| Fix some compiler warnings. |
|
More...
|
about 9 years ago
|
| Don't try and shortcut a typedef that refers to an unresolved type by making it refer to the target of the unresolved type. This is correct for nested typedefs, but not for other type references. |
|
More...
|
over 11 years ago
|
| Always search the whole of the type list when fixing up type references, as we can't guarantee any particular order of backward and forward references in the DWARF so just searching part of the list is not safe. |
|
More...
|
over 11 years ago
|
| Avoid buffer overflows with long type names. |
|
More...
|
almost 12 years ago
|
| Handle a typedef that refers back to a previous typedef that itself was a forward reference that hasn't been resolved yet. |
|
More...
|
almost 12 years ago
|
| Ignore any debug file with a .gnu_debugaltlink section for now, as this means that dwz multi-file compression has been used and libdwarf doesn't handle this yet and it will cause all sorts of crashes. |
|
More...
|
about 12 years ago
|
| Update libdwarf to 20140208 release |
|
More...
|
about 12 years ago
|
| Update to 20140131 release of libdwarf. |
|
More...
|
about 12 years ago
|
| Give up on trying to share file records for header files as it means we can't work out which CU to read for a inline function because we may find the wrong parent file. |
|
More...
|
about 12 years ago
|
| Only warn if the address goes backward within a single function. |
|
More...
|
about 12 years ago
|
| Because we now share header files between multiple source files we can't just mark every function in the header as done as soon as we load line numbers for one of the source files. |
|
More...
|
about 12 years ago
|
| We strip a leading "./" from file names when adding file records so we need to do the same when searching, or we won't match them. |
|
More...
|
about 12 years ago
|
| Try the DWARF4 standard DW_AT_linkage_name attribute as well as the DW_AT_MIPS_linkage_name extension when looking for the mangled name of a function. |
|
More...
|
about 12 years ago
|
| Don't try and compare, and possible update, a function's name if we don't have one as gcc sometimes seems to output automatically generated constructors with no name. |
|
More...
|
about 12 years ago
|
| Avoid creating duplicate fil_t records for the same header file when it is referenced from more than one source file in the same executable or shared library. |
|
More...
|
about 12 years ago
|
| Compute high_pc correctly for functions when it is recorded in the DWARF as an offset from low_pc. |
|
More...
|
about 12 years ago
|
| Beef up the DWARF unwinder a bit. |
|
More...
|
about 12 years ago
|
| Handle alignment correctly in 64 bit ELF images. |
|
More...
|
about 12 years ago
|
| Don't try and use DWARF variable location data for globals |
|
More...
|
about 12 years ago
|
| Add support for DWARF variables whose location varies during the course of a function's execution. |
|
More...
|
about 12 years ago
|
| Don't try and examine an attribute that we failed to read. |
|
More...
|
about 12 years ago
|
| Install the abort handler after the setjmp context is created. |
|
More...
|
about 12 years ago
|
| Don't terminate a DWARF unwind just because FP becomes zero - the correct termination condition is when there is an unwind record which does not specify an unwind rule for the PC. |
|
More...
|
about 12 years ago
|
| Remove the user abort function before exiting wait_with_intr or we might wind up jumping to a sigjmp context that is no longer valid if a signal arrives between exiting wait_with_intr and the user abort handler being removed. |
|
More...
|
about 12 years ago
|
| Make sure the user abort function only fires once. |
|
More...
|
about 12 years ago
|
| Only remove the SIGIO handler after we have cleared the async flag on the socket, or we might get a signal in the interim that will then cause the process to abort. |
|
More...
|
about 12 years ago
|
| Avoid reading past the end of strings when comparing tokens. |
|
More...
|
about 12 years ago
|
| Fix compiler warnings. |
|
More...
|
about 12 years ago
|
| Handle SP relative variables in the C interpreter. |
|
More...
|
about 12 years ago
|
| Don't abort if a variable location is too complicated to deal with. |
|
More...
|
about 12 years ago
|
