< bitcoin-git>
[bitcoin] MarcoFalke opened pull request #16737: test: Establish only one connection between nodes in rpc_invalidateblock (master...1908-testConnectOnce) https://github.com/bitcoin/bitcoin/pull/16737
< bitcoin-git>
[bitcoin] MarcoFalke opened pull request #16739: ci: Pass down $MAKEJOBS to test_runner.py, other improvements (master...1908-ciScratch) https://github.com/bitcoin/bitcoin/pull/16739
< emilengler>
Does SaveBlockToDisk works with buffers?
< sipa>
what are buffers?
< emilengler>
I added a usleep to it but it saves multiple blocks at once
< emilengler>
I added a 10 seconds usleep function at the end of SaveBlockToDisk
< emilengler>
But the log prints lots of them every 10 seconds
< emilengler>
Like not one UpdateTip per 10 seconds, around 50 UpdateTip per second
< sipa>
what is the code you added?
< emilengler>
Correction, I didn't used a usleep I used this_thread::sleep_for
< emilengler>
sipa: Yes in the log it takes pauses for around 10 seconds
< emilengler>
Before something else happens
< sipa>
emilengler: oh
< sipa>
block activation is not the same as block accepting
< sipa>
blocks are stored on disk when they're syntactically valid, and the headers of their parents are known etc
< sipa>
full validation only happens when that block becomes part of the best chain
< sipa>
due to out-of-order downloading that may be at a different point in time than when it was downloaded
< warren>
Checking if this is still true: since sha1 collisions are now possible signed git tags are unsafe since they sign only the sha1 githash, but a signed git commit is OK because it signs a lot more data?
< sipa>
warren: the hash used in a signature doesn't need to be collision resistant, only preimage resistant
< sipa>
i take that back
< sipa>
i'm not sure
< emilengler>
sipa: Sorry for the late response but what do you mean with out-of-order downloading? Or more a general question, where does the download actually happens or are there any good graphics for symbolization?
< sipa>
emilengler: net_processing mostly
< emilengler>
Thanks
< sipa>
we don't download blocks in order
< emilengler>
Why?
< sipa>
emilengler: bitcoin.stackexchange.com
< sipa>
search there, and if you don't find an answer ask it there; i'll answer it personally if needed
< emilengler>
FWIW, I got the answer. It downloads parallel for a fast-as-possible download if I got it correctly
< sipa>
harding: yeah, not much changed fundamentally
< jeremyrubin>
instagibbs: one that throws std::overflow_error would be *really* bad /s
< jeremyrubin>
instagibbs: for instance something that can't be distinguished from urandom('8') << 10, because then the tail 10 bits are always 0
< jeremyrubin>
instagibbs: so even though the top bits are solid, the bottom bits always being 0 makes it fail.
< bitcoin-git>
[bitcoin] promag opened pull request #16740: qa: Relax so that the subscriber is ready before publishing zmq messages (master...2019-08-slowjoiner) https://github.com/bitcoin/bitcoin/pull/16740
< meshcollider>
sipa: you mean second preimage right
< meshcollider>
warren: annotated tags have more info than just the commit ID, you must be thinking of lightweight tags