Hui Zhu wrote:
> On Tue, Aug 25, 2009 at 01:51, Michael Snyder<msnyder@vmware.com> wrote:

>> Now I can comment on the dump/load part of the patch.
>> It does not seem as if previous issues have been addressed:
>> 1) Core file and record file are separate, no way to verify
>>   that they correspond to each other.
> 
>> 2) Previous log not cleared when loading a new log.
> For it, I think we don't need worry about it.  Because I make the
> record load together with record_open.  So each time when we load
> record, the record list is empty.

I would like to see it made explicit, please.


>> Have you considered combining this patch with the idea that I
>> proposed -- storing the record log in the core file?  Seems like
>> that would be very easy now -- gdb already has the core file open,
>> and there is a global pointer to the BFD.
>> See gdbcore.h:extern bfd* core_bfd;
> 
> Cool.  If we just want record_core_ops support load, we can remove the
> cmd "record load".  :)
> 
>> In fact, just for fun, I've merged the two patches in my
>> local tree.  It seems to work quite well.  I can just post
>> it here if you like, or here's a hint to show you how it goes:
> 
> 
> Why not?  Please post it.  I can not wait to try it.  :)

Great -- attached.