* [rfa] Fix crashes on complex types
@ 2002-03-17 10:21 Daniel Jacobowitz
2002-03-19 11:23 ` Elena Zannoni
0 siblings, 1 reply; 3+ messages in thread
From: Daniel Jacobowitz @ 2002-03-17 10:21 UTC (permalink / raw)
To: gdb-patches
The complex type support has bitrotten a bit. We used to set
TYPE_TARGET_TYPE properly in all the places we would create such types,
but a few more have crept in during the intervening years. This patch
just makes sure that field is valid, so that we can use it to print
later. This fixes PR gdb/422.
OK to commit?
--
Daniel Jacobowitz Carnegie Mellon University
MontaVista Software Debian GNU/Linux Developer
2002-03-17 Daniel Jacobowitz <drow@mvista.com>
Fix PR gdb/422.
* c-lang.c (c_create_fundamental_type): Handle FT_COMPLEX,
FT_DBL_PREC_COMPLEX, and FT_EXT_PREC_COMPLEX.
* dwarf2read.c (read_base_type): Set TYPE_TARGET_TYPE for
complex types.
* stabsread.c (rs6000_builtin_type): Likewise.
(read_sun_floating_type): Likewise.
Index: c-lang.c
===================================================================
RCS file: /cvs/src/src/gdb/c-lang.c,v
retrieving revision 1.10
diff -u -p -r1.10 c-lang.c
--- c-lang.c 2002/02/13 18:49:29 1.10
+++ c-lang.c 2002/03/17 18:17:24
@@ -338,6 +338,30 @@ c_create_fundamental_type (struct objfil
TARGET_LONG_DOUBLE_BIT / TARGET_CHAR_BIT,
0, "long double", objfile);
break;
+ case FT_COMPLEX:
+ type = init_type (TYPE_CODE_FLT,
+ 2 * TARGET_FLOAT_BIT / TARGET_CHAR_BIT,
+ 0, "complex float", objfile);
+ TYPE_TARGET_TYPE (type)
+ = init_type (TYPE_CODE_FLT, TARGET_FLOAT_BIT / TARGET_CHAR_BIT,
+ 0, "float", objfile);
+ break;
+ case FT_DBL_PREC_COMPLEX:
+ type = init_type (TYPE_CODE_FLT,
+ 2 * TARGET_DOUBLE_BIT / TARGET_CHAR_BIT,
+ 0, "complex double", objfile);
+ TYPE_TARGET_TYPE (type)
+ = init_type (TYPE_CODE_FLT, TARGET_DOUBLE_BIT / TARGET_CHAR_BIT,
+ 0, "double", objfile);
+ break;
+ case FT_EXT_PREC_COMPLEX:
+ type = init_type (TYPE_CODE_FLT,
+ 2 * TARGET_LONG_DOUBLE_BIT / TARGET_CHAR_BIT,
+ 0, "complex long double", objfile);
+ TYPE_TARGET_TYPE (type)
+ = init_type (TYPE_CODE_FLT, TARGET_LONG_DOUBLE_BIT / TARGET_CHAR_BIT,
+ 0, "long double", objfile);
+ break;
case FT_TEMPLATE_ARG:
type = init_type (TYPE_CODE_TEMPLATE_ARG,
0,
Index: dwarf2read.c
===================================================================
RCS file: /cvs/src/src/gdb/dwarf2read.c,v
retrieving revision 1.50
diff -u -p -r1.50 dwarf2read.c
--- dwarf2read.c 2002/03/14 22:53:35 1.50
+++ dwarf2read.c 2002/03/17 18:17:25
@@ -2979,6 +2979,18 @@ read_base_type (struct die_info *die, st
type = init_type (code, size, type_flags, DW_STRING (attr), objfile);
if (encoding == DW_ATE_address)
TYPE_TARGET_TYPE (type) = dwarf2_fundamental_type (objfile, FT_VOID);
+ else if (encoding == DW_ATE_complex_float)
+ {
+ if (size == 32)
+ TYPE_TARGET_TYPE (type)
+ = dwarf2_fundamental_type (objfile, FT_EXT_PREC_FLOAT);
+ else if (size == 16)
+ TYPE_TARGET_TYPE (type)
+ = dwarf2_fundamental_type (objfile, FT_DBL_PREC_FLOAT);
+ else if (size == 8)
+ TYPE_TARGET_TYPE (type)
+ = dwarf2_fundamental_type (objfile, FT_FLOAT);
+ }
}
else
{
Index: stabsread.c
===================================================================
RCS file: /cvs/src/src/gdb/stabsread.c,v
retrieving revision 1.28
diff -u -p -r1.28 stabsread.c
--- stabsread.c 2002/03/08 17:19:39 1.28
+++ stabsread.c 2002/03/17 18:17:25
@@ -2978,10 +2978,14 @@ rs6000_builtin_type (int typenum)
case 25:
/* Complex type consisting of two IEEE single precision values. */
rettype = init_type (TYPE_CODE_COMPLEX, 8, 0, "complex", NULL);
+ TYPE_TARGET_TYPE (rettype) = init_type (TYPE_CODE_FLT, 4, 0, "float",
+ NULL);
break;
case 26:
/* Complex type consisting of two IEEE double precision values. */
rettype = init_type (TYPE_CODE_COMPLEX, 16, 0, "double complex", NULL);
+ TYPE_TARGET_TYPE (rettype) = init_type (TYPE_CODE_FLT, 8, 0, "double",
+ NULL);
break;
case 27:
rettype = init_type (TYPE_CODE_INT, 1, 0, "integer*1", NULL);
@@ -4511,6 +4515,7 @@ read_sun_floating_type (char **pp, int t
int nbits;
int details;
int nbytes;
+ struct type *rettype;
/* The first number has more details about the type, for example
FN_COMPLEX. */
@@ -4525,9 +4530,12 @@ read_sun_floating_type (char **pp, int t
if (details == NF_COMPLEX || details == NF_COMPLEX16
|| details == NF_COMPLEX32)
- /* This is a type we can't handle, but we do know the size.
- We also will be able to give it a name. */
- return init_type (TYPE_CODE_COMPLEX, nbytes, 0, NULL, objfile);
+ {
+ rettype = init_type (TYPE_CODE_COMPLEX, nbytes, 0, NULL, objfile);
+ TYPE_TARGET_TYPE (rettype)
+ = init_type (TYPE_CODE_FLT, nbytes / 2, 0, NULL, objfile);
+ return rettype;
+ }
return init_type (TYPE_CODE_FLT, nbytes, 0, NULL, objfile);
}
^ permalink raw reply [flat|nested] 3+ messages in thread* Re: [rfa] Fix crashes on complex types
2002-03-17 10:21 [rfa] Fix crashes on complex types Daniel Jacobowitz
@ 2002-03-19 11:23 ` Elena Zannoni
2002-03-19 11:27 ` Daniel Jacobowitz
0 siblings, 1 reply; 3+ messages in thread
From: Elena Zannoni @ 2002-03-19 11:23 UTC (permalink / raw)
To: Daniel Jacobowitz; +Cc: gdb-patches
Daniel Jacobowitz writes:
> The complex type support has bitrotten a bit. We used to set
> TYPE_TARGET_TYPE properly in all the places we would create such types,
> but a few more have crept in during the intervening years. This patch
> just makes sure that field is valid, so that we can use it to print
> later. This fixes PR gdb/422.
Looks good.
>
> OK to commit?
>
Any chance you can sneak in a testcase? Steal the gcc one?
Elena
> --
> Daniel Jacobowitz Carnegie Mellon University
> MontaVista Software Debian GNU/Linux Developer
>
> 2002-03-17 Daniel Jacobowitz <drow@mvista.com>
>
> Fix PR gdb/422.
> * c-lang.c (c_create_fundamental_type): Handle FT_COMPLEX,
> FT_DBL_PREC_COMPLEX, and FT_EXT_PREC_COMPLEX.
> * dwarf2read.c (read_base_type): Set TYPE_TARGET_TYPE for
> complex types.
> * stabsread.c (rs6000_builtin_type): Likewise.
> (read_sun_floating_type): Likewise.
>
> Index: c-lang.c
> ===================================================================
> RCS file: /cvs/src/src/gdb/c-lang.c,v
> retrieving revision 1.10
> diff -u -p -r1.10 c-lang.c
> --- c-lang.c 2002/02/13 18:49:29 1.10
> +++ c-lang.c 2002/03/17 18:17:24
> @@ -338,6 +338,30 @@ c_create_fundamental_type (struct objfil
> TARGET_LONG_DOUBLE_BIT / TARGET_CHAR_BIT,
> 0, "long double", objfile);
> break;
> + case FT_COMPLEX:
> + type = init_type (TYPE_CODE_FLT,
> + 2 * TARGET_FLOAT_BIT / TARGET_CHAR_BIT,
> + 0, "complex float", objfile);
> + TYPE_TARGET_TYPE (type)
> + = init_type (TYPE_CODE_FLT, TARGET_FLOAT_BIT / TARGET_CHAR_BIT,
> + 0, "float", objfile);
> + break;
> + case FT_DBL_PREC_COMPLEX:
> + type = init_type (TYPE_CODE_FLT,
> + 2 * TARGET_DOUBLE_BIT / TARGET_CHAR_BIT,
> + 0, "complex double", objfile);
> + TYPE_TARGET_TYPE (type)
> + = init_type (TYPE_CODE_FLT, TARGET_DOUBLE_BIT / TARGET_CHAR_BIT,
> + 0, "double", objfile);
> + break;
> + case FT_EXT_PREC_COMPLEX:
> + type = init_type (TYPE_CODE_FLT,
> + 2 * TARGET_LONG_DOUBLE_BIT / TARGET_CHAR_BIT,
> + 0, "complex long double", objfile);
> + TYPE_TARGET_TYPE (type)
> + = init_type (TYPE_CODE_FLT, TARGET_LONG_DOUBLE_BIT / TARGET_CHAR_BIT,
> + 0, "long double", objfile);
> + break;
> case FT_TEMPLATE_ARG:
> type = init_type (TYPE_CODE_TEMPLATE_ARG,
> 0,
> Index: dwarf2read.c
> ===================================================================
> RCS file: /cvs/src/src/gdb/dwarf2read.c,v
> retrieving revision 1.50
> diff -u -p -r1.50 dwarf2read.c
> --- dwarf2read.c 2002/03/14 22:53:35 1.50
> +++ dwarf2read.c 2002/03/17 18:17:25
> @@ -2979,6 +2979,18 @@ read_base_type (struct die_info *die, st
> type = init_type (code, size, type_flags, DW_STRING (attr), objfile);
> if (encoding == DW_ATE_address)
> TYPE_TARGET_TYPE (type) = dwarf2_fundamental_type (objfile, FT_VOID);
> + else if (encoding == DW_ATE_complex_float)
> + {
> + if (size == 32)
> + TYPE_TARGET_TYPE (type)
> + = dwarf2_fundamental_type (objfile, FT_EXT_PREC_FLOAT);
> + else if (size == 16)
> + TYPE_TARGET_TYPE (type)
> + = dwarf2_fundamental_type (objfile, FT_DBL_PREC_FLOAT);
> + else if (size == 8)
> + TYPE_TARGET_TYPE (type)
> + = dwarf2_fundamental_type (objfile, FT_FLOAT);
> + }
> }
> else
> {
> Index: stabsread.c
> ===================================================================
> RCS file: /cvs/src/src/gdb/stabsread.c,v
> retrieving revision 1.28
> diff -u -p -r1.28 stabsread.c
> --- stabsread.c 2002/03/08 17:19:39 1.28
> +++ stabsread.c 2002/03/17 18:17:25
> @@ -2978,10 +2978,14 @@ rs6000_builtin_type (int typenum)
> case 25:
> /* Complex type consisting of two IEEE single precision values. */
> rettype = init_type (TYPE_CODE_COMPLEX, 8, 0, "complex", NULL);
> + TYPE_TARGET_TYPE (rettype) = init_type (TYPE_CODE_FLT, 4, 0, "float",
> + NULL);
> break;
> case 26:
> /* Complex type consisting of two IEEE double precision values. */
> rettype = init_type (TYPE_CODE_COMPLEX, 16, 0, "double complex", NULL);
> + TYPE_TARGET_TYPE (rettype) = init_type (TYPE_CODE_FLT, 8, 0, "double",
> + NULL);
> break;
> case 27:
> rettype = init_type (TYPE_CODE_INT, 1, 0, "integer*1", NULL);
> @@ -4511,6 +4515,7 @@ read_sun_floating_type (char **pp, int t
> int nbits;
> int details;
> int nbytes;
> + struct type *rettype;
>
> /* The first number has more details about the type, for example
> FN_COMPLEX. */
> @@ -4525,9 +4530,12 @@ read_sun_floating_type (char **pp, int t
>
> if (details == NF_COMPLEX || details == NF_COMPLEX16
> || details == NF_COMPLEX32)
> - /* This is a type we can't handle, but we do know the size.
> - We also will be able to give it a name. */
> - return init_type (TYPE_CODE_COMPLEX, nbytes, 0, NULL, objfile);
> + {
> + rettype = init_type (TYPE_CODE_COMPLEX, nbytes, 0, NULL, objfile);
> + TYPE_TARGET_TYPE (rettype)
> + = init_type (TYPE_CODE_FLT, nbytes / 2, 0, NULL, objfile);
> + return rettype;
> + }
>
> return init_type (TYPE_CODE_FLT, nbytes, 0, NULL, objfile);
> }
^ permalink raw reply [flat|nested] 3+ messages in thread* Re: [rfa] Fix crashes on complex types
2002-03-19 11:23 ` Elena Zannoni
@ 2002-03-19 11:27 ` Daniel Jacobowitz
0 siblings, 0 replies; 3+ messages in thread
From: Daniel Jacobowitz @ 2002-03-19 11:27 UTC (permalink / raw)
To: Elena Zannoni; +Cc: gdb-patches
On Tue, Mar 19, 2002 at 02:22:12PM -0500, Elena Zannoni wrote:
> Daniel Jacobowitz writes:
> > The complex type support has bitrotten a bit. We used to set
> > TYPE_TARGET_TYPE properly in all the places we would create such types,
> > but a few more have crept in during the intervening years. This patch
> > just makes sure that field is valid, so that we can use it to print
> > later. This fixes PR gdb/422.
>
> Looks good.
OK, I'll check it in later today. Thanks!
> > OK to commit?
> >
>
> Any chance you can sneak in a testcase? Steal the gcc one?
Well, I don't know how. I'll think about it. The problem is that
complex types in C require either C99 or at least GCC; I'll see if I
can handle the testcase not compiling gracefully.
--
Daniel Jacobowitz Carnegie Mellon University
MontaVista Software Debian GNU/Linux Developer
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2002-03-19 19:27 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2002-03-17 10:21 [rfa] Fix crashes on complex types Daniel Jacobowitz
2002-03-19 11:23 ` Elena Zannoni
2002-03-19 11:27 ` Daniel Jacobowitz
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox