Re: [dsdp-dd-dev] SPIRIT vs a new standard XML schema? (was Debugmodelan

[Daniel Jacobowitz <drow@xxxxxxxxx>]

On Tue, Jun 06, 2006 at 09:54:32AM -0700, Spear, Aaron wrote:
> 2) It is completely in our control today.  Because of the fact that
> existing SPIRIT files lack some information we might need, the
> translators/parsers could be smart enough to prompt to fill in the gaps
> (I can hear the groans already, but that will be reality until the world
> is on board with us)

Something to think about.  In practice, I expect that there will always
be debuggers which need to do one or more of these things:

  - Specially recognize certain units in the debug information,
    possibly by name.  GDB does this to find which set of registers in
    the target description correspond to its a priori knowledge of the
    architecture, since I did not try to completely eliminate same.
    Example: to locate the correct registers for argument passing.

  - Augment the vendor-provided information with specific configuration
    files written by the debugger vendor, that are not necessarily
    suitable to be code inside the debugger.  For instance,
    associations of units which should be considered and displayed
    together even though they're separate on the bus-derived SPIRIT
    layout.  You might want to be able to show or hide units in
    a hierarchical layout _other_ than SPIRIT's (think the Windows
    "view by connection" versus "view by type" here).

That means that there's a high chance the debugger will need to
preprocess the files anyway.

Of course, that doesn't have to affect this decision; any debugger
wanting those features still has the option of inventing its own
format and translating SPIRIT - or whatever else you choose - to
that!

-- 
Daniel Jacobowitz
CodeSourcery