Get rid of settings merging code in util/system.cpp repeated 5 places,
inconsistently:
- ArgsManagerHelper::GetArg
- ArgsManagerHelper::GetNetBoolArg
- ArgsManager::GetArgs
- ArgsManager::IsArgNegated
- ArgsManager::GetUnsuitableSectionOnlyArgs
Having settings merging code separated from parsing simplifies parsing somewhat
(for example negated values can simply be represented as false values instead
of partially cleared or emply placeholder lists).
Having settings merge happen one place instead of 5 makes it easier to add new
settings sources and harder to introduce new inconsistencies in the way
settings are merged.
This commit does not change behavior in any way.
fae43a97ca test: Seed test RNG context for each test case, print seed (MarcoFalke)
Pull request description:
Debugging failing unit tests is hard if the failure is non-deterministic and the seed is not known.
Fix that by printing the seed and making it possible to set the seed from outside.
ACKs for top commit:
davereikher:
Tested ACK fae43a97ca
Tree-SHA512: 33d848dd1f4180d3664ecf60e9810c2a93590c05276b2c46b1e4fe6e376b45916a46b90c803bb602750ab666da3a05ce499e550024685a90b8cc38fab6667cb8
fa4c6fa9b1 doc: Add documentation for new test/lib (MarcoFalke)
faec28252c scripted-diff: test: Move setup_common to test library (MarcoFalke)
Pull request description:
Sorry for clickbait, this is only a move-only scripted-diff commit and one documentation commit.
Longer term, someone who knows something about build systems can make this an actual library. Motivation for this is that each module gets compiled for each target that includes it. For example, setup_common is compiled 27 times (for the fuzz suite) and another 3 times for the other tests (bench, unit test, gui)
ACKs for top commit:
practicalswift:
ACK fa4c6fa9b1 -- diff looks correct and Travis is happy
jonatack:
ACK fa4c6fa9b1 with the reserve that the commit messages (and PR description) contain the motivation for this change. Built, ran tests, light code review.
ryanofsky:
Code review ACK fa4c6fa9b1. I didn't realize `lib` was actually name of existing directory, not a new name. But in any case this looks good and nice to have one scripted diff instead of two.
Tree-SHA512: 2e176df90c60578276e4a6dc83ff57ff59d8e666ecf30c5ceacb8c326725da91baa4cac3dfa7a2e1605f58122a3e3e27e4938ff33e3a0ce7ea53afffebbf57a4
286f197704 Add util_ArgParsing test (Russell Yanofsky)
Pull request description:
ArgsManager test coverage for parsing of integer and boolean values is
currently very poor and doesn't give us a way of knowing whether changes to
ArgsManager may unintentionally break backwards compatibility, so this adds a
new test to catch regressions.
ACKs for top commit:
promag:
ACK 286f197, more surprising results 😱
laanwj:
ACK 286f197704
Tree-SHA512: 9e1db3ef87e55abbc280af60c088f35765a1f9e2ec20507ad0c1992027b875490016868dcb8cc287e6df279dd0e00f10550901af3de3d36287867249e0bd8207
ArgsManager test coverage for parsing of integer and boolean values is
currently very poor and doesn't give us a way of knowing whether changes to
ArgsManager may unintentionally break backwards compatibility, so this adds a
new test to catch regressions.
This was causing a lot of test cases not not be very meaningful because
multiple configuration options were combined into one line.
The changes in test output with this fix make sense and look like:
```diff
- testnet=1 regtest=1 || test
+ testnet=1 regtest=1 || error: Invalid combination of -regtest, -testnet and -chain. Can use at most one.
```
Issue was reported and debugged by
Wladimir J. van der Laan <laanwj@protonmail.com> in
https://github.com/bitcoin/bitcoin/pull/17385#issuecomment-550033222
Belts and suspenders: make sure outgoing log messages don't contain
potentially suspicious characters, such as terminal control codes.
This escapes control characters except newline ('\n') in C syntax.
It escapes instead of removes them to still allow for troubleshooting
issues where they accidentally end up in strings.
Remove testcase generating code from util_SettingsMerge so it can be reused in
new tests.
The hash value expected in util_SettingsMerge changes as a result of this, but
only because the testcases are generated in a different order, not because any
cases are added or removed. It is possible to verify this with:
SETTINGS_MERGE_TEST_OUT=new.txt test/test_bitcoin --run_test=util_tests/util_SettingsMerge
git checkout HEAD~1
make test/test_bitcoin
SETTINGS_MERGE_TEST_OUT=old.txt test/test_bitcoin --run_test=util_tests/util_SettingsMerge
diff -u <(sort old.txt) <(sort new.txt)
The new output is a little more readable, with simpler testcases sorted first.
Followup to #15869. Treat "-wallet" as the network-specific argument in test
instead of "-server", to make test output clearer and be more consistent with
bitcoind. Update embedded hash to match changed output from this.
