SafeRepr.test_brokenReprIncludesID

[tom.prince]

In #5544, exarkun suggests that assertion made by twisted.python.test.test_reflectpy3.SafeRepr.test_brokenReprIncludesID is too weak.

It should probably assert about what exactly the string should look like, rather than just that it has a particular substring.

[hash14]

Stricter testing of id() in test_brokenReprIncludesID

[hash14]

Note that the patch assumes that className and formatter.name in twisted.python._reflectpy3._safeFormat will always be BrokenType and repr respectively. Not sure if this is a safe assumption, please advise...

[tom.prince]

Thanks for your contribution.

When calling safe_repr, the formatter is going to be repr and BrokenType is a name coming from the test-suite, so in this test, it will always be that.

1. There should be spaces around operators like %.
2. A comment on why only the first line of the repr is looked at would be helpful for people looking at the code in the future.

I'll fix these two things up, and then merge.

[tomprince]

(In [37111]) Branching to improve-SafeRepr.test_brokenReprIncludesID-6265.

[tomprince]

(In [37112]) Apply patch from hash14.

Refs #6265.

[tom.prince]

Also, there was a twistedchecker error: String formatting operations should always use a tuple for non-mapping values

[tomprince]

(In [37116]) Merge improve-SafeRepr.test_brokenReprIncludesID-6265: Assert about the actual format of the repr.

Author: hash14 Reviewers: tom.prince Fixes: #6265