[PING*2][PATCH v3 0/2] [gdb] untie and validate inheritance location

Bruno Larsen blarsen@redhat.com
Mon Nov 1 13:17:20 GMT 2021


ping

On 10/19/21 16:15, Bruno Larsen wrote:
> ping
> 
> On 9/30/21 16:26, Bruno Larsen wrote:
>> This started with what was thought of as an incorrect inheritance
>> location expression, but turned out to be a correct, but unusual
>> expression, and uncovered an incorrect assumption when calculating
>> offsets of baseclasses.
>>
>> This patchset is the new iteration of that solution, that changed
>> gnuv3_baseclass_offset calculation to allow for non-virtual inheritance
>> that has a non-trivial location expression and also moved the basic
>> validation to this function. The check in value_contents_copy_raw was
>> changed to an assert.
>> Finally, the test was expanded to have one incorrect trivial location,
>> and two non-trivial location expressions, one correct and one incorrect.
>>
>> Bruno Larsen (2):
>>    [gdb] Untied inheritance virtuality and loc kind
>>    [gdb] Add Simple offset validation when calculating baseclass_offset
>>
>>   gdb/gnu-v3-abi.c                              |  23 +-
>>   .../gdb.dwarf2/dw2-inheritance-locexpr-1.exp  | 233 +++++++++++++++++
>>   .../gdb.dwarf2/dw2-inheritance-locexpr-2.exp  | 235 ++++++++++++++++++
>>   .../gdb.dwarf2/dw2-inheritance-locexpr.c      |  69 +++++
>>   gdb/value.c                                   |   8 +-
>>   5 files changed, 561 insertions(+), 7 deletions(-)
>>   create mode 100644 gdb/testsuite/gdb.dwarf2/dw2-inheritance-locexpr-1.exp
>>   create mode 100644 gdb/testsuite/gdb.dwarf2/dw2-inheritance-locexpr-2.exp
>>   create mode 100644 gdb/testsuite/gdb.dwarf2/dw2-inheritance-locexpr.c
>>
> 
> 


-- 
Cheers!
Bruno Larsen



More information about the Gdb-patches mailing list