fix: fix failing tests

[szevzol]

My first intention was to use your library in my production config handling, to take advantage of classes in the code when reading config data. (Thank you for your work!)

My issue was that I had "null" values in my yaml file and dataconf cannot handle them today as far as I've seen. However I removed those null values and decided to see how I can improve dataconf to handle null.

To be able to do that first I would like to fix failing unit tests.

I'm working on Win 10, using python 3.9.6.

[zifeo]

comma should not be required in the spec, likely the issue is with windows using \r\n and \n, can you try with that?

[szevzol]

Even if I try with \r\n, the result is the same:

FAILED tests/test_multi.py::TestMulti::test_complex - pyparsing.exceptions.ParseSyntaxException: , found '='  (at char 13), (line:2, col:3)

[zifeo]

Make sure you did not upgrade vyparsing by mistake: https://github.com/zifeo/dataconf/blob/main/poetry.lock#L449

[szevzol]

I have pyparsing 3.1.1.

I'm going to debug this test case in details hopefully tonight.

[szevzol]

Very strange. This fails:

assert multi.string("a { b = 1\nc = 2 }").string("a { c = 3\nd = 4 }").on( A ) == A(a=N(b=1, c=3, d=4))

But this, basically the same assertion, but passes:

assert multi.string("a { b = 1\nc = 2 }").string("a { d = 4\nc = 3 }").on( A ) == A(a=N(b=1, c=3, d=4))

Difference is that in the second string c and d swapped.

[zifeo]

there are a lot of weirdness in the latest pyhocon change (that bumped pyparsing): https://github.com/chimpler/pyhocon/issues?q=is%3Aissue+is%3Aopen+0.3.60, I would stick at pyparsing v2 and those issues should be gone

[szevzol]

I can confirm that my testcase is passing with pyhocon==0.3.59 and pyparsing==2.4.7. Should dataconf depend on these versions instead?

[zifeo]

Yes please and we can merge your PR 👍

[szevzol]

Fine! is it enough to update version in pyproject.toml?

[szevzol]

Updated also poetry.lock

[zifeo]

@szevzol thanks for the contribution and happy 2024