digitalmars.D.bugs - [Issue 4212] New: DWARF: void arrays cause gdb errors
- d-bugmail puremagic.com (35/35) May 20 2010 http://d.puremagic.com/issues/show_bug.cgi?id=4212
- d-bugmail puremagic.com (15/15) May 20 2010 http://d.puremagic.com/issues/show_bug.cgi?id=4212
- d-bugmail puremagic.com (56/56) May 20 2010 http://d.puremagic.com/issues/show_bug.cgi?id=4212
- d-bugmail puremagic.com (12/12) May 20 2010 http://d.puremagic.com/issues/show_bug.cgi?id=4212
http://d.puremagic.com/issues/show_bug.cgi?id=4212
Summary: DWARF: void arrays cause gdb errors
Product: D
Version: 1.057
Platform: Other
OS/Version: Linux
Status: NEW
Severity: normal
Priority: P2
Component: DMD
AssignedTo: nobody puremagic.com
ReportedBy: nfxjfg gmail.com
(This bug report is for dmd 1.061)
This code:
extern (C) void abort();
struct X {
union {
void[] a;
void[a.sizeof] b;
}
}
void foo(X b) {
abort();
}
void main() {
foo(X.init);
}
Makes gdb output "Dwarf Error: Cannot find DIE at 0x0 referenced from DIE at
0x68 [in module /tmp/test/dr]" when the abort() jumps back into the debugger,
and you request a stacktrace with "bt".
I suspect something is wrong with static void[] arrays.
--
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
May 20 2010
d-bugmail puremagic.com
http://d.puremagic.com/issues/show_bug.cgi?id=4212
Robert Clipsham <robert octarineparrot.com> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |robert octarineparrot.com
20:21:57 BST ---
Test case is for D1, with D2 it fails with an error that has no line number
information. Reduced test case for D2:
----
void[8] foo = void;
void main(){}
----
--
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
May 20 2010
http://d.puremagic.com/issues/show_bug.cgi?id=4212
Robert Clipsham <robert octarineparrot.com> changed:
What |Removed |Added
----------------------------------------------------------------------------
Keywords| |patch
20:45:35 BST ---
The patch below fixes both test cases:
Index: backend/dwarf.c
===================================================================
--- backend/dwarf.c (revisión: 494)
+++ backend/dwarf.c (copia de trabajo)
-1558,6 +1558,13
DW_AT_type, DW_FORM_ref4,
0, 0,
};
+ static unsigned char abbrevTypeArrayVoid[] =
+ {
+ DW_TAG_array_type,
+ 1, // child (the subrange type)
+ DW_AT_sibling, DW_FORM_ref4,
+ 0, 0,
+ };
static unsigned char abbrevTypeSubrange[] =
{
DW_TAG_subrange_type,
-1573,7 +1580,6
DW_AT_type, DW_FORM_ref4,
0, 0,
};
- unsigned code1 = dwarf_abbrev_code(abbrevTypeArray,
sizeof(abbrevTypeArray));
unsigned code2 = (t->Tflags & TFsizeunknown)
? dwarf_abbrev_code(abbrevTypeSubrange2,
sizeof(abbrevTypeSubrange2))
: dwarf_abbrev_code(abbrevTypeSubrange,
sizeof(abbrevTypeSubrange));
-1581,12 +1587,15
unsigned idxsibling = 0;
unsigned siblingoffset;
nextidx = dwarf_typidx(t->Tnext);
+ unsigned code1 = nextidx ? dwarf_abbrev_code(abbrevTypeArray,
sizeof(abbrevTypeArray))
+ : dwarf_abbrev_code(abbrevTypeArrayVoid,
sizeof(abbrevTypeArrayVoid));
idx = infobuf->size();
infobuf->writeuLEB128(code1); // DW_TAG_array_type
siblingoffset = infobuf->size();
infobuf->write32(idxsibling); // DW_AT_sibling
- infobuf->write32(nextidx); // DW_AT_type
+ if (nextidx)
+ infobuf->write32(nextidx); // DW_AT_type
infobuf->writeuLEB128(code2); // DW_TAG_subrange_type
infobuf->write32(idxbase); // DW_AT_type
--
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
May 20 2010
http://d.puremagic.com/issues/show_bug.cgi?id=4212
Walter Bright <bugzilla digitalmars.com> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
CC| |bugzilla digitalmars.com
Resolution| |FIXED
22:07:14 PDT ---
changeset 495 (dwarf.c changes)
--
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
May 20 2010