Bug 4039 – Another possible DWARF issue

Status
RESOLVED
Resolution
FIXED
Severity
normal
Priority
P2
Component
dmd
Product
D
Version
D1 (retired)
Platform
Other
OS
Linux
Creation time
2010-04-01T13:34:00Z
Last change time
2014-02-15T02:17:55Z
Assigned to
nobody
Creator
nfxjfg
Blocks
4044

Comments

Comment #0 by nfxjfg — 2010-04-01T13:34:21Z
$ cat a.d module a; import p.b; void main() { } $ cat p/b.d module p.b; void moo() { } $ dmd a.d p/b.d -c -gc $ dmd a.o b.o $ dwarfdump a > /dev/null dwarfdump ERROR: dwarf_offdie: DW_DLE_DIE_ABBREV_LIST_NULL (151) $ objdump --dwarf a > /dev/null objdump: Warning: .debug_info offset of 0x2f in .debug_pubnames section does not point to a CU header. Please note: dwarfdump may or may not be buggy by itself, and what objdump outputs is just a warning. It may be that this is not really an issue. dmd 1.057 with patch applied from http://d.puremagic.com/issues/show_bug.cgi?id=3987#c10
Comment #1 by nfxjfg — 2010-04-01T13:35:00Z
...and Tango as stdlib.
Comment #2 by robert — 2010-04-01T14:18:07Z
It's reproducible with dmd 2.042 and phobos, so it's probably an issue with how the dwarf info is merged between object files.
Comment #3 by bugzilla — 2011-04-05T22:24:35Z
Curiously, neither dwarfdump nor objdump complain about either a.o or b.o individually.