Expected behaviour re outer table border with nonzero gutter?

[mjasperse]

Hi guys,
Just wanted to clarify whether it is intended that the exterior border of a table is fragmented by the gutter. This seems like a bug to me, but the test suites have results committed indicating it is correct:

e.g. test_outside_border_width()

e.g. test_table_with_gutter_and_padding_and_outer_border_width()

I've changed it in my local version, but if this is the intended result I will move the change to outside the library.

Cheers,
Martijn

[Lucas-C]

Hi @mjasperse and thank you for this report! 👍

To me this seems like a bug.
I looked at the existing test reference PDF files, and to me it would make more sense for them to have an "unbroken" outer border.

Pinging @RubendeBruin about this, as he contributed most of the padding implementation, and may have extra useful information to share regarding this 😊

But it'd be nice to see how you fixed that, so feel free to submit a PR!

[RubendeBruin]

When the outer border width is not specified then I think the border should be fragmented.

I don't think there was really a use case for the combination of outer border width and padding when I implemented it. Having it continuous would be more efficient.

[mjasperse]

Before I submit a PR, I want to check what you think of the following change to handle this combination of options:

This adds the gutter space inside the outer border, which is compatible with how HTML renders tables with nonzero border-spacing.

[RubendeBruin]

My preference for this season would be:

:-)

But for the rest of the year I think your proposal is a good one.
Just one remark/question: Is the padding supposed to be to the centerline of the border or to the edge of the border? In the picture you posted it appears to be the centerline:

[mjasperse]

@RubendeBruin thanks for the observation! I added a padding amount of outer_border_width/2 so that the spacing is more consistent as you pointed out.