rebase

rebase

handle_header_range -> handle_block_headers

handle_header -> handle_block_header

We are not using libmultiprocess

Symbol check is now failing. Not sure why it previously didn't pick up libminiupnpc, but now it does.

  FAILED: src/qt/CMakeFiles/symbol-check-bitcoin-qt /bitcoin/distsrc-87b0a4640229-x86_64-linux-gnu/src/qt/CMakeFiles/symbol-check-bitcoin-qt
21:14:08   cd /bitcoin/distsrc-87b0a4640229-x86_64-linux-gnu/src/qt && /home/bitcoinabc/.guix-profile/bin/python3.10 /bitcoin/contrib/devtools/symbol-check.py /bitcoin/distsrc-87b0a4640229-x86_64-linux-gnu/src/qt/bitcoin-qt
21:14:08   /bitcoin/distsrc-87b0a4640229-x86_64-linux-gnu/src/qt/bitcoin-qt: libminiupnpc.so.17 is not in ALLOWED_LIBRARIES!
21:14:08   /bitcoin/distsrc-87b0a4640229-x86_64-linux-gnu/src/qt/bitcoin-qt: failed LIBRARY_DEPENDENCIES

rename API (headers -> block-headers, pb.Header -> pb.BlockHeaders)

fix client.py

use /block-header for the path, BlockHeader for the protobuf message and chronik_block_header.py for the functional test, for consistency with other endpoints and in case we need another type of header in the future
Discussed in private with Tobias.

hold cs_main while fetching all the headers instead of holding and releasing it for each individual header, to avoid getting an inconsistent chain

I'm not very familiar with the chronik plugin system, so just to clarify, is the CScript class now something that chronik users will be encouraged to use when writing plugins? If so, do we have a clear idea of how we will handle API deprecation in case we want to refactor the test framework in the future (how long in advance we will give users notice and maintain the legacy API, how we will log deprecation notifications)?

rebase and shorten chronik_bridge::ffi -> ffi

idiomaticity

The setup_scripts/chronik-client_websocket.py failure above seems unrelated to me. It still is a bit suspicious, as I haven't seen an unrelated chronik ws related failure in a while.

@bot chronik-client-integration-test

I'm wondering if the way each individual CBlockIndex is fetched with the cs_main lock could be an issue. I guess there could be more efficient ways to do it, but mostly I wonder if there is a risk of returning an inconsistent list of headers if a reorg happens between two locks.

sigh

rededuplicate boundary checks, keep them early in the function this time

revert deduplicate

This is weird. It looks intermittent

review

nit: use another constant for the number of blocks, make it smaller because it does not change the validity of the test

add a comment about boundary check

review

the change in how dnsseed.dump is written (int(rep.fGood) vs rep.reliabilityStatus == ReliabilityStatus::OK ? 1 : 0) is initially wrong, because it seems to me reliabilityStatus is not actually initialized before the second time we connect to the node. Still investigating.

tests

One comment nit/typo.

--unused include