Report memory limit and allocated memory in longrepr

[petr-tik]

Make the longrepr more descriptive and enrich the xml output with this information.

Refactor limit_memory to return a _MemoryInfo type, which can be presented as a PytestSection and a longrepr string.

Remove item.nodeid from longrepr since longrepr is only used by junit-xml parser and the structure of the xml document makes the item.nodeid in the error message redundant.

Testing performed Add a test that parses the resulting xml file to make sure the expected string was written.

Issue number of the reported bug or feature request: #3

[gaborbernat]

@petr-tik can you try to resolve the merge conflicts?

[petr-tik]

@gaborbernat i have. please take a look when you can

[petr-tik]

Thanks for the detailed review with code style and refactor suggestions.

Addressed all of them. Only have 1 REVIEW comment about the name of the dataclass, happy to remove the comment, if you find the type name descriptive enough.

[petr-tik]

do you want me to edit the PR message before merging?

[gaborbernat]

If you have any amendments you think makes sense, sure 👍

[petr-tik]

i wanted to tidy it up and leave a descriptive commit message in main

[gaborbernat]

Go ahead with that then 👍

[petr-tik]

Done. I am guessing you are the one with squash and merge rights?

[pablogsal]

Thanks a lot @petr-tik for the fantastic work and for persuing getting this fixed ❤️

And thanks a lot @gaborbernat for the review! 💪

[petr-tik]

oh, my PR message wasn't included in the commit.

Thanks for the detailed review. Absolute pleasure