NAME   
    trash -- move file(s) into trashcan directory or remove them  
   
SYNOPSIS   
    trash [<option>]... <file>...   -- trash the files   
    trash -h   -- for this message 
   
OPTIONS 
   -h  -- for this help message  
   -s  -- suppress all diagnostics except for total failure 
       to remove a file (but see -f) and error exit messages. 
   -v  -- verbose (default) 
   -f  -- exit 0 even if can't remove files. 
       This also suppresses diagnostics about removal failure.
   -rm -- remove if we can't rename (default)  
   -no-rm -- do NOT remove if we can't rename 
   -same-name -- try first to rename using same filename; 
       if this fails, then use a made up new name. 
       Rename to same name is failed if file by this name exists 
       already, but this check is prone to race conditions. 
   -no-same-name -- use only made up new names in trashcan dir. 
       This is the default. 
   -nfs-safe -- ignore .nfs* files when doing removals. 
       This is default and avoids some problems with the dirty details 
       of some nfs provided filesystems.  
   -no-nfs-safe -- don't treat .nfs* files specially.  
   -   -- all following arguments are to be treated as files to remove 
           (otherwise args beginning with `-' are options).  
   
DESCRIPTION   
  
The trash program may be used as an optimized equivalent to the 
traditional unix `rm -rf' to remove files / directories. 
The optimization occurs when files (including directories) 
can be renamed into a trashcan directory (which must be on the 
same filesystem i.e. same disk partition).  
For removing a directory with many files, it is far faster to simply 
rename it into a trashcan directory. 
Of course, it must eventually be removed from the trashcan directory 
to avoid filling up the fileysystem.  
In addition to speed, there are two other advantages: 
-- It is possible to rename directories which cannot be removed by 
a non-root user due to the presence in the directory of files owned 
by another user.  
-- Recovery of files that were unintentially trashed may be simplified. 
  
The trash program attempts renames only into trashcan directories 
that are on the same filesystem as the file to be removed. 
If the trash program cannot rename away a file, it attempts to 
(recursively) remove it. 
Unlike some damaged versions of rm, the trash program first attempts 
to make directories writeable before removing files from them. 
  
The trash program knows the locations of trashcan directories from the 
environmental variable TRASH which should contain a colon-delimited 
set of paths for the trashcan directories (same format as for PATH). 
If this env. variable is missing, then it can only do actual deletions. 
  
It is recommended that the system administrator set the value for TRASH 
in the commonly used shell startup files, and keep this in synch with 
trashcan directories on each filesystem, and with a root cron job run  
e.g. nightly to remove contents of the trashcan directories. 
It is also possible for an individual user to do something similar, 
but without root permission it may sometimes be impossible to remove 
all of the contents of a trashcan directory.  
   
When renaming into a trashcan directory, the default is to make up a 
new name (and to print this name to stderr). 
The new name is composed of various information which put together is 
likely to be unique, and a check for existing file is made to avoid 
overwriting; still race conditions are possible. 
The current format for made up names is:  
       YYYYMMDD.hhmmss.prog.pid.iter.filename  
Where: YYYY is year, MM is month 01-12, DD is day 01-31, hh is hour 00-23, 
   mm is minute 00-59, ss is second 00-59, 
   prog is name of program that trashed the file, if known (limit 20 chars), 
   pid is process id on of program on computer it ran on, 
   iter is added for uniqueness, and filename is filename of renamed file 
   (limit 30 chars).