Created attachment 28320 [details] compileable c++ test case and dwarf dump For strong typed enum, there is no type information: #include <iostream> enum UI: unsigned int { None = 0, Single = 1, Multiple = 0xFFFF0000U }; <1><26aa>: Abbrev Number: 69 (DW_TAG_enumeration_type) DW_AT_name : UI DW_AT_byte_size : 4 DW_AT_decl_file : 1 DW_AT_decl_line : 3 DW_AT_sibling : <26cc> Should have a DW_AT_type : <unsigned int type> under DW_TAG_enumeration_type
Guess we'd need to either move ENUM_FIXED_UNDERLYING_TYPE_P bit (and possibly ENUM_UNDERLYING_TYPE convention) out of the FE to tree.h, or add a langhook to query that info from the FE (won't work well with LTO though, but there are tons of things that don't either).
*** Bug 55447 has been marked as a duplicate of this bug. ***
This problem can trace back from gcc 4.4.4.
But it isn't a regression, so Summary shouldn't be changed for that fact. You can fill in Known to work: (guess empty) and Known to fail:.
Confirming.
I think this is a duplicate of PR debug/16063. With the proposed patch for that applied the debuginfo for the UI enum looks like: [ 243e] enumeration_type name (string) "UI" byte_size (data1) 4 type (ref4) [ 2463] decl_file (data1) 1 decl_line (data1) 3 sibling (ref4) [ 2463] [ 244d] enumerator name (strp) "None" const_value (data1) 0 [ 2453] enumerator name (strp) "Single" const_value (data1) 1 [ 2459] enumerator name (strp) "Multiple" const_value (data4) 4294901760 [ 2463] base_type byte_size (data1) 4 encoding (data1) unsigned (7)
(In reply to Mark Wielaard from comment #6) > I think this is a duplicate of PR debug/16063. I agree. *** This bug has been marked as a duplicate of bug 16063 ***