"No page selected" after saving from editor

[cmb]

Hello Developers,

if the page heading contains more than one whitespace in a row, after saving the page will not be found. This is caused by the latest change to rfc(), which reduces all consecutive whitespace to a single space (for cleaner URLs). I suggest to change cmsimple/adm.php line 458:

Code: Select all

array_splice($temp, -1, 1, uenc(preg_replace('/\s+/isu', ' ', trim(strip_tags($matches[1]))))); 

But I found another issue: it does still not work, when a space is entered at the end of the heading. tinyMCE converts this to a UTF-8 non-breaking space, and this is not stripped of by trim()!

But even worse: in cmsimple/cms.php line 595 the UTF-8 non-breaking space is being partially replaced by a space (\xC2\xA0 -> \xC2\20) resulting in an invalid UTF-8 byte sequence. It seems that browsers do ignore the \xC2, but of course it's a bug (I've introduced it ) to replace a single byte with the upmost bit set in any UTF-8 string. BTW: the same happens with the former version (/\s+/isu), at least if no specific locale is set. So the regex has to be replaced with a more elaborate one (perhaps even regarding other Unicode whitespace characters). Or should mb_ereg() be preferred?

PHP and UTF-8

Christoph

[svasti]

Lot's of reasons to bring out a 1.5.4 before introducing new features.

[cmb]

I've found out, that a similar stripping of whitespace occurs twice in plugins/pluginloader/page_data/page_data_router.php (line 174 and 216 in r247). So I implemented a small function xh_rmws(), which replaces all (consecutive) unicode whitespace characters by a single space. To additionally trim all unicode whitespace from the beginning and end of a string, it has to be called this way:

Code: Select all

trim(xh_rmws($str))