Comment 10 for bug 38986

You were, I were. The reason is that it will take a humongous amount of memory, disk space and CPU to index the whole disk, which will create a full text index of files whose content with 98% likelihood I will never be interested in. Beagle « magic » works only for the file name for almost all system files, for the content there have to be filters for every file type (I wish there'd be a latex filter for example). As for feature regression there is nothing beagle can do for use cases that can be accessed with what you can see in the screenshot /except/ “file name” and “full-text” for very specific mime types.

I mean g-s-t doesn't even have a .desktop file...

For all I care let this sink in for a while. There will be new bugs opened about this, I'm sure.