0021410: text looks OK in edit of note but renders badly - MantisBT - Signup temporarily disabled due to spam

ID	Project	Category	View Status	Date Submitted	Last Update

0021410	mantisbt	ui	public	2016-07-22 09:07	2016-08-28 01:12

Reporter	geoff@streeter.gr	Assigned To	dregad
Priority	normal	Severity	minor	Reproducibility	always
Status	closed	Resolution	fixed
OS	Linux	OS Version	Kubuntu 14.04
Product Version	1.2.3
Target Version	1.3.1	Fixed in Version	1.3.1

Summary	0021410: text looks OK in edit of note but renders badly
Description	text looks OK in edit of note but renders badly
Steps To Reproduce	Wrap <\pre> tags around the text in the file and use it as a note
Tags	No tags attached.
Attached Files	partner.desc (3,393 bytes)

geoff@streeter.gr 2016-07-22 09:09 reporter ~0053689	Valid structures on a component: Normal: pointer inside name space points to its follow ┌───────────────────────────────────────────────┐ │NSR FOLLOW │ │ ↑ │ │ * │ └───────────────────────────────────────────────┘ Acceptable: pointer inside name space points to its NSR ┌───────────────────────────────────────────────┐ │NSR FOLLOW │ │↑ │ │* │ └───────────────────────────────────────────────┘ Pointer inside subspace points to its parent. Normal. ┌───────────────────────────────────────────────┐ │NSR FOLLOW │ │ ↑ │ │ │ ┌───────┐ │ │ └────────* │ │ │ └───────┘ │ │ │ └───────────────────────────────────────────────┘ Pointer inside subspace points to its parent. Acceptable. ┌───────────────────────────────────────────────┐ │NSR FOLLOW │ │↑ │ ││ ┌───────┐ │ │└────────* │ │ │ └───────┘ │ │ │ └───────────────────────────────────────────────┘ Pointer inside subspace points to another name space. Inside original, outside subspace. Normal. ┌───────────────────────────────────────────────┐ │NSR FOLLOW │ │ ┌───────────────┐ │ │ ┌───────┐ │NSR FOLLOW │ │ │ │ │ │↑ │ │ │ │───────────────┘ │ │ │ └───────┘ └───────────────┘ │ │ │ └───────────────────────────────────────────────┘ Pointer inside subspace points to another name space. Inside original, outside subspace. Acceptable. ┌───────────────────────────────────────────────┐ │NSR FOLLOW │ │ ┌───────────────┐ │ │ ┌───────┐ │NSR FOLLOW │ │ │ │ │ │ ↑ │ │ │ │──────────────────────┘ │ │ │ └───────┘ └───────────────┘ │ │ │ └───────────────────────────────────────────────┘

geoff@streeter.gr 2016-07-22 09:10 reporter ~0053690	The incomplete boxes and extra lines are not supposed to be there.

cproensa 2016-08-08 07:01 developer ~0053757	The issue occurs at string_nl2br() function Disabling wrapping pre text shows the correct line breaks: $g_wrap_in_preformatted_text = OFF;

cproensa 2016-08-08 08:45 developer ~0053758 Last edited: 2016-08-08 08:46	this may fix it: note adding "u" as modifier for the regexp. --- a/core/string_api.php +++ b/core/string_api.php @@ -123,7 +123,7 @@ function string_nl2br( $p_string, $p_wrap = 100 ) { -$t_output .= preg_replace( '/([^\n]{' . $p_wrap . ',}?[\s]+)(?!<\/pre>)/', "$1\n", $t_piece ); +$t_output .= preg_replace( '/([^\n]{' . $p_wrap . ',}?[\s]+)(?!<\/pre>)/u', "$1\n", $t_piece ); seems like the 100 character count for line wrap is counted in bytes, for extended chars, unless the option is set? I'm not such expert in regexp+unicode, anyone with better knwoledge can confirm?

dregad 2016-08-08 09:02 developer ~0053759	Correct, without the 'u' flag, pcre treats strings as sequences of 1-byte chars. IMO we should get rid of $g_wrap_in_preformatted_text, and add CSS 'overflow-x' attribute to 'pre' tags.

cproensa 2016-08-09 07:49 developer ~0053766	IMO we should get rid of $g_wrap_in_preformatted_text, and add CSS 'overflow-x' attribute to 'pre' tags. Can we fix the 'u' in current 1.3.x and remove the option in v2?

dregad 2016-08-09 09:41 developer ~0053772	Follow-up issue: 0021594

MantisBT: master-1.3.x ba89d0e1 2016-08-09 05:16 dregad Details Diff	Prevent unwanted line breaks in <pre> tags By default, Mantis wraps text within <pre> tags at 100 chars (see $g_wrap_in_preformatted_text option). If the text within the tag contains Unicode extended characters, the regex performing the wrap does not count them properly, as pcre works on bytes, causing line breaks to be added in the wrong place. Adding 'u' modifier to the regex fixes the behavior, as suggested by cproensa. Fixes 0021410		Affected Issues 0021410
	mod - core/string_api.php		Diff File