Re: storeio and libstore bugs

[Roland McGrath]

> when reading past the end of a storeio provided file, you get EIO because
> offset is equal to the size of the underlying store, but the only offset
> check is in dev_rw:

EIO is reasonable for a read some distance off the end of the store.
Reading precisely at the end should report EOF (return success with 0 bytes).
I checked in a change that should do that.

> BTW, I have fixed a couple of bugs in the concat store.  It has never worked!
> Now it does, and encoding/decoding, too.  Other stores that don't have an
> open function (mvol, ileave) can not work anymore, too, because the argp
> parser checks that the store has an open function.  Should we add open
> functions for all the stores lacking them?

The concat and ileave stores are created automagically by store_parsed_open
code when there are multiple stores given.  So when the child stores are
all of the same type, you just do -T device hd1s1 hd2s1.  But adding open
functions makes it possible to use different types for the children, with
-T concat foo:name1 bar:name2.  So adding those others too is fine.