We have a JS file that contains our Ajax App string bundles.
It has special characters, such as ® and Japanese characters. The file is encoded in UTF-8.
When a diff between the workspace version and the checked-in version (Perforce) is made, the left side of the diff shows the file as if it's encoded in ASCII (possibly OS default), even though the default set for IJ is UTF8 and the file is encoded in UTF8.
The right side (local copy) shows up encoded correctly in UTF8.
What that causes is that all the lines with special or Japanese characters appear to be modified, rendering the diff in this case irrelevant.
As a result, the end user has to leave IntelliJ and do the diff in the Perforce client (P4V in my case).
In Demetra it works fine. It broke in Selena, since the 1st EAP build.
Environment: Vista / XP Pro, JRE6, Perforce
Amnon
If I ask IJ to show me the file history and I compare 2 versions on Perforce, it is fine (left and right show Japanese characters).
But, if I have the file checked out and compare it by pressing Ctrl+D in the Changes panel, the left side shows squares, and the right shows Japanese, which is strange, as the right side is the checked-out workspace copy and the left side is the checked in version.