23100+ entries in 0.034s
mod6: What I was hoping for is maybe a build script. Would that be muti-gig?
mod6: asciilifeform: can you make a repeatable build environment for The Foundation to use going forward? Something put together by you, blessed by you, so we're all building/testing/using something that "fits in head"?
mod6: like... say.. 3-4 weeks ago when we were testing stuff.
mod6: I also seem to recall, and maybe I should dig for it, someone else built for gentoo before anyone was discussing 32bit.
mod6: yeah, I believe I did. But I can't find a build log for it.
mod6: <+nubbins`> what *has* 0.5.3.1 worked on so far? debian, darwin, that it? << what's blowing my mind is why i was able to build just fine on aws Gentoo AMI. Unless it was packed full of stuff in there that we don't want to begin with. Which must be the case.
mod6: anyway, that's the deal. we need something that we can work with; one supported environment, blessed by alf.
mod6: i thought everyone wanted to create a common gentoo configuration.
mod6: are you talking to me?
mod6: yeah, what we need now is a build environment as discussed yesterday. or we really can't move forward.
mod6: it probably does not
mod6: /usr/lib/gcc/x86_64-pc-linux-gnu/4.8.3/../../../../x86_64-pc-linux-gnu/bin/ld: cannot find -lcrypto
mod6: /usr/lib/gcc/x86_64-pc-linux-gnu/4.8.3/../../../../x86_64-pc-linux-gnu/bin/ld: cannot find -lssl
mod6: yeah,it's weird. it seemed to puke on this:
mod6: there is another error beyond that one??
mod6: next part here... in your email, it seems that you apply your patch "asciilifeform-kills-integer-retardation.patch", and this doesnt seem to completely solve it...
mod6: so need to run in gdb and disassemble.
mod6: root@debian-test:~/release-test/bitcoin-v0_5_3_1/bitcoin/src# readelf -a bitcoind | grep -i SSL
mod6: yeah i get nothing from these: root@debian-test:~/release-test/bitcoin-v0_5_3_1/bitcoin/src# readelf -a bitcoind | grep -i ssl
mod6: did you see my paste?
mod6: asciilifeform: is there a definitive way to check the binary for which SSL is linked in?
mod6: my hunch is that this is not working as expected. I need alf to verify for me.
mod6: alf might know. i bet it's the wrong goddamn one
mod6: I don't know for sure though.
mod6: 'OpenSSL 1.0.1g 7' << is probably referring to the output seen here: OpenSSL 1.0.1g 7 Apr 2014
mod6: obviously, i don't have a clue as to what's going on there.
mod6: so if that's not working as expected, I (we) need help getting the makefile in order so that it does this correctly & the auto.sh script if that's part of the problem also.
mod6: my deb machine was installed in december, but maybe the system wide OpenSSL 1.0.1g was upgraded in feb. The version the system wide gives me is April 7 2014. The xxd output looks like its indicating the one from the system.
mod6: we can change that up. that's easy enough.
mod6: fair assessment. i thought the same thing at first, and even had all the libs on our site. then started thinking conversely about it.
mod6: so that's why that isthe way it is.
mod6: someone also mentioned that we (the bitcoin foundation) should host the openssl, bdb, and boost libs on our website for dl. i decided against that as i like someone to feel like they're getting a copy from the originator. (this is from yesterday)
mod6: (the virtualbox one, in particular.)
mod6: i can take a look at mine in xxd, just a moment.
mod6: <+mircea_popescu> mod6 i thought you tested it in vms ? << yup. one virtual box deb 6 vm, one aws deb6 AMI and a friend had a separate deb6 ami. my gentoo testing was on gentoo aws AMI also.
mod6: <+mircea_popescu> the idea is, for someone wanting to copy your environment, to be able to do so with minimal work << i think going forward this is the only thing we'll support/build from/work on, something that we agree is "good"
mod6: <+ascii_field> has any of you lot checked to see that the executable resulting from this build ACTUALL contains the openssl built by auto.sh ? << i never did, this could be a problem too
mod6: i guess it was doing something strange on my end, or maybe debian 6 has something strange it does?
mod6: <+mircea_popescu> mod6 no, actually, there's no way to have these issues INTO GCC!!! << here. what?
mod6: I don't know why it worked for me.
mod6: mp: im not sure I follow
mod6: i think, maybe im not the right guy for this. or at minimum, I need someone who can look over my shoulder on this. there's no way we should have these issues 5 months into a project.
mod6: no good answers there.
mod6: mircea_popescu: nope. none. my deb6 environments (yes more than one) all compiled it fine with gcc 4.4.5. i didn't have problems on gentoo either fwiw. but not sure what the gentoo AMI had for a compiler. my guess is that I was fooled by gcc.
mod6: so re: all the uint32_t, sorry that you're all having problems with this. I guess I missed a lot of stuff.
mod6: asciilifeform: any luck with that error?
mod6: ah good deal, thanks.
mod6: someone should publish a "how-to-build-non-derp-gentoo" receipe
mod6: i've used gentoo exactly 2x: for testing bitcoind hah
mod6: oh. ok, didn't even know that.
mod6: (is it the only hold out left?)
mod6: what if gentoo adopts systemd next week?
mod6: what about all the "linux is dead" memes?
mod6: You think that gentoo would be better than openbsd?
mod6: BingoBoingo: all of this was very recently tested on gentoo. both 64 & 32 bit.
mod6: something downlevel anyway
mod6: the build im trying to work through on openbsd uses like (I can't remember off the top of my head) 4.2.x or something
mod6: most of our testing we were successful with seemed to be gcc 4.4.5
mod6: asciilifeform: huh. im not sure what'd that be. i did a lot of testing on amd64, ever saw that error. i did see something similar just now though on openbsd, but i suspect that something totally different because, well, openbsd.
mod6: anyway, food for thought.
mod6: iirc ascii mentioned it, said he'd maintain the pogo project. it's not a horrible idea.
mod6: these projects are diverging.
mod6: we really need two branches of code now i think.
mod6: if you wanna do that cross compiling stuff, you gotta use ascii's original one.
mod6: ok, so remember, version that we have for auto.sh (v0.0.5 & v0.0.5-32) are NOT to be used for cross compiling.
mod6: so i got everything to compile on obsd, but getting segfaults when trying to run the static binary. core dump looks like its something with threading...
mod6: there is a version number in the top of the auto.sh script, which version are you using?
mod6: many of us were able to build, not just me this time ;)
mod6: yes, was tested on a hand full of environments ( as stated in the ADDRESS ): deb6 (64/32), gentoo (64/32), ubuntu 12/14 (64) and apparently macosx too. not sure what you're running into.
mod6: <+ascii_field> ben_vulpes, mod6, et al: auto.sh leads to 'util.h:650:8: error: ‘uint32_t’ does not name a type' on my boxes. << what os/arch?
mod6: one year i got gotten iirc
mod6: did he fake me out last year?
mod6: <+BingoBoingo> kakobrekla: Best log error evar << that is pretty good hehe
mod6: good deal. im excited about a hardened version of bitcoin to run on obsd!
mod6: BingoBoingo: cool. ive got openssl/bdb/boost built now on obsd 5.6, then just gotta apply sublte obsd changes to v0.5.3.1 and then try the static build. if all works & pulls blocks, I'll make patches: one for the v0.5.3.1 source and one for `auto.sh' (which needs a few tweaks). maybe you can just point at that instead. anyway, no worries.
mod6: line 858 of wallet.cpp from v0.5.3.1
mod6: BingoBoingo: thank's for the OpenBSD reminder about arc4random, a note about your blog though... you do say 7.x or 8 or w/e.. in v0.5.3/v0.5.3.1 the function that contains that part to change from rand() to arc4random() is called: CWallet::SelectCoinsMinConf
mod6: sweet, thanks nubbins`
mod6: i'll see if i can do some sorcery in the next few days and get it resolved.
mod6: There's a lot of stuff on our "to-do" list, but it'd be nice to knock out some OpenBSD support before we get into doing other heavy lifting.
mod6: but all of that is easy, i just need to dig in to figure out how to get the proper configure args for openbsd -- iirc last time I tried, I tried to build a statically linked binary and then openssl compile failed. and I just got busy with the release so I havent gotten back to it yet.
mod6: so it looks like i've got 1 small tweak to the makefile, and for some reason it doesnt seem to like "DB_LOG_AUTO_REMOVE" in db.cpp and there are a few include statements needed.
mod6: naw. its just a few additional include statements in a few of the source files iirc
mod6: Hopefully I'll have something figured out for that soon. And yup, updates will go to the mailing list. :]