GDB gnats: c++ bugs

Rick Moseley rmoseley@redhat.com
Tue Aug 26 21:13:00 GMT 2008


Jan Kratochvil wrote:
> Hi Sami,
>
> On Tue, 26 Aug 2008 17:14:19 +0200, Sami Wagiaalla wrote:
>   
>> There are more bugs of course but these are the ones I'll be looking at  
>> for now.
>>     
>
>   
> When you are at it there are also:
> KFAIL: gdb.cp/templates.exp: constructor breakpoint (PRMS: gdb/1062)
> KFAIL: gdb.cp/templates.exp: destructor breakpoint (PRMS: gdb/1112)
>   
Tom had me do a search of the gdb bugzilla for all C++ bugs with "ptype" 
errors and 1112 popped up on that list.  I'm also looking at the 
following to verify they are all still valid since most were posted > 3 
years ago:

1111 - ptype prints return value for constructors in templated classes

1112 - can't breakpoint in  destructor

1113 - ptype claims that destructors take an int arg

1593 - problem with nested namespaces

1734 - gdb does not store an empty class

1783/2271/2273(all seem to be same) - gdb can't lookup class typedef 
when in methods of class
    *NOTE - Jim Ingham filed 1783 with a lot of detail and leaves this 
ominous note:
          "Not sure how to fix this without breaking other legitimate 
cases of namespace lookup, however"

1817 - double-colon not understood in typecast

There are also a couple of bugs in the gcc data base concerning gdb:

7932 - Emit debug information about non-type templates

30161 - gcc should generate dwarf info about template.


I have worked through all of the gdb bugs, except 1593, verifying they 
are all still around.  I have yet to look at the gcc ones, not sure yet 
if there is any connection between those 2 and some of gdb bugs.

I'll try to come up with "finger in the breeze" estimates tomorrow.

Rick




More information about the Archer mailing list