X-Git-Url: http://git.rot13.org/?a=blobdiff_plain;f=Fuse.pm;h=9a9bd9f3feabe4d1f51a357112efb8c8c1024745;hb=edb5c38bcf65f8176b6886ad705d02ac7ba9b8c6;hp=d81e2b5e5cd45d8b7708892e1d822899bca86fbe;hpb=e4a72b0e315193dc5405a0fb20d0fef73d416094;p=perl-fuse.git diff --git a/Fuse.pm b/Fuse.pm index d81e2b5..9a9bd9f 100755 --- a/Fuse.pm +++ b/Fuse.pm @@ -187,29 +187,20 @@ many valid keys. Most of them correspond with names of callback functions, as described in section 'FUNCTIONS YOUR FILESYSTEM MAY IMPLEMENT'. A few special keys also exist: - -debug => boolean - =over 1 -This turns FUSE call tracing on and off. Default is 0 (which means off). - -=back +=item debug => boolean -mountpoint => string +This turns FUSE call tracing on and off. Default is 0 (which means off). -=over 1 +=item mountpoint => string The point at which to mount this filesystem. There is no default, you must specify this. An example would be '/mnt'. -=back - -mountopts => string - -=over 1 +=item mountopts => string -This is a comma seperated list of mount options to pass to the FUSE kernel +This is a comma separated list of mount options to pass to the FUSE kernel module. At present, it allows the specification of the allow_other @@ -219,11 +210,7 @@ need 'user_allow_other' in /etc/fuse.conf as per the FUSE documention mountopts => "allow_other" or mountopts => "" -=back - -threaded => boolean - -=over 1 +=item threaded => boolean This turns FUSE multithreading on and off. The default is 0, meaning your FUSE script will run in single-threaded mode. Note that single-threaded mode also @@ -247,11 +234,7 @@ you're using are also thread-safe. built with USE_ITHREADS, or if you have failed to use threads or threads::shared.) -=back - -nullpath_ok => boolean - -=over 1 +=item nullpath_ok => boolean This flag tells Fuse to not pass paths for functions that operate on file or directory handles. This will yield empty path parameters for functions @@ -348,6 +331,14 @@ Here are the meaning of the fields: (The epoch was at 00:00 January 1, 1970 GMT.) +If you wish to provide sub-second precision timestamps, they may be +passed either as the fractional part of a floating-point value, or as a +two-element array, passed as an array ref, with the first element +containing the number of seconds since the epoch, and the second +containing the number of nanoseconds. This provides complete time +precision, as a floating point number starts losing precision at about +a tenth of a microsecond. So if you really care about that sort of thing... + =head3 readlink Arguments: link pathname. @@ -724,7 +715,7 @@ Keep in mind that read and write are from the client perspective, so read from our end means data is going *out*, and write means data is coming *in*. It can be slightly confusing. -=head1 poll +=head3 poll Arguments: Pathname, poll handle ID (or undef if none), event mask, (optional) file handle