log☇︎
2700+ entries in 0.24s
ascii_butugychag: at least until we have a smarter patch applicator
punkman: patch that wasn't in release will have to be rebased to be in next release
mod6: Thanks again for posting your report & corresponding patch. Great work.
trinque: could make him eat a genesis v-patch of himself
assbot: Logged on 28-12-2015 00:39:31; asciilifeform: in the sense that presently patch Q is said to have patch P as an antecedent if ~any file~ in Q was modified by P
mircea_popescu: http://log.bitcoin-assets.com/?date=28-12-2015#1354663 <<< genesis.patch. ☝︎
asciilifeform: ben_vulpes: your question is elementarily answerable by looking at the hashes. each file hash exists solely because it came out of another patch. apply recursively.
asciilifeform: BUT only because a patch can affect multiple files, can it have multiple ~immediate~ antecedents.
ben_vulpes: a patch may have multiple antecedents, correct?
asciilifeform: in the sense that presently patch Q is said to have patch P as an antecedent if ~any file~ in Q was modified by P ☟︎
ben_vulpes: 1) only press the head patch in question (this is how i wrote my press)
punkman: ben_vulpes: if you want to press a minimal version of a patch, you can recursively find all antecedents of a HEAD patch
ben_vulpes: rebooting an old thread http://log.bitcoin-assets.com/?date=14-11-2015#1323387: asciilifeform: since a patch has multiple antecedents, would an accurate description of the correct patch-selection algo for press be something along the lines of "given a patch, find its antecedents. presssed version must contain the descendants of all of those antecedents."? ☝︎
mod6: different patch though
mod6: it was this patch that overcame that issue: bitcoin-v0_5_3-db_config.6.vpatch
mod6: And this statement is not accurate: ``With the “rm_checkpoints” patch, the wedge issue at Block 252450 was overcome.''
mod6: read this section: [ Tabling of Checkpoints Patch ]:
mod6: ;;later tell pete_dushenski Although I posted this patch in December of '14, it was quickly tabled by February -- and I wrote about this in the SoBA: http://therealbitcoin.org/ml/btc-dev/2015-February/000041.html
punkman: asciilifeform: but original was 'debug sanity' and did 2 very distinct, separable things, in 1 patch, which imho is mistake << agreed, but most of the wx crap was related to debug
asciilifeform: but original was 'debug sanity' and did 2 very distinct, separable things, in 1 patch, which imho is mistake
assbot: Logged on 24-12-2015 20:19:11; asciilifeform: punkman: i like this patch except imho 1) patches ought to do One Thing 2) printf hardcoded to stdout is Wrong Thing
mod6: ok cool. my latest patch should do just that.
mod6: all: btw, if you do use the above test patch for V, please remember you may need to run a dos2unix on the dpaste [dpaste.com/12M99KB.txt]
asciilifeform: and then punkman doesn't end up writing a patch that needlessly conflicts (or vice versa)
asciilifeform: punkman: i like this patch except imho 1) patches ought to do One Thing 2) printf hardcoded to stdout is Wrong Thing ☟︎
mod6: If you add the above patch, you should end up with:
punkman: in the debug_sanity patch
asciilifeform: ;;later tell punkman was it you who submitted the patch that removes all remaining wxisms ? if so, would you consider rewriting as a vpatch ?
asciilifeform: must understand that 'patch' is simply one possible operation for a vtron
assbot: [BTC-dev] openbsd patch ... ( http://bit.ly/1kdP6EL )
mod6: that would have had to have been placed in the local pdir (patch dir) and sdir (seals dir) manually - it's not in http://thebitcoin.foundation/v/ yet
mircea_popescu: when one executes v-magic upon the genesis patch one must end up with a ready-to-go bitcoin.
mircea_popescu: why's it not in teh patch tree ;/
asciilifeform: (a seal, recall, is a signature of a particular patch)
BingoBoingo: pete_dushenski: As penance you should fire up a trb node to test the new version string patch
ben_vulpes: it's vastly more likely that someone's going to need to crap out a patch for eg openssl than replace it wholesale.
asciilifeform: ;;later tell jurov http://therealbitcoin.org/ml/btc-dev/attachments/20151219/mempool_dev2_a654caa4f28ed6f78906fef28060c2f46470f067.patch << why not vpatch ?
asciilifeform: the only kind of thing that is 'in own tree' is a patch that has no common ancestor with ~anything~ in the genesis tree
asciilifeform: just as if i had cut out a single line of a single file, and a subsequent patch relies on that line.
mod6: Scenario: Guy creates vpatch that utilizes irc.h (removed in vpatch asciilifeform_ver_no_5_4_and_irc_is_gone_and_now_must_give_ip.vpatch) must ensure that Guy's patch is on it's own tree?
mod6: <+asciilifeform> it would end up as an altogether other tree in the graph << so said patch to add documentation dir would be its own root?
mod6: oh, i see, so like if a patch further down the line (in time) remove a file that a specific sub tree relies upon, that those become removed?
asciilifeform: (a patch that uses a deleted file can be legit so long as it is not on the tree branch that contained the deletion)
asciilifeform: mod6: test it with my latest patch
asciilifeform: http://log.bitcoin-assets.com/?date=20-12-2015#1348260 << this is necessary (apparently i have to explain it) because WE DON'T HAVE a known-good 'patch' util ☝︎
mod6: if we wait until the end, we'll save some cycles -- and mathematically the hashes wouldn't come out right if somehow the files were diddled inbetween. but I can see some merit in checking after each patch is pressed.
asciilifeform: and incidentally BingoBoingo, should he make use of this patch, can easily set another
asciilifeform: BingoBoingo: if the latest patch is ratified by the latter, then yes
asciilifeform: this is a press of the programmable-versionstring patch and down.
BingoBoingo: ben_vulpes: LibreSSL node also runs with ~1GB of ram utilized at startup after last BDB locks patch because OpenBSD malloc something or other
shinohai: Also asciilifeform ty for Programmable Version Strings patch
mod6: <+ascii_field> https://bitnodes.21.co/nodes/?q=/therealbitcoin.org:0.9.99.99 << experimental programmable-versionstring patch is in test, and seen even here << cool!
ascii_field: https://bitnodes.21.co/nodes/?q=/therealbitcoin.org:0.9.99.99 << experimental programmable-versionstring patch is in test, and seen even here
pete_dushenski: youth, to the extent that they have no patch of anything to call their own, are inclined to jump aboard any train promising any sort of future
pete_dushenski: as well as "boomers" very strong desire to maintain, if not grow, the little patch of earth they've thus far claimed for themselves
ben_vulpes: "no listen, we're super inclusive. get yourself a gpg key, get in the wot, read the logs, submit a patch, and we'll talk about it."
mod6: Not a lot really remains left for v054. I'm in the process of getting all of the 3rd party deps, listed and then will sign and find a place for them on the website. Then I need to update my build script so it pulls and verifies all of that stuff from our own host. Beyond that, I just need to publish the v054-RELEASE patch I've been sitting on. Then Mr. Vulpes & I will need to sign all the vpatches and post 'em to the mailing list.
ascii_field: i had the motherfucking 'play with barbarians' patch
jurov: the patch adds -l dl to LIBS
mod6: if you get that igprof patch working jurov, let me know.
jurov: the jettison patch had it the other way
ascii_field: jurov, phf: i recommend getting that patch of mine which lets you request EXACT bytes of heap used total at ANY TIME
phf: jurov: the goal is to get bitcoind working with pogo's limited memory. the problem is that a running bitcoind grows in memory use as a result of normal operations. we know that some processes claim a lot of memory by design, like mempool, so first step is to get a reliable way of cleaning out mempool. ascii wrote that patch, but discovered that in practice zapmempool doesn't reduce memory use.
assbot: Logged on 19-08-2015 14:32:04; asciilifeform: thestringpuller: this patch does not work as described, on account of boost idiocy (removal of items from the hash does not invoke their destructures! believe)
phf: jurov: the overal goal though is to flush the mempool, but simply measuring the memory between zapmempool shows that the patch specifically doesn't do it. there's either additional source of leak, or there's a leak in mempool, or, and that's the most likely case, the patch in question doesn't touch al lthe places where mempool has data
assbot: Logged on 19-08-2015 14:32:04; asciilifeform: thestringpuller: this patch does not work as described, on account of boost idiocy (removal of items from the hash does not invoke their destructures! believe)
phf: i think maybe it's worthwhile to patch the logger so it doesn't report until the client actually does a read/write. if you do a port scan on yourself, will have same effect
asciilifeform: there is only one way to reliably broadcast 1) when a particular patch came to exist 2) what the present state of the art may be.
assbot: LKML: Alan Cox: Re: [PATCH] kdesu broken ... ( http://bit.ly/20IGXcV )
assbot: LKML: Linus Torvalds: Re: [PATCH] kdesu broken ... ( http://bit.ly/20IGXcT )
mod6: <+shinohai> mod6: have you had any sucess with bsd yet ? << yeah, OpenBSD, achieved a full sync a few months back. had to patch it by hand.
jurov: no patch needed since 0.8
jurov: ascii_field: electrum-server depends on bitcoin having "txindex" which 0.5.4 not sure if has, 0.7 had and the el. patch was not heavy
phf: luckily for you today only you can get a new and improved PATCH from me that will at least let you create context, since i only finished it like last night, i don't yet know if it will actually encrypt, but it might
asciilifeform: http://log.bitcoin-assets.com/?date=05-11-2015#1317131 << might want to patch trb's version thing instead ☝︎
asciilifeform: trinque: there was a patch. iirc it didn't make it in for some trivial reason
trinque: you recommend a separate proggy rather than patch to bitcoind?
asciilifeform: whole thing is a trivial example of 'can operate entirely as a small patch on a bitcoind'
assbot: diffutils.git - GNU diff and patch utilities ... ( http://bit.ly/1P36dGM )
ascii_field: davout: this kind of braindamage has been around for some years, yes. but until now i did not know that it has reached the point where you have to patch ~the printer itself~ rather than merely the eeprom in the cartridge.
assbot: Logged on 28-10-2015 16:56:05; ascii_field: funkenstein_: i recommend dpasting your patch for review before signing formally
ascii_field: funkenstein_: i recommend dpasting your patch for review before signing formally ☟︎
mod6: vdiff.sh a b > your_patch_name.vpatch
assbot: Logged on 28-10-2015 16:30:06; funkenstein_: so after a number of tests I was going to submit the fix on this patch, but after using V more often it has become clear that it is annoying to have patches made from different places in the directory
funkenstein_: of course V still can be used in either case to apply any patch you like made from anywhere, just a minor annoyance
funkenstein_: so after a number of tests I was going to submit the fix on this patch, but after using V more often it has become clear that it is annoying to have patches made from different places in the directory ☟︎
funkenstein_: in this instance, i assume a properly measured replacement of the thing is preferred to an additional patch to fix
funkenstein_: so if anybody wants a signed antipatch for my broken patch, let me know. otherwise, I assume the thing can be simply dropped.
mod6: yeah. that patch won't be added into the mirror. now if someone wants to try out his patch locally, they'll have to pull his vpatch, and add his key to the wot manually and then see what happens if they want to press out a branch that includes that vpatch.
asciilifeform: e.g. funkenstein's latest patch will need an antipatch
assbot: LKML: Matthew Garrett: [PATCH 04/12] x86: Lock down IO port access when trusted_kernel is true ... ( http://bit.ly/1GAWd4H )
asciilifeform: i had one when i was in school, and it worked great. but today - nope. because go and try to install freebsd 4.7 - where do you get the packages? gonna patch by hand ?
ascii_field: see his patch, http://therealbitcoin.org/ml/btc-dev/2015-October/000172.html
ascii_field: derived from the patch set.
ascii_field: (i write a patch, sign it, throw in the hopper, everyone who has my name ticked - sees a tree with that patch applied. but OLD LINKS MUST CONTINUE TO WORK)
ascii_field: jurov: thing is, every time a new patch comes to exist, this is a new version.
mod6: so with V, you can use a command called "press" that will patch up through a given HEAD.
ascii_field: (or they will turn into nonsense with every new patch)
ascii_field: all of this is because i wanted to show funkenstein why his patch is catastrophically mistaken
assbot: WordPress blogger patch foot-drag nag: You're tempting hackers • The Register ... ( http://bit.ly/1ZWhrAo )