STY: Spaces in wrong place

[ShaharNaveh]

• closes #xxxx
• tests added / passed
• passes black pandas
• passes git diff upstream/master -u -- "*.py" | flake8 --diff
• whatsnew entry

[datapythonista]

looks good, couple of comments, but nothing important

[datapythonista]

Not a big deal, but I think I'd leave the \n at the end, seems clearer and more consistent with the spaces at the end

[datapythonista]

I think in this case it's clearer the other way round. I know we can't have exceptions if we want to validate it in the CI, but I'm wondering if we could just apply the rule for a single space, and leave those other cases to common sense.

[ShaharNaveh]

I know we can't have exceptions if we want to validate it in the CI

Actually I have thought of a way to do so (ish).

currently the script is validating if there is no space at the end of the line and a leading space at the begging of the next one, but what if it having an exception to check the number of leading spaces and according to that, the script will know if this is a style violation or an intentional.

If we say for example that if the string is starting with 3 spaces, it will consider it as intentional indentation, and will not report errors.

I am not very good with words so here is some test cases.

---

Will not pass:

foo = (
            "bar"
             " baz" # One space
)

foo = (
            "bar"
             "  baz" # Two spaces
)

---

Will pass:

foo = (
            "bar"
             "   baz" # Three spaces
)

---

Any thoughts?

[datapythonista]

I see your point, but doing something so complex is out of scope for pandas. I'd rather don't validate this, that making contributors have to know of such a thing.

I'd simply check the case where there are no spaces at the end of the first line, and there is just one space in the next. This will capture written paragraphs, where we should never have one space next to another, and will ignore everything else, which may include csv data and other stuff.

[datapythonista]

It feels like in this case the previous option is more natural. Since every line represents the content of a line if this was the content of a file.

So, for example, if we have a file:

XX
 X

It probably makes more sense to have:

data = (
    "XX\n"
    " X\n"
)

Than

data = (
    "XX\n "
    "X\n"
)

What do you think?

[ShaharNaveh]

Hmm I do get you're point.

so if I can check if the line is ending with a \n, it's OK for the next line to begin with a space? right?

correct me if I misunderstood

[datapythonista]

Yes, I think the rule could be that if one line finishes with \n, we don't really care what's in the next, it should always be ok.

But if you think it makes sense, you can revert these few lines for now, so we can merge this. And then we can see in more detail in the other PR (but I think it should just be that condition).

[ShaharNaveh]

But if you think it makes sense, you can revert these few lines for now, so we can merge this. And then we can see in more detail in the other PR (but I think it should just be that condition).

Sure!

[ShaharNaveh]

Yes, I think the rule could be that if one line finishes with \n, we don't really care what's in the next, it should always be ok.

Logic for this test case is now implemented at a4ec49e

[datapythonista]

Thanks @MomIsBestFriend , let's wait for the CI, but ready to be merged.

[TomAugspurger]

Thanks!