[Tests] Require exact match in assert_start_raises_init_eror (jnewbery)

[maflcko]

Extracted from #12379, because the changes are important on their own.

This allows for exact testing, since the match can be specified with a strict regex. Internal details (such as exact formatting of the error message) can still be fuzzed away by regex wildcards.

[laanwj]

I did a similar thing in #12302 initially and no one agreed :/

[maflcko]

Hmm, I didn't notice the initial version of #12302, otherwise it would have been a Concept ACK from me

[laanwj]

Yeah doesn't matter, utACK

[jnewbery]

Tested ACK 5812273

Should be tested on other platforms in case stderr is different. See #12379 (comment) for example

I did a similar thing in #12302 initially and no one agreed :/

I think everyone agreed, but I argued that we should do the minimal fix required to get the v0.16 rc building and passing the tests.

[maflcko]

@Sjors Do they still fail for you?

[jnewbery]

This is wrong. Should be elif not partial_match and re.fullmatch(...)

or nest the whole thing one level deeper:

if partial match:
    if re.search(...) is None:
        #assert
else:
    if re.fullmatch(...) is None:
        #assert

[maflcko]

Added a commit to make the tests pass on windows

[jnewbery]

Tested ACK fae1374. Works for me.

[Sjors]

@MarcoFalke on MacOS 10.13.3. wallet_encryption.py still fails with a timeout. wallet_multiwallet.py and wallet_multiwallet.py --usecli now pass.

Running the full suite now; will update this comment if something new fails.

[laanwj]

Tested ACK fae1374

@MarcoFalke on MacOS 10.13.3. wallet_encryption.py still fails with a timeout.

Huh - we don't use any remotely complex regexes, so I wonder how this can cause timeout.