< bitcoin-git>
bitcoin/master 68c280f Jon Atack: test, refactor: abstract the feature_nulldummy blockheight values
< bitcoin-git>
bitcoin/master ccd976d Jon Atack: test: use 327 fewer blocks in feature_nulldummy
< bitcoin-git>
bitcoin/master 6e22b52 MarcoFalke: Merge #21373: test: generate fewer blocks in feature_nulldummy to fix time...
< bitcoin-git>
[bitcoin] MarcoFalke merged pull request #21373: test: generate fewer blocks in feature_nulldummy to fix timeouts, speed up (master...fix_feature_nulldummy_test_timeout) https://github.com/bitcoin/bitcoin/pull/21373
< bitcoin-git>
[bitcoin] jnewbery closed pull request #21562: [net processing] Various tidying up of PeerManagerImpl ctor and dtor (master...2021-03-peer-manager-impl-ctor-dtor) https://github.com/bitcoin/bitcoin/pull/21562
< bitcoin-git>
[bitcoin] jnewbery reopened pull request #21562: [net processing] Various tidying up of PeerManagerImpl ctor and dtor (master...2021-03-peer-manager-impl-ctor-dtor) https://github.com/bitcoin/bitcoin/pull/21562
< gwillen>
how do "artifacts" work on Cirrus CI? Is there a link somewhere from a job to download them, like in Travis?
< gwillen>
I'm seeing something in the docs about configuring them in the cirrus.yml file, but I don't see any of those lines in the bitcoin one, and I don't immediately see any artifact links from the jobs.
< wumpus>
i think you're right we're not exporting anything from our CI jobs
< gwillen>
ahh okay, thanks wumpus
< wumpus>
any specific reason you'd want to access the results? what is built is not really like the release binaries, it's specific for testing
< gwillen>
I was thinking more about detailed logfiles than results, didn't Travis used to make those available?
< gwillen>
(in the interest of full disclosure, the actual problem I'm trying to solve is with Elements, but since we just merged the switch from Travis to Cirrus from Core, I'm just starting to understand how to work with Cirrus and I wanted to make sure I understand how Core is using it)
< wumpus>
the task link has the full logs afaik https://cirrus-ci.com/task/5065870943715328 (it's the "View more details on Cirrus CI" link at the bottom of the github details screen)
< hebasto>
MarcoFalke: but this is a workaround only, right?
< MarcoFalke>
hebasto: It restores the clang-11 behavior
< wumpus>
#endmeeting
< aj>
MarcoFalke: restores clang-11 behaviour or just ignores negative constraints entirely?
< MarcoFalke>
no idea. We simply can't use them as designed because by default they'll spit out too many warnings and they are too invasive to be useful
< aj>
*nod*
< sipa>
it seems to me that in properly designed code (which we certainly aren't close to...), negative locks shouldn't be needed except perhaps inside within-module code
< sipa>
at least on a boundary between modules the caller shouldn't need to know or care about what locks the callee is going to need
< aj>
sipa: they're kind of/theoretically useful for being sure a switch from RecursiveMutex to Mutex is safe
< sipa>
i guess you could have callbacks that leave and re-enter a module, while an internal lock is being held
< aj>
callbacks lose the annotations anyway though, generally
< sipa>
we do have runtime locks too
< sipa>
they're possibly more useful for dealing with this
< aj>
Assume( x->y == 3 ) doesn't work if x is GUARDED_BY something :(
< aj>
well, no need for the ->y i suppose
< jnewbery>
Not being able to Assume() anything that's GUARDED_BY a mutex makes me sad
< sipa>
why does that not work?
< aj>
sipa: Assert and Assume use a lambda function for the code they're checking, which then isn't annotated with the lock, so can't access GUARDED things
< sipa>
hmm
< sipa>
sad
< aj>
could probably have an AssertWithLock(L, E) that adds the annotation for L i suppose, but...
< luke-jr>
would have expected the implicit variable stuff for lambdas, to also imply locks
< sipa>
luke-jr: my understanding is that the law in florida can't take effect until there is a federal change... for some weird reason states can decide whether or not they adopt DST, but they can't decide their timezone or to always have DST (only never DST, or DST between specific dates)
< luke-jr>
I guess it's tantamount to saying FL is Central time
< luke-jr>
which moves the timezone border
< sipa>
right, but that too isn't possible without a change to federal law
< luke-jr>
but still, the people of FL could just ignore the fed decision, and use whatever time they want? :P
< luke-jr>
it's no different from saying office hours are 1 hour offset