This is kind of bizarre.

The self referencing [video] tag works well, but when used with node or index parameters, it displays "video not available". The bug happens only if the tag is preceded with more than 5 Cyrillic letters.

This works:

Some latin text
[video:node=100:index=0]

This falis:

Some измешан text
[video:node=100:index=0]

This works:

Некој miksed text
[video:node=100:index=0]
CommentFileSizeAuthor
#7 flash_mb3.patch2 KBmiopa
#6 flash_mb2.patch3.56 KBmiopa
#2 flash_mb.patch4.27 KBmiopa

Comments

dmjossel’s picture

dmjossel commented

Well... crap. And here I am about to try and test this on a site with Russian as the default language...

miopa’s picture

miopa commented
Status: Active » Needs review
StatusFileSize
new flash_mb.patch4.27 KB

The bug comes from use of single-byte strpos and substr_replace functions on multi-byte strings.

Because there is no multi-byte substr_replace function built in PHP, I included in the patch one that I found in the User Contributed Notes on http://www.php.net/substr_replace.

travist’s picture

travist commented

Does this work for non-multi-byte character configurations? Can anyone verify?

miopa’s picture

miopa commented

It should work, the configuration is irrelevant, the affected functions apply to the body text. But i don't have any non utf-8 installation to check.

attheshow’s picture

attheshow commented

At first glance, this looks like a fairly lengthy function just to do string replacement. Is there really no existing method of doing this without maintaining our own custom function for it?

miopa’s picture

miopa commented
StatusFileSize
new flash_mb2.patch3.56 KB

I noticed that I forgot to replace one occurrence of substr_replace, so here is an update patch.

miopa’s picture

miopa commented
StatusFileSize
new flash_mb3.patch2 KB

I've analyzed the code and found a way to fix it without the added function. The thing is, I'm not sure if it's worse to maintain one extra function or this mash-up of single-byte and multi-byte string pointers that brings to memory the long forgotten horrors of C++.