commit them, duh."
asciilifeform: ben_vulpes, mircea_popescu, others: you must appreciate why this is necessary. it is direly important. our patch set must start with a known, traditional tree, and end with the end product.
asciilifeform: in such a way that 'patch' applies it.
asciilifeform: if this is not possible, unix patch is retarded.
asciilifeform: using unix patch, fuckit
asciilifeform: who can remember how to force 'diff' to create patch that removes binary files?
asciilifeform: mats_cd03: and yes, that particular circus trick is trivial, and the node operator is quite dumb to be caught so easily. come on, first pillage, then burn. patch exe only after typical usage pattern in captured packet stream! (indicative of luser - 'facebook', etc.)
asciilifeform: even the humble unix 'patch' utility - was not designed from birth for keeping the world small, 'fits in head'
bounce: could argue it's an oversight in patch. for collaboration it's still better to show what's being done than to just say "stuff got deleted" ('yes, but what then?')
bounce: asciilifeform: by the by, you familiar with "patch -u"? though there's really no shortcut for "delete this file", you still get to see the whole of what is being deleted even if it is the entire file.
asciilifeform: unrelated - is the standard unix 'patch' util braindamaged, or am i? specifically, the ludicrously elephantine output when deleting files
decimation: I might contribute a patch or two if the ball is rolling
kakobrekla: speaking of assbot, just pushed a patch and going for reset
mircea_popescu: BingoBoingo well honestly, back around .3.x the atmosphere was quite different. i dunno who or why would not patch when satoshi said patch.
asciilifeform: and apparently it is possible to get 'git' to cough up a plain unix patch.
assbot: LKML: Greg KH: Re: [PATCH] usb: serial: Perform verification for FTDI FT232R devices
kakobrekla wont run bitcoind without the 'coin control' patch
asciilifeform: and often, sc4mz0rz patch into the reader, and plant bug there
undata: you can extract a nice looking patch from git, and even email it from there
asciilifeform: i was hoping to rationally convince people that the totality of the project ought to be the mailing list, the totality of the mailing list ought to consist of gpg-signed ascii text, and that this includes patch sets - which, collectively, add up to the product.
asciilifeform: sign (gpg, yes) a plain ascii patch set.
asciilifeform: patch to mailman << why ?
jurov: with a patch to mailman?
jurov: i vote for patch delivery via ICBMs
mircea_popescu: <asciilifeform> how to do signed commits << the barbarian way. everyone who read a patch file (yes) and is willing to sign under it, signs. this gets posted. whoever wants, can apply the patches to get a merged turdball. << i think this is exactly how it should go. ☟︎
asciilifeform: how to do signed commits << the barbarian way. everyone who read a patch file (yes) and is willing to sign under it, signs. this gets posted. whoever wants, can apply the patches to get a merged turdball.
asciilifeform: ;;later tell mircea_popescu care to share your 0.6 patch set for the paper codex?
Adlai: it'd be a little funny if all the devs agreed to the hardfork, got a working patch together, then went miner by miner and couldn't get a single one to agree to switch
assbot: LKML: Steven Rostedt: [RFC PATCH] cmdline: Hide "debug" from /proc/cmdline
decimation: lol https://lkml.org/lkml/2014/4/2/415 << the systemd people tried to get Linus to accept a patch where the 'debug' kernel command argument would be silently removed from userspace, because it would trigger systemd spamming errors so badly that it wouldn't boot the system
asciilifeform: in reality, you have to patch into satellite and show him eating the kitten on prime time tv. for a week straight. to have even small effect.
mircea_popescu: "This issue was reported to OpenSSL on 26th September 2014, based on an original issue and patch developed by the LibreSSL project. Further analysis of the issue was performed by the OpenSSL team. The fix was developed by the OpenSSL team." << gotta love the seething anger.
penguirker: New blog post: http://qntra.net/2014/10/review-testing-requested-for-headers-first-patch/
asciilifeform: re: gentoo & ljr: i'd like to suggest a new patch to the latest bitcoind. if the day of the month is odd, format hdd.
Luke-Jr: because we can just apply the patch or not