< michaelfolkson>
Is this worth putting in CONTRIBUTING.md? (the fanquake guidance)
< laanwj>
definitely (if there isn't anything like that in there yet)
< michaelfolkson>
OK happy to add this to my to do list (assuming fanquake doesn't want to open the PR)
< fanquake>
I'm not fussed. I'd like to think it's obvious that when you open a PR, you should take 2 minutes to summarize your change, including relevant / useful info, and save reviewers having to ask dumb things like "what is the bug", "how does it fix it" etc.
< laanwj>
i agree, but mentioning it explicitly in a document might be useful to point people to instead of having to type it out time after time
< michaelfolkson>
fanquake: Right but sometimes people struggle with what needs explaining to others when it is obvious to them. Guidance often helps as a reminder not to take shortcuts with explanations
< michaelfolkson>
I'll open the PR adding fanquake as a co-author then
< bitcoin-git>
[bitcoin] hebasto opened pull request #22124: 0.20: Update translations after closing 0.20.x on Transifex (0.20...210602-0.20-tr) https://github.com/bitcoin/bitcoin/pull/22124
< hebasto>
the Transifex.com labeled 33 languages with the "no translators" label. Those languages look abandoned as zero strings are translated for all versions. Suggesting to remove them to keep our repository tide. Or is this a topic for a weekly meeting?
< laanwj>
hebasto: the transifex import script is supposed to strip empty translations, doesn't it do this anymore?
< laanwj>
that would be the place to solve it, i think it's okay to leave the languages on transifex so that people who want to translate them can do so
< hebasto>
ah, right -- no empty translations in the repo, it works
< laanwj>
i tend to universally accept requests to add base languages (e.g. "nl"), but careful about adding dialects "nl_EN", too easy to come up with nonsense ones and i don't have time to investigate this
< hebasto>
sounds reasonable to me
< laanwj>
though i wish the language requests came with a description field so people can provide more information about the language they want to translate to, this would make it easier to recognize non-flippant requests
< laanwj>
unfortunately translations provide a fairly easy vandalism angle for people that want to mess with things, because it's inherently hard to check for language you don't know
< hebasto>
as usually in open-source, translationss are done and kept in a good state by people who care about it. If nobody cares, the result is unpredicted
< hebasto>
*translations
< laanwj>
right, it's just even more complicated by language barriers
< ufotofu>
are the build instructions for linux slightly out of date? the past couple of times i've attempted building have resulted in a libboost related error that was resolved by installing libboost-all-dev (on ubuntu).
< laanwj>
as far as i know they are up to date, which version of ubuntu are you using
< ufotofu>
not sure which package it needed specifically, out of the many libboost-all-dev installs
< ufotofu>
21.04
< laanwj>
libboost-all-dev is al extreme sledgehammer so we would like to avoid that
< ufotofu>
yeah understood
< ufotofu>
this occurred on ppc64le and now on riscv64 too. can't remember if i ran into it on x86_64.
< promag>
fanquake: yup my bad
< laanwj>
that might explain things, i suppose most people following the instrucitons will be using x86_64, ppc and risc-v are usually cross-compiled, anyhow let us know if you figure out which package is missing
< laanwj>
it might also be a bug in the platform-specific parts of the boost detection am scripting (vaguely remember that a architecture-specific path is injected somewhere)
< ufotofu>
pretty sure i ran into it on arm64 too so you're probably right
< laanwj>
thanks-that looks like boost_thread isn't linked correctly, which is weird, i guess this is for 0.21? fwiw the master branch no longer uses boost thread at all so might have solved this issue
< laanwj>
it might be that you followed the instructions on the master branch, which don't mention installing boost-thread-dev because it is no longer needed
< ufotofu>
aha ok that makes sense
< ufotofu>
ok sorry for wasting your time >_>
< hebasto>
np, happy your problem is solved
< laanwj>
no problem, i think it's a bug that configure doesn't report the missing library but it fails at link time, though probably not worth fixing anymore
< bitcoin-git>
bitcoin/master 883a5c7 fanquake: test: remove BasicTestingSetup from allocator unit tests
< bitcoin-git>
bitcoin/master 0559065 fanquake: test: remove BasicTestingSetup from amount unit tests
< bitcoin-git>
bitcoin/master 4c389ba fanquake: test: remove BasicTestingSetup from arith_uint256 unit tests
< bitcoin-git>
[bitcoin] laanwj merged pull request #22086: test: remove BasicTestingSetup from unit tests that don't need it (master...remove_unneeded_test_setup) https://github.com/bitcoin/bitcoin/pull/22086
< achow101>
hebasto: the best way would be to get them to be able to replicate the crash on testnet or regtest and then send us the offending wallet file
< bitcoin-git>
[bitcoin] hebasto opened pull request #22133: build, qt: Make QWindowsVistaStylePlugin available again (regression) (master...210602-style) https://github.com/bitcoin/bitcoin/pull/22133