Description:
------------
Can we have the extra features added to strtr() that are in the GNU strtr? They'd be really useful.

In particular, the ability to delete certain characters from the string [currently we must resort to str_replace, which gets ugly], and to squeeze repeated characters to a single one [requires preg_replace() at the moment] 

This could be done by adding some predefined constants such as 
 STRTR_DELETE, STRTR_SQUEEZE_REPEATS 



Reproduce code:
---------------
In GNU tr, we can do things like

echo "hello world" | tr -s "l" 
#results in "helo world"             (duplicate 'l' removed)

echo "hello world" | tr -d 'od'      ('o' and 'd' removed)
#results in "hell wrl"

echo "hello world" | tr -s "wl" "zl"  (translate AND squeeze)
#results in "helo zorld"


(Support for character classes might be nice too)

Thank you.