Version: (using KDE Devel) Installed from: Compiled sources For example, Konqueror often detects bad MIME types from text files: A C header is detected as Objective-C header, a binary file is detected as HTML file because there is the "<html>" string somewhere at the beginning of the file... We could have an option to make MIME type detection use the following algorithm: - Use the file extension. - If the file extension is not associated with anything, then don't try to guess the MIME type. - If there is no file extension, try to guess the MIME type.
Yes, "guessing" ("mime magic") is unreliable, but it's still very useful in many cases. The solution is not to disable guessing, but to 1) improve the guessing 2) use extensions where guessing can't do it right After updating to KDE4 (the mime implementation is completely different, it's now based on freedesktop.org shared-mime-info), please report any wrong "mimetype guessing" with example files.