1) Improve the '-S' option algorithm, to get indirect names, from
DW_AT_specification and DW_AT_abstract_origin attributes, to use
those names in the comparison for member and inlines functions.
2) Correct incorrect calculation of 'pointer_size' for proprietary
EM_PPC64 machine type.
3) Changes required when porting the samples to Windows
4) Include additional information for the Windows version.
BLD
BLDLIBDWARF
BLDTESTDIR
CREATINGARELEASE
Makefile.in
README
UPDATEDWARFDUMPVERSION.sh
configure (regenerated)
configure.in
dwarfdump2/NEWS
This adds improved error reporting in case
of major libdwarf error.
dwarfgen/createirepfrombinary.cc
These are not normally used, but they
enable simple tests of the malloc-fail case.
dwarfgen/TESTmallocfail
dwarfgen/fakemalloc.c
These implement and document new handling
of malloc-space-exhausted special case
for Dwarf_Error. Using a global.
This has no effect on any client code,
it just enables sensible reports (from libdwarf)
in case the malloc arena exhausted.
libdwarf/dwarf_alloc.c
libdwarf/dwarf_error.c
libdwarf/dwarf_error.h
libdwarf/libdwarf.h.in
Document malloc-fail special case
error handling.
libdwarf/libdwarf2.1.mm
Regenerated: v 2.23 libdwarf/libdwarf2.1.pdf
and dwarf_add_AT_with_ref_sig8() to libdwarf producer function list.
Documented the new functions in libdwarf2p.1.mm.
If we already have a DW_AT_sibling attribute we take care not
to add a second one.
dwarfgen now supports some 'class reference' forms (local references
and sig8 references, not yet global references).
Fixed dealloc for dwarf_get_pubtypes to avoid duplicate free.
and adding an error code now used by the producer.
dwarfdump: adding support to print .debug_pubtypes
dwarfgen: adding support to read and write .debug_pubnames
and .debug_pubtypes.
are really misnamed. Now libdwarf2.1.mm exposes the real meaning
and libdwarf2.1.mm (in commentary) does too.
dwarfgen has fixes so it builds 64bit.
Omissions in dwarfgen which were not processing FORMs right
are corrected.
