On 16/05/14 16:19, Chris Green wrote:
Exactly! What I want is a 'find' that uses a database like 'locate'. For me 'find' is perfect in what it does and how it allows me to tune the search but it's too slow if one is searching across a big disk.
Back in the late 1990s when I used to work at Suffolk County Council, we used to run a nightly cronjob which simply did "find / -print > /files.txt". It would then be trivial to do a virtually-instant "grep" for something in this file index, with the understanding that any files added that day would obviously not be included. But then you'd have only just created them, so you'd probably remember, or your application's "open recent" would still have a reference.
You could even write a "locate" wrapper that did both a conventional locate and appended results from a grep in your file index.
KDE also has things like Nepomuk, which index most user-space files for searching (presumably including by file name), but I always turn them off as they always seem to hammer my system at start up.
Simon