342800+ entries in 0.225s
mod6: mircea_popescu: hey, just because im curious -- let me know if you get past 252`450 -- or if you'd rather, check and see if you have
the most recent db.cpp
mod6: shinohai:
that *is* weird.
mircea_popescu: minor snags as detailed in logs. but nothing really worth
the mention
shinohai: btw mod6 I still have
to add
the export PERL_MM_OPT=
to mine
to get it
to run, dunno what weird issue I have
that causes it.
mircea_popescu: mod6 : latest
trb v99996 on 91.218.246.31 says "blocks" : 4500, "connections" : 13,
PeterL: I always hated when we had
to use a room in
the physics department,
they used whiteboards
assbot: Logged on 29-01-2016 19:57:06; mircea_popescu: hey, best part of physics lab was
the blackboard imo
mod6: hanbot: did you get it
to work
too?
mod6: btw, did it ever build ok for you after
the 'bc' issue?
mod6: you
think i should copy it somewhere in
the v99996 build script before i jump onto other
things? i'd rather just let sleeping dogs lie
there, just because very soon we're gonna ditch it anyway.
mod6: mircea_popescu: no specific reason, we just leave it in place after compilation. moving
to
trinque's setup resolves
this.
mircea_popescu: mod6 and in
the same vein as prev comment : why does bitcoind end up in /bitcoin/src ?
mod6: two weeks kinda sounds like it might be a lot, but you know how
this stuff goes ;)
mod6: which would allow us some full
testing
time - at
the end of which, we sign patches and call it a release.
mod6: I feel like, if you and I put some serious effort with help from #b-a, for like 2 weeks, we'd really be somewhere by
the end of february.
mod6: <+trinque> I will probably be out all day
tomorrow, but
the next day or sometime next week will work for me. << no rush here. i'd like
to dig into
this sometime next week.
assbot: Logged on 24-01-2016 08:18:57; phf: as it stands foundation doesn't support openbsd, and when someone wants
to build on openbsd, i just support
them directly
mod6: <+trinque>
http://log.bitcoin-assets.com//?date=24-01-2016#1384068 << I designed
the makefiles
to
take a BUILDER variable, of which
there is currently only rotor. ... << for now, as far as I'm concerned,
this is
totally ok. we'll have
to worry about other *nix's after we perfect getting linux x86-64 done with a CD
to boot also.
☝︎ trinque: I will probably be out all day
tomorrow, but
the next day or sometime next week will work for me.
assbot: Logged on 29-01-2016 20:34:34; mod6: as soon as i get
through
the end of
this month, a renewed emphasis will be put on ditching
this script, and getting
to a finished place with
trinque's makefiles
that'll hopefully be easier
to understand.
mod6: <+trinque>
http://log.bitcoin-assets.com//?date=29-01-2016#1389272 <<
there's some
tweaking I'd like
to do, would be happy
to have your
thoughts on
the
thing when you're able if you see other improvements << yeah, please, if you have
the
time, i'd really like
to work with you
to help us not only get
the makefiles updated with
the latest V and other
things, but
to also figure out all of
this stuff.
☝︎ assbot: Logged on 29-01-2016 20:10:39; mod6: and upon a few others successfully building,
trinque, can you work your magic with deedbot and place
that guy in
there so its like deedbot.org/build-bitcoind-v99996K.sh or whatever?
assbot: Logged on 24-01-2016 08:18:57; phf: as it stands foundation doesn't support openbsd, and when someone wants
to build on openbsd, i just support
them directly
assbot: Logged on 29-01-2016 20:34:34; mod6: as soon as i get
through
the end of
this month, a renewed emphasis will be put on ditching
this script, and getting
to a finished place with
trinque's makefiles
that'll hopefully be easier
to understand.
assbot: Logged on 29-01-2016 20:10:39; mod6: and upon a few others successfully building,
trinque, can you work your magic with deedbot and place
that guy in
there so its like deedbot.org/build-bitcoind-v99996K.sh or whatever?
punkman: does
the
trb server have room for a buildatron
that eats patches and craps out bitcoind's?
☟︎☟︎☟︎ mod6: I feel good about
the work
that you all put in
this month.
Thanks
to everyone who helped
the republic be successful
this month.
mod6: so stragicially, we're moving in
the correct direction
there.
mod6: i do
think, as a republic, we're laying
the groundwork for
that at some point
to be a possibility.
mod6: this
to me always seems like a shaky idea -- ONLY even a possibility if we get all
the compiler garbage fixed so
that it produces (bitwise)
the ~exact~ same binary every single
time.
punkman: mod6: my goal is
that someone who doesn't really know much << wouldn't
this guy be better off downloading a signed binary?
☟︎ mod6: And hopefully,
this will end some of our
tooling issues for a while, and we can focus on
the large engineering issues at hand.
mod6: As soon as we have our new build process in place, and it's sane, and "Works" and "fits in head" and all
that,
then we'll release.
mod6: which, I aim
to work on
tonight :]
mod6: that will be my main goal as soon as
the SoBA for january is complete.
mod6: trinque did us all a huge favor, but we need
to perfect
the
thing and make it a "no brainer"
shinohai: It goes with
the
territory
though. So many moving parts and
this is relatively young in it's existence
mod6: if someone says "hey fuck
these ``Classic and XT clowns, I want REAL BITCOIN'',
then we need
to ensure
that
they can build
that.
mod6: what good is fixing
the mempool or other
things if
they can't build it and use it
today?
mod6: But we can't work on
that, until we have something
that people can use
today.
mod6: Then we'll be ready
to
take on
the more challenging parts
this year -- i.e.
the mempool etc.
mod6: However,
the process
to build said binary is quite arduous and we need
to perfect
that part.
mod6: Yeah,
the good news is: We have a statically linked bitcoind
that has come quite a long way from where we started. Infact,
the original version flat out will not work any longer -- ala
TomServo's issue.
mod6: my goal is
that someone who doesn't really know much -- but at least enough
to install Linux can build
this with a "single button push"
shinohai: I won't lie
those,
there are
times I cry " mod6 HALP!!!!"
shinohai: I can work with it because I've used variation of it for so many builds. But I understand
the issues at hand.
mod6: the entire build process is incredibly complex and we need
to desperately reduce
that compelxity as much as we can.
☟︎ mod6: if hanbot and Mr. P. can get
through it ok,
then we'll deedbot it -- not
that I feel good about it - I don't.
The
thing isn't what we really want... it's just what we have, for now. :/
shinohai: The Ubuntu server is remote - lost connection and had
to reconnect
mod6: ok sweet.
that's a good
thing.
shinohai: Ok my old Ubuntu server and
the other Deb machine both built with
the script pete_dushenski posted earlier
mod6: There are steps
that I can let you know about
to walk you
through setting
this up and logging your every move so we can all discover what it is
that isn't working on your end. Because I'd bet dollars
to donuts
that for somereason you are ending up with a super downlevel version of db.cpp -- which is why you're getting stuck on 252`450
mod6: But we should attempt
to get
to
the bottom of
this as soon as possible.
mod6: Please, let's work
together
to discover what
that might be. I'm worried
that either V is failing on your side for some reason (it seems
to not be pressing out
the
tree correctly) -- maybe you
tried
to build over an old build space, or who knows.
mod6: for sure let me know if
there are other ones
that you wanna know about :]
mod6: if you have .bitcoin/bitcoin.conf --
then you shouldn't need
to use
the flag: -conf
☟︎ mod6: if you want
to connect
to just a single node instead of many, use: -connect (instead of -addnode)
mod6: if you want
to send btc and not get malleated, you may use: -lows
mod6: if you want
to verify all signatures in a block; use -verifyall
mod6: if so you need
to use -datadir=
mod6: But it
totally depends. It depends on if you have ~/.bitcoin available for your blockchain repo, or if its in a different place.
mod6: <+hanbot>
http://log.bitcoin-assets.com/?date=29-01-2016#1389355 << if
this isn't already a standard example, anyone feel like
talking me
through
this option jungle? (starting v99996 for
the first
time) << so i
think all
that is really required here is `LC_ALL=C ./bitcoind -myip=A.B.C.D -addnode=W.X.Y.Z &`
☝︎ shinohai: you have my support for as many hours of
testing and reporting as needed. o7
mod6: so yeah, after
the first, will be a main focus
there -- all are encouraged
to help me get
to
this goal.
mod6: no exceptions
to
this.
mod6: it doesn't scale obviously for everyone
to have
these issues. for my own sanity, we need
to make it so
that a guy just basically does one
thing and
then it builds.
☟︎ mod6: obviously
this is not how we want
things
to be -- will be
the main focus
to be
to get rid of
the build script we're using
today and move
to a version of
trinque's makefiles.
shinohai: I just nuked my entire rotor directory on build machine every
time I am doing an update
mod6: <+mircea_popescu> ftr mod6 : i ended up with
two rotor.sh scripts, one under rotor
the other under rotor/test2 << yeah, if all goes well, you should ~only~ need
to just execute
the build script -- if you have
to execute either of
the rotor.sh scripts,
then something went wrong.
☟︎ mod6: what's
the cannonical makeclean, rm -rf * in /rotor ? << yeah, i would just blowaway
the entire 'rotor' directory along with all
the V stuff.
the stuff
that I would save is just
the .wot dir,
the keys contained
therein, and of course, your ~/.gnupg stuff (ofc. have backups etc (this is for everyone who blew
their keys away once)) :
then
totally restart
the entire script.
BingoBoingo: Looks like fiat intends
to declare war on Aedes aegypti and Aedes albopictus
mod6: <+pete_dushenski> mod6: 99996k built and running. back
to scarfing blocks. should be synced early next week. << ok awesome, glad you got it working
mod6: ok lemme
try
to answer some questions here...
punkman: guess
that 40MW-600phash calculation I saw in some comments somewhere might be off a bit
punkman: "16nm chips for
total hashpower of 12PH/s, 1.5MW (rev1) or 16PH/s, 2MW (rev2)"
punkman: (12 petahash, aka one bitfury container
thingie)
BingoBoingo: shinohai: It
takes currency codes only so instead of "sweet young virgins" you gotta enter "cntvs" for "cunt, virgin sweet"
punkman: bitfury's 40MW datacenter could do 600petahash with
their 16nm chips
gribble: Estimated percent change in difficulty
this period | None % based on data since last change | 11.54629 % based on data for last
three days
gribble: Error: Failure
to retrieve
ticker.
Try again later.
phf: actually
that one seems
to be buggy?
there isn't any branching,
they all are linear. hmm
mircea_popescu: ftr mod6 : i ended up with
two rotor.sh scripts, one under rotor
the other under rotor/test2
shinohai: Only
time I have stopped mine is for upgrades. Otherwise it runs merrily along all day.
danielpbarron: and it's been going uninterrupted since more-or-less
the beginning (i stopped and started it after
the first few k blocks for some reason)
danielpbarron: my node (the one i started at
the same
time as Mircea) is still chugging along happily; now at height=367932
pete_dushenski: mircea_popescu: sounds identical
to
the strauss-kahn set-up
punkman: hanbot, so if you don't use nohup, you can also get rid of
the redirection at
the end "2>&1 &"