www.digitalmars.com         C & C++   DMDScript  

digitalmars.D - Debug information for enumerator values

reply "Iain Buclaw" <ibuclaw ubuntu.com> writes:
I've implemented this in gdc.

https://github.com/D-Programming-GDC/GDC/commit/021dda8feaba282fec60600729ba8abc2c64cf81


---
enum enum_byte   : byte   { kE1, kE2, kE3 }
enum enum_ubyte  : ubyte  { kE1, kE2, kE3 }
enum enum_short  : short  { kE1, kE2, kE3 }
enum enum_ushort : ushort { kE1, kE2, kE3 }
enum enum_int    : int    { kE1, kE2, kE3 }
enum enum_uint   : uint   { kE1, kE2, kE3 }
enum enum_long   : long   { kE1, kE2, kE3 }
enum enum_ulong  : ulong  { kE1, kE2, kE3 }

void main()
{
   enum_byte var_byte;
   enum_ubyte var_ubyte;
   enum_short var_short;
   enum_ushort var_ushort;
   enum_int var_int;
   enum_uint var_uint;
   enum_long var_long;
   enum_ulong var_ulong;
}
---

(gdb) print var_byte
$1 = test.enum_byte.kE1

(gdb) print var_ubyte
$2 = test.enum_ubyte.kE1

(gdb) print var_short
$3 = test.enum_short.kE1

(gdb) print var_ushort
$4 = test.enum_ushort.kE1

(gdb) print var_int
$5 = test.enum_int.kE1

(gdb) print var_uint
$6 = test.enum_uint.kE1

(gdb) print var_long
$7 = test.enum_long.kE1

(gdb) print var_ulong
$8 = test.enum_ulong.kE1

(gdb) print (ulong)var_ulong
$9 = 0

(gdb) print (byte)'test.enum_uint.kE2'
$10 = 1 '\001'

(gdb) print ('test.enum_ulong')3
$11 = (test.enum_ulong.kE2 | test.enum_ulong.kE3)

(gdb) print ('test.enum_ulong')2
$12 = test.enum_ulong.kE3

(gdb) print ('test.enum_ulong')4
$13 = (unknown: 4)


What do you think?  Is <module>.<name>.<member> too verbose, or 
just right? :-)

Regards
Iain
Sep 17 2013
next sibling parent "Tove" <tove fransson.se> writes:
On Tuesday, 17 September 2013 at 09:52:37 UTC, Iain Buclaw wrote:
 (gdb) print ('test.enum_ulong')3
 $11 = (test.enum_ulong.kE2 | test.enum_ulong.kE3)

 (gdb) print ('test.enum_ulong')2
 $12 = test.enum_ulong.kE3

 What do you think?  Is <module>.<name>.<member> too verbose, or 
 just right? :-)

 Regards
 Iain

Kickass! I think it's "just right"... _BUT_ in case of multiple values, I would prefer something like this: $11 = test.enum_ulong(kE2 | kE3)
Sep 17 2013
prev sibling next sibling parent Iain Buclaw <ibuclaw ubuntu.com> writes:
On 17 September 2013 11:02, Tove <tove fransson.se> wrote:
 On Tuesday, 17 September 2013 at 09:52:37 UTC, Iain Buclaw wrote:
 (gdb) print ('test.enum_ulong')3
 $11 = (test.enum_ulong.kE2 | test.enum_ulong.kE3)

 (gdb) print ('test.enum_ulong')2
 $12 = test.enum_ulong.kE3

 What do you think?  Is <module>.<name>.<member> too verbose, or just
 right? :-)

 Regards
 Iain

Kickass! I think it's "just right"... _BUT_ in case of multiple values, I would prefer something like this: $11 = test.enum_ulong(kE2 | kE3)

Need to open a bug report on 'wishlist' items in gdb D support. As I plan to get round to hacking it sometime soon... -- Iain Buclaw *(p < e ? p++ : p) = (c & 0x0f) + '0';
Sep 17 2013
prev sibling parent Iain Buclaw <ibuclaw ubuntu.com> writes:
On 17 September 2013 10:52, Iain Buclaw <ibuclaw ubuntu.com> wrote:
 I've implemented this in gdc.

 https://github.com/D-Programming-GDC/GDC/commit/021dda8feaba282fec60600729ba8abc2c64cf81

Now done the same for manifest constants. https://github.com/D-Programming-GDC/GDC/commit/6b40c481068530374abb0819b80c7d2ef09072fd --- enum A { a = 42, b = 45 } enum foo = A.a; void main() { enum bar = A.b; enum baz = &main; return; } -- (gdb) print 'test.foo' $1 = enum.A.a (gdb) print bar $2 = enum.A.b (gdb) print baz $3 = (void (* const)(void)) 0x402a00 <D main> (gdb) p (int)bar $4 = 45 Note that these values are purely in the debug code only - they are not addressable in memory, and are not emitted in non-debug object files/binaries - meaning there is no bloat added. (gdb) p &'test.foo' Can't take address of "test.foo" which isn't an lvalue. (gdb) p &baz Can't take address of "baz" which isn't an lvalue. Hope this helps those who use CTFE a lot. ;-) -- Iain Buclaw *(p < e ? p++ : p) = (c & 0x0f) + '0';
Sep 17 2013