Fix truncate of table cells containing ANSI

[mikelorant]

Table cells (including the header) may contain ANSI codes. When determining the width of these cells, ANSI was not excluded.

The reflow package has the same truncate function that is aware of ANSI. This improves cell width calculation and correctly truncates.

[mikelorant]

@charmbracelet Can I get a review for this very small bug fix. Thanks.

[mikelorant]

@maaslalani Would you be able to do a review of this fix? Thanks.

[maaslalani]

Hey @mikelorant, this looks great! Do you mind adding a test for ANSI truncation so this doesn't regress again?

[mikelorant]

Take a look at the tests file and you'lll see why this is going to be a really big problem.

Options:

• Switching the tests to use golden files in a testdata directory.
• Accept confusing ANSI codes in the plain text multi line strings.
• Add a separate test specifically to test strings that use lipgloss.NewStyle(). This gets more difficult with tests are termenv detects no ANSI capability when running tests but can be worked around.

[maaslalani]

Take a look at the tests file and you'lll see why this is going to be a really big problem.

Options:

• Switching the tests to use golden files in a testdata directory.
• Accept confusing ANSI codes in the plain text multi line strings.
• Add a separate test specifically to test strings that use lipgloss.NewStyle(). This gets more difficult with tests are termenv detects no ANSI capability when running tests but can be worked around.

I would go with the 3rd option to test strings that have lipgloss.NewStyle or hand code ANSI sequences in the strings (with the help of a few constants from termenv). If we go with the lipgloss.NewStyle() we'll likely have to force the color profile to be set to ANSI256 in one of the tests so the ANSI sequences show up.

[mikelorant]

Had a little fun with the test by making the header an incredibly long ANSI string.

Have confirmed this test fails on the master branch without the fix.

❯ go test ./...
--- FAIL: TestTableANSI (0.00s)
    table_test.go:975: expected:

        ┌───────────┬────────┬──────┐
        │   Fruit   │ Color  │ Code │
        ├───────────┼────────┼──────┤
        │ Apple     │ Red    │ 31   │
        │ Lime      │ Green  │ 32   │
        │ Banana    │ Yellow │ 33   │
        │ Blueberry │ Blue   │ 34   │
        └───────────┴────────┴──────┘

        got:

        ┌───────────┬────────┬──────┐
        │   Fruit   │ Color  │   C│
        ├───────────┼────────┼──────┤
        │ Apple     │ Red    │ 3…   │
        │ Lime      │ Green  │ 3…   │
        │ Banana    │ Yellow │ 3…   │
        │ Blueberry │ Blue   │ 3…   │
        └───────────┴────────┴──────┘
FAIL
FAIL	github.com/charmbracelet/lipgloss/table	0.515s
FAIL

With the fix, test passes:

=== RUN   TestTableANSI
--- PASS: TestTableANSI (0.00s)

[maaslalani]

Nice work on this one @mikelorant, thank you so much!