Readline read_history and write_history addition

[Yitzchak Gale]

Sterling Clover wrote:
>> ...as we've got a built-in standard for
>> handling precisely the sorts of IO errors that are returned, it seems
>> silly not to use it.

IO exceptions wreak havoc on programs using any instance
of MonadIO other than IO itself. At best, you need a major
reorganization of your entire program. At worst, it can make
it completely impossible to write a library.

The built-in standard that you are referring to is fine
for when something happens that makes it impossible
for the program to continue - out of memory, hard disk
failure, etc. But in cases where it is possible to return
a meaningful value and continue, why be so disruptive?

As Robert Dockins pointed out, it is very easy for
those who prefer an IO exception to throw one
themselves. The opposite can be difficult or impossible.

Thanks,
Yitz