[Dwarf-Discuss] Vendor extensions in .debug_macinfo

David Anderson davea42 at earthlink.net
Fri Jul 29 14:37:33 PDT 2011


ISSUE 1

In the new 6.3.1.5 text:

"Each so defined opcode is valid for subsequent entries
until the terminating entry with type code 0, including any sequences
included from those entries using DW_MACRO_transparent_include."

seems awkward,  it is really unclear whether any transparent_include
terminating entry terminates the DW_MACRO_define_opcode or not.


My attempts at better wording have not been fruitful so far.
I think you mean 'nested include terminating entries don't terminate
DW_MACRO_define_opcode'  but since a nested include could define its own
opcodes (right?) it gets messy to say properly.


-----------------------------------------
ISSUE 2

6.3.5 Legacy macro information




"just a series of macinfo entries.  There are 5 legacy macinfo opcode types:"

Remove the noise word "just", it adds nothing.

---------------------------------------------
ISSUE 3

Final  comment: IMO we should be  careful to say
that any unused bytes in the section are all zero.

That still leaves a problem in that detecting
the next section header after  an arbitrary
series of zero bytes  is endian-dependent (version
number being 2 bytes, of course).

Or we could disallow unused bytes entirely.
Or we could say all bytes must be part of valid .debug_macro
sequences (not necessarily saying that all are referenced from
debug_info).
-----------------

Summary:
Other than the above three issues  .debug_macro seems fine to me.

DavidA.

-- 
On two occasions I have been asked [by members of Parliament!], 'Pray, Mr. Babbage, if you put into the machine wrong figures, will the right answers come out?' I am not able rightly to apprehend the kind of confusion of ideas that could provoke such a question.  -- Charles Babbage




More information about the Dwarf-Discuss mailing list