RE: [Bacula-devel] A few bugs in 1.37.25 with mainframe tape driv es

[David B. <db...@si...>]

> It looks like your tape drive is not following the tape drive 
> standard, which 
> says that it must return an EOF status *TWO* times when 
> reading. At that 
> point, the program knows it is at the end of the tape. If the 
> program reads 
> again, then and only then should it get an error (i.e. your 
> errno==28).

If you can point me at some independent documentation for this, I'll go
hammer on the IBM people until they fix it. Since this is Linux, the
driver should behave like any other Linux driver. 

(Sticky part about this one: the 359x tape driver is the only Linux/390
device driver that IBM has kept OCO and doesn't distribute the source.
This means I have to make a business case for *them* to fix it rather
than just fixing it. Grr.)