R
Roedy Green
Has anyone looked inside File.list to see how the native method is
implemented on Windows?
I notice it that my own File.list code to scan directories seems to
run considerably slower than C utilities, e.g. looking for junk files
to delete. The hard disk also seem to "buzz" as if doing many small
seeks when in runs my code.
I have my hard disk well defragged with the most of the directories
contiguous and lumped together to make such scanning efficient.
I also notice that Take Command, which maintains a simple flat file
mirror listing all the directories on disk can find scan for a string
and jump to a directory ever so much faster than java can scan.
I am wondering if this is just the nature of Java-OS mismatched
impedance, or there is something that could be done to speed up
File.list. It is doing anything stupid?
implemented on Windows?
I notice it that my own File.list code to scan directories seems to
run considerably slower than C utilities, e.g. looking for junk files
to delete. The hard disk also seem to "buzz" as if doing many small
seeks when in runs my code.
I have my hard disk well defragged with the most of the directories
contiguous and lumped together to make such scanning efficient.
I also notice that Take Command, which maintains a simple flat file
mirror listing all the directories on disk can find scan for a string
and jump to a directory ever so much faster than java can scan.
I am wondering if this is just the nature of Java-OS mismatched
impedance, or there is something that could be done to speed up
File.list. It is doing anything stupid?