[andreynering]

• Ref: feat(table): v2 table bubble bubbles#772
• Once ready and merged, these fixes will be incorporated by the table bubble.
• Follow-up issue: [v2] Consider adjusting the YOffset setting to offset terminal rows instead of table cells/rows #532

fix(table): do not print final empty row when overflow

Before

After

fix(table): fix headers with custom vertical padding

Note that we're still always truncating headers to 1 line, but now we're respecting any custom vertical padding.

Before

After

Long Headers

fix(table): fix rendering for bordered cells

Before

After

fix(table): fix horizontal shrink with outline borders only

Before

After

fix(table): fix layout for tables with inner borders only

Before

After

feat(table): rework height rendering logic to fix related bugs

I suggest to checkout and take a look at the golden files at table/testdata/TestTableHeightShrink to understand the fixes done here.

In short, the behavior was unreliable before, in particular when BorderRow(true) was set or when cell had custom margin / padding.

[andreynering]

I know this line looks complex, but it's a result of extensive testing and debugging, so we're covered (even by unit tests).

I can consider separating it into variables if you think it'd help.

[andreynering]

I improved this slightly since, but still open to opinions if you have.

[andreynering]

This one ended up being a small regression due to being tricky.

In theory the intended behavior is debatable, but I'm open to work on restoring the previous behavior.

In short, the previous behavior was forcing the table to show more records than asked if yOffset is too high. (For example, if the user asked to offset 90 out of 100 total rows, we'll still show more then 10 records if we're space "height" available for it).

[andreynering]

OK, I was actually able to restore the previous behavior and also fix some edge cases (but not all).

[andreynering]

Hey @Broderick-Westrope,

Just an update from charmbracelet/bubbles#601 (review) to let you know that your tests helped us to identify many bugs on table rendering and I have fixed them here.

This will be ported over to the Bubbles' Table component because it'll soon use this under the hood.

[meowgorithm]

Incredible work, @andreynering. This LGTM. Just to have it documented, how is the table API changing in this PR?

[andreynering]

Just to have it documented, how is the table API changing in this PR?

These functions were exposed as public because they're needed on my Bubbles PR:

• func DataToMatrix(data Data) (rows [][]string)
• func (t *Table) GetData() Data
• func (t *Table) GetHeaders() []string
• func (t *Table) GetBorderTop() bool
• func (t *Table) GetBorderBottom() bool
• func (t *Table) GetBorderLeft() bool
• func (t *Table) GetBorderRight() bool
• func (t *Table) GetBorderHeader() bool
• func (t *Table) GetBorderColumn() bool
• func (t *Table) GetBorderRow() bool
• func (t *Table) GetHeight() int
• func (t *Table) GetYOffset() int
• func (t *Table) FirstVisibleRowIndex() int
• func (t *Table) LastVisibleRowIndex() int
• func (t *Table) VisibleRows() int

[bashbunni]

NICE. Confirmed all test results look good, code is way tidier, and all of the table examples run 👌

[bashbunni]

We should probably also document that headers don't wrap in the README(?)

[meowgorithm]

Good call. I’d document it in the GoDocs.