That was a while ago; since upgrading to 1.1.1, this screen garbage has returned with a vengeance. attaching a file showing what it looks like
scrambled-2.bmp
function convert_urls_into_links(&$text) {
/// Given some text, it converts any URLs it finds into HTML links.
/// Make lone URLs into links. eg http://moodle.com/
$text = eregi_replace("(:space:|^|\(|\[)(:alnum:+)://([^[:space:]]*)([[:alnum:]#?/&=])",
"\\1<a href=\"\\2://\\3\\4\" target=\"newpage\">\\2://\\3\\4</a>", $text);
/// eg www.moodle.com
$text = eregi_replace("(:space:|^|\(|\[)www\.([^[:space:]]*)([[:alnum:]#?/&=])",
"\\1<a href=\"http://www.\\2\\3\" target=\"newpage\">www.\\2\\3</a>", $text);
}
What does yours look like? How did you get the Moodle files onto your server?
Very interesting. I've checked the regex on my server visually character by character three times now and it looks like these are identical. I downloaded the zip file of 1.1.1 onto my windows 98 pc and then uploaded it to the linux server I'm using via a program called WinSCP3. Haven't yet found any visible corruption in any of the zipped files.
Something must be going on, however, because I'm also getting lots of time outs on the main site page, and the error message always says its at line 650 in weblib.php, which is the end of this function. Should I check the code byte by byte?
This is fairly bizarre stuff. It turns out that in the weekly outline content box, a left angle bracket is getting spit out for every character printed in the box, i. e. instead of printing
< a title="edit summary" href=
it prints
< << <a < <t <i <t <l <e <= <" <E <d <i <t < <s <u <m <m <a <r <y <" < <h <r <e <f <=
It also does this for the course descriptions on the site's home page.
After some experimenting, however, I have found that it does not do this when the language is set to Chinese (simplified or traditional) or Japanese. If the language is set to anything else, it spits out a extra left bracket in these two places.
My sysop (a friend) saw this and said "strange!" What do you all say?