Re: RFC: Represent lexical blocks covering discontiguous ranges

[Jim Blandy <jimb at codesourcery dot com>]

Joel Brobecker <brobecker at adacore.com> writes:
>> This patch extends GDB's blockvector structure so that it can properly
>> represent functions and lexical blocks whose code doesn't occupy a
>> single contiguous range of addresses.  In DWARF debuging information,
>> this allows GDB to handle DW_AT_ranges attribute when it appears on
>> DW_TAG_lexical_block, DW_TAG_try_block, and DW_TAG_catch_block
>> debugging information entries.  When the debugging information
>> provides only a single contiguous range for each block, the code falls
>> back to the old representation.
>> 
>> What do folks think?
>
> I think this is really cool. I guess my only question is: Why do we
> keep the two parallel approaches to blocks? Could we simply only use
> your map and get rid of the old structure?

We could, although it would waste a little space.  Let me chew on that
a bit.