Line break issues in some cases

[ml1nk]

In situations where a line break is at the beginning inside a formatter like [[;;]\n] the line before the break gets ignored. For example A[[;;]\n]B becomes B without any sign of A.

The pull requests adds some tests to detect the problem an tries to correct it by adding/removing a char if necessary.

The problem is that $.terminal.iterate_formatting won't trigger an callback after the last bracket of a line when there is no character following. There are cases where it is correct (just remove 'array[i + 1][0] !== "["' to see the relevant cases fail with an empty line extra, but in our case it leads to ignoring a line.

By adding a char (for example " ") in some situations* we get the necessary callback. By removing the char and marking it as the last_iteraction it adds the line correctly.

There is probably a better way to fix it, but at least the origin of the bug is visible. If the fix isn`t working in all cases or a better way found, i will reduce the pull request to the tests only.

*

line[line.length - 1] === "]" && // ends without content
i < array.length - 1 && // there is a next entry (the line was separated with \n)
array[i + 1][0] !== "["  // \n was inside the formatting

[coveralls]

Coverage increased (+0.07%) to 79.387% when pulling 3a80675 on ml1nk:devel into 9e757de on jcubic:devel.

[jcubic]

I think it's better to fix iterate_formatting then patch usage if it, I kind of fixed the issue with this snippet:

                } else if (i === string.length - 1) {
                    callback({
                        count: count,
                        index: i,
                        formatting: formatting,
                        length: 0,
                        text: in_text,
                        space: space
                    });
                }

it fixed the issue but break tests because it create empty string for normal cases, will try to fix the issue or just use filter(Boolean) if I'll don't found better solution.

[jcubic]

ok, I've fixed the empty string issue, I've just removed last_bracket check from split_equal the code is much nicer now.

[ml1nk]

Thank you for the better bugfix, hopefully the PR saved you some time debugging.