During some web development at work we’ve discovered some very annoying bug in IE6 and IE7 which (bug) I’d like to describe.
I’ve also done a small video to show everything in detail.
Ok, so we have an IE6 and we need to follow a link containing some fancy UTF-8 characters. No problem – it works ok as long as we paste the link in address bar and click enter.
The problems occur when we access the link in another way. For example, if it is on some page and we click it – under some unknown circumstances the UTF-8 characters get crapped. This problem is always occurring in IE6 (but all works ok in IE7) when I access out portal on our dev server, but somehow never occurs when I do so on local environment. It also don’t occur if I copy link to buffer and paste to address bar, or open link in new window using context menu option.
“You cannot reproduce it so you cannot prove it” some might say. Ok. But I can reproduce it in IE6 and IE7 using another way of accessing link.
The bug is really annoying because it clearly shows that everything in IE is done through arse. You have a link with parameter encoded – why decode it and then encode back in some unpredictable manner? You already have it, why all this redundant decode/encode/decode cycle? Bizarre.
And I cannot really think of a good workaround for this problem. If you know a good workaround for this please write about it in comments (thanks in advance).
UPD: I’ve just discovered that auto-overflow style for adding scroller to “Lenore: more neat bugs I found” image didn’t work for IE6 (although it works under Safari, Opera and FF3).