The result of the phpunit is following.

$ phpunit 
PHPUnit 5.6.0 by Sebastian Bergmann and contributors.

...................................F                              36 / 36 (100%)

Time: 60 ms, Memory: 4.00MB

There was 1 failure:

1) Test_Table::test_column_value_too_long_with_multibytes
Failed asserting that 142 matches expected 80.

/Users/miyauchi/www/php-cli-tools/tests/test-table.php:52

FAILURES!
Tests: 36, Assertions: 98, Failures: 1.

I wrote a sample test for investigating this problem.
I found that str_pad() isn't working as expected for multibyte text.
I am going to ask people for the solution of this problem.

	public function test_strpad() {
		$max = 'みなさんこんにちは!';
		$texts = array(
			'こんにちは',
			'Hello',
		);

		foreach ( $texts as $text ) {
			$new[] = str_pad( $text, mb_strwidth( $max ) );
		}

		// The text width has to be 19.
		$this->assertSame( mb_strwidth( $max ), mb_strwidth( $new[0] ) ); // Fail.
		$this->assertSame( mb_strwidth( $max ), mb_strwidth( $new[1] ) );
	}

Result:

1) Test_Table::test_strpad
Failed asserting that 14 is identical to 19.

👌 thanks @miya0001

Oh ...orz
Splitting the text by mb_substr() is not working as expected in this case.

mb_substr( 'こんにちは', 0, 2 ) => "こん"
mb_substr( 'Hello', 0, 2 )     => "He"

@miya0001 Did you test whether the problem is still present with the changes that @gitlost had done in #107 ?

I am trying with this master branch, I think it hasn't solved.

I saw the problem with following test.

	public function test_column_value_too_long_with_multibytes() {

		$constraint_width = 80;

		$table = new cli\Table;
		$renderer = new cli\Table\Ascii;
		$renderer->setConstraintWidth( $constraint_width );
		$table->setRenderer( $renderer );
		$table->setHeaders( array( 'Field', 'Value' ) );
		$table->addRow( array( 'この文章はダミーです。文字の大きさ、量、字間、行間等を確認するために入れています。この文章はダミーです。文字の大きさ、量、字間、行間等を確認するために入れています。この文章はダミーです。文字の大きさ、', 'こんにちは' ) );
		$table->addRow( array( 'Lorem Ipsum is simply dummy text of the printing and typesetting industry.', 'Hello' ) );

		$out = $table->getDisplayLines();
		print_r( $out );
		for ( $i = 0; $i < count( $out ); $i++ ) {
			$this->assertEquals( $constraint_width, mb_strwidth( $out[$i] ) + 1 );
		}
	}

I am trying to find a solution to split multibyte text as fixed length for now.

Hi @miya0001 , I see what's happening - my fault - Ascii::row() is passing the width of the column to safe_substr() which is doubled for East Asian chars.

I was working on fixes related to #106 but am trying to fix this now - I'm thinking that maybe add a flag to safe_substr() to indicate the $length is a width, which is what I'm trying in gitlost#3 (which I accidentally pushed here), but although it's working locally it's failing on Travis so it's not the full story - maybe you can look at it...

Ok gitlost#3 is now passing Travis but it's not a great solution - it doubly loads unicode/regex.php for a start.

I'll see if there's a better solution (probably have to refactor the eaw_regex stuff I think) or let us know if you have one - perhaps there should be a safe_str_split() function or similar that does all the grimy stuff Ascii::row() is now doing...

Hi @gitlost 😄

Hmm, table looks broken on your test result.
I couldn't found the reason that why test is passed...

I see str_pad is not working for multibyte text, so I wrote a simple function to solve this problem.

https://github.com/miya0001/php-cli-tools/blob/4339c5906bc0543cb14a3fcc88cec755cb0ab7fd/lib/cli/cli.php#L257-L261

Following is a test for it. It is passed.

https://github.com/miya0001/php-cli-tools/blob/4339c5906bc0543cb14a3fcc88cec755cb0ab7fd/tests/test-cli.php#L47-L53

Thanks!

Oh! Sorry, it is just a problem of the CSS!

It looks nice!
I want to learn how did you fix this problem. 😄

Thanks!

I close this PR. :)

Hi no problem - I was wondering about that and didn't realize it was the CSS.

The fix isn't very elegant and is slow at the moment - it breaks the string into UTF-8 characters into an array and then UTF-8 character by UTF-8 character exams it to see if it's an East Asian char and then adjusts the length by 2 or 1 accordingly - there must be a better way!

I couldn't found the solution to split multibyte text as fixed length, so your solution is only way for now.
I am going to write a blog that introduces your solution. 😄

I am happy to see solving this problem.
Thanks for great work!

Thank you @miya0001! - I'll keep thinking about it though (sleeping on things is usually a good approach!) and see if there's something nicer...