feat: use Intl.Segmenter as default wordSplit method

[canstand]

Description

#10604

Hi @asturur, I will try to make it happen.
The plan is divided into two steps:

1. Refactor the existing word segmentation method, and retain the separator in the returned array;
2. Introduce Intl.Segmenter as the default method, and use the previous method as fallback.

In Action

The first step of the refactoring is to change the regular expression into a capture group to retain the separator, and filter out the useless empty array items in the wordSplit return value.

wrapLine no longer needs to insert infix. To keep the results as consistent as possible with the previous ones, the single space at the beginning and end of the line is intentionally ignored.

There are two inconsistencies after the refactoring:

• Before: \t and \r are replaced with spaces;
• After: \t and \r are retained

• Before: one space is removed regardless of the number of spaces at the beginning and end of the line;
• After: if there is more than one space at the beginning and end of the line, it remains unchanged

Is it necessary to maintain complete consistency with the pre-refactoring state?
Edit: Just noticed that if it's not exactly the same, there will be issues with character counts etc.

[codesandbox]

Review or Edit in CodeSandbox

Open the branch in Web Editor • VS Code • Insiders

Open Preview

[github-actions]

Build Stats

file / KB (diff)	bundled	minified
fabric	914.681 (+0.019)	301.494 (+0.044)

[canstand]

Did a little research. The current version wraps lines, by default, each soft line break will swallow a whitespace character ([ \t\r]).

To maintain compatibility (consistent rendering results) and support languages ​​that do not break lines with spaces, it is necessary to remember whether spaces are removed at each line wrap.

I prefer to introduce a break change and adopt a new line wrapping method: pre-wrap.

I looked at the line breaking method of Inkscape, as shown below:

It behaves in accordance with CSS whitespace white-space: pre-wrap;:

• Spaces and tabs are preserved, and hang trailing whitespaces (not limited by the line box width).

Looking forward to feedback.

[stale]

This pull request has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs 😔. Thank you for your contributions.