< bitcoin-git>
bitcoin/master 067fba5 Wladimir J. van der Laan: devtools: Always use unabbreviated commit IDs in github-merge.py
< bitcoin-git>
bitcoin/master 7524376 fanquake: Merge #16201: devtools: Always use unabbreviated commit IDs in github-merg...
< bitcoin-git>
[bitcoin] fanquake merged pull request #16201: devtools: Always use unabbreviated commit IDs in github-merge.py (master...2019_06_ghmerge_unabbrev) https://github.com/bitcoin/bitcoin/pull/16201
< bitcoin-git>
[bitcoin] practicalswift opened pull request #16212: addrdb: Remove temporary files created in SerializeFileDB. Fixes non-determinism in unit tests. (master...SerializeFileDB-non-determinism) https://github.com/bitcoin/bitcoin/pull/16212
< fanquake>
Apparently 350'000 lines is too many for a gist
< wumpus>
I can imagine
< wumpus>
neither http nor browsers are really optimized for viewing huge text files
< wumpus>
this is all too clear with travis too
< fanquake>
yea, which is also why viewing a 10'000s line travis raw log in the brwoser also sucks
< wumpus>
^^ that
< jonasschnelli>
Currently testing semaphore 2 CI. A complete X86_64 depends build with make check takes 22mins (open source plan).
< jonasschnelli>
And it looks like the max execution time can be extended to "hours"
< wumpus>
jonasschnelli: wow
< jonasschnelli>
Ubuntu 18.04 as base system
< jonasschnelli>
They offer a 3.4GH 8 CPU system with 16GB ram (paid plan)
< jonasschnelli>
CLI tool to connect via SSH and debug the instance as well
< jonasschnelli>
nice cache tool (which I'm currently testing)
< jonasschnelli>
sigh: semaphore2 doesn't support public access to the build log (yet)
< jonasschnelli>
(or I haven't figured out how to enable it)
< jonasschnelli>
Anyone know why in appveyor function name var (C __func__) get a "W" appended?
< jonasschnelli>
there are missing essential features like public access to PR build logs
< promag>
achow101: ^
< gertjaap>
is there a make switch i should use to get more verbose warning while compiling? i got a comment on my PR that there's a compiler warning, but i'm not seeing it myself, even after doing make clean
< sipa>
gertjaap: not really, though clang is known for having more readable errors than gcc
< jonasschnelli>
gertjaap: maybe look at the travis output log
< sipa>
oh a missing warning; that's probably just due to different compiler version or so
< gertjaap>
i would enjoy catching these things locally before pushing and waiting for travis to tell me what's wrong.
< gertjaap>
i'm using gcc and it's version 7.4
< jonasschnelli>
gertjaap: sure. Though you can always enable travis on your personal fork/branch
< jonasschnelli>
MarcoFalke: what features are we missing with Travis? The semaphore founders are interested to know and it looks like that end of the year, they will have a competitive solution for our needs?
< jonasschnelli>
More recent os-images is probably one...
< MarcoFalke>
We run in docker, so os images don't matter
< MarcoFalke>
freebsd would be nice
< aseem>
Hi there, I submitted my first PR (https://github.com/bitcoin/bitcoin/pull/16207) but travis-ci tests are failing. Any pointers on how to go about debugging it? Especially on environments that I don’t have access to. Sometimes restarting them helps in other repos, not sure how to do that here or if that’ll even help?
< aseem>
(Full disclosure, this is a tiny change and I know the community frowns upon those. I’m not trying to get a PR merged for the sake of it. From what I can tell, it’s the easiest fix for the issue referenced?). Any feedback welcome.
< instagibbs>
aseem, depends on the issue, lately there have been a number of eroneous failures, let me look
< instagibbs>
"Error! Initial build successful, but not enough time remains to run later build stages and tests" caching issue, I'll restart the tests, hopefully passes this time now that it has cached
< aseem>
instagibbs: Thanks for a quick response. Will await results
< aseem>
Actually looks like they passed! :)
< sipa>
cfields, jnewbery, dongcarl: i've modified the permissions you had in bitcoin/bitcoin to be the new "triage" permission rather than "write"; let me know if this breaks anything
< bitcoin-git>
[bitcoin] darosior opened pull request #16217: getrawtransaction: inform about blockhash argument when lookup fails (master...getrawtransaction_error_blockhash) https://github.com/bitcoin/bitcoin/pull/16217