[Console]Improve formatter for double-width character #10714
Conversation
`$formatter->formatBlock()` formats a broken block when some double-width character is set.
Some character has doubled-width so `mb_strlen` does not return actual string width (just returns count of chars). By using `mb_strwidth` (http://www.php.net/manual/en/function.mb-strwidth.php), it can property measure the actual width of string.
When `Application` renderes Exception, it also writes broken block when message contains double-width (more generally, multi-byted) character. This commit fixes above.
|
I found that rendering error uses logic independent of |
| // str_split is not suitable for multi-byte characters, we should use preg_split to get char array properly. | ||
| // additionally, array_slice() is not enough as some character has doubled width. | ||
| // we need a function to split string not by character count but by string width | ||
| $str_split_width = function ($string, $width) { |
There was a problem hiding this comment.
please make it a private method rather than a closure which need to be recreated by PHP for each call to renderException
There was a problem hiding this comment.
above $str_split_width, $strlen also exists as closure. Shall I move $strlen as private method either?
There was a problem hiding this comment.
and could you please give advice for method naming,
$strlen
->private strlenfor consistency (Helperhas a function with same name)
->private strwidthas this does not return length of string but width- $str_split_width
-> how aboutprivate splitStringByWidth()?
There was a problem hiding this comment.
agreed about splitStringByWidth rather than $str_split_width.
and for $strlen, it could indeed be renamed to a private method stringWidth as it does not compute the length anymore after your change
There was a problem hiding this comment.
Thanks for advice.
Shall I leave the name of Helper#strlen unchanged (which I changed its behavior, too) unlike Application's, as changing that name introduces BC break?
There was a problem hiding this comment.
yeah, it cannot be renamed as it is protected
| } | ||
| $originalEncoding = mb_detect_encoding($string); | ||
|
|
||
| $convertedString = false === $originalEncoding ? $string : mb_convert_encoding($string, 'utf8', $originalEncoding); |
There was a problem hiding this comment.
To be consistent with stringWidth, you should fallback to str_split if mb_string does not support the encoding
There was a problem hiding this comment.
Thanks, I've overlooked the code in stringWidth. I've changed variable's name, too.
|
Thank you @denkiryokuhatsuden. |
…kiryokuhatsuden) This PR was squashed before being merged into the 2.3 branch (closes #10714). Discussion ---------- [Console]Improve formatter for double-width character | Q | A | ------------- | --- | Bug fix? | yes | New feature? | no | BC breaks? | yes for one that expecting current broken output | Deprecations? | no | Fixed tickets | | Tests pass? | yes | License | MIT | Doc PR | EDIT: fixed the table above not to remove irrelevant line As mb_strlen just returns "how many chars in the string", formatting with double-width character is bit broken. The test I add is skipped when mbstring extension is not loaded. I'm afraid if some of you cannot properly display japanese string. (表示するテキスト just means "Some text to display") Commits ------- a52f41d [Console]Improve formatter for double-width character
|
@denkiryokuhatsuden on which PHP version have you been testing this? |
|
@jakzal tested on php5.5. Are there some tests failing? |
|
@denkiryokuhatsuden yes, they were never passing on PHP 5.5. See #10897. |
This PR was merged into the 2.3 branch. Discussion ---------- [Console] Fix a console test | Q | A | ------------- | --- | Bug fix? | yes | New feature? | no | BC breaks? | no | Deprecations? | no | Tests pass? | yes | Fixed tickets | - | License | MIT | Doc PR | - re #10714 Every wide character in the fixture file is actually 3 ansi-characters long. Commits ------- 61108b9 Disable 5.6 until it is stable again. 8cadb49 Update the fixtures.
This PR was merged into the 2.3 branch. Discussion ---------- Explicitly define the encoding. | Q | A | ------------- | --- | Bug fix? | yes | New feature? | no | BC breaks? | no | Deprecations? | no | Tests pass? | yes | Fixed tickets | - | License | MIT | Doc PR | - Credits for discovering it go to @nicolas-grekas. Cheers! Travis for PHP 5.6 cannot be enabled yet as there's one more test failing. re #10714 #10714 Commits ------- 619ff58 Explicitly define the encoding.
EDIT: fixed the table above not to remove irrelevant line
As mb_strlen just returns "how many chars in the string",
formatting with double-width character is bit broken.
The test I add is skipped when mbstring extension is not loaded.
I'm afraid if some of you cannot properly display japanese string.
(表示するテキスト just means "Some text to display")