From: Doug Evans <dje@google.com>
To: Jan Kratochvil <jan.kratochvil@redhat.com>
Cc: Tom Tromey <tromey@redhat.com>, gdb-patches@sourceware.org
Subject: Re: [patch] fix exp/12117
Date: Thu, 14 Oct 2010 01:55:00 -0000 [thread overview]
Message-ID: <AANLkTinksJaBz98tN3duEyzrT=GTi-PezsgeYEbi1ngq@mail.gmail.com> (raw)
In-Reply-To: <20101013211008.GA23114@host1.dyn.jankratochvil.net>
[-- Attachment #1: Type: text/plain, Size: 2694 bytes --]
On Wed, Oct 13, 2010 at 2:10 PM, Jan Kratochvil
<jan.kratochvil@redhat.com> wrote:
> Also I do not like much the pr[0-9]+.exp names, one gets fimiliar with any
> name but I never remember PR numbers. [in fact an idea by Roland McGrath]
I often just go with the flow for things such as this, in the absence
of specific rules.
If we want to declare a specific rule of no more such prNNN naming,
that's fine by me.
> @@ -1388,7 +1389,7 @@ check_typedef (struct type *type)
> /* It is dangerous to call lookup_symbol if we are currently
> reading a symtab. Infinite recursion is one danger. */
> if (currently_reading_symtab)
> - return type;
> + return make_cv_type (is_const, is_volatile, type, NULL);
>
> name = type_name_no_tag (type);
> /* FIXME: shouldn't we separately check the TYPE_NAME and
I left this out for a later pass to be conservative and avoid having
to deal with any unintended consequences.
But if one wants to handle this in the same patch, ok.
> @@ -1398,7 +1399,7 @@ check_typedef (struct type *type)
> if (name == NULL)
> {
> stub_noname_complaint ();
> - return type;
> + return make_cv_type (is_const, is_volatile, type, NULL);
> }
> sym = lookup_symbol (name, 0, STRUCT_DOMAIN, 0);
> if (sym)
Ditto.
> @@ -1407,11 +1408,10 @@ check_typedef (struct type *type)
> TYPE_TARGET_TYPE (type) = alloc_type_arch (get_type_arch (type));
> }
> type = TYPE_TARGET_TYPE (type);
> + is_const |= TYPE_CONST (type);
> + is_volatile |= TYPE_VOLATILE (type);
> }
>
> - is_const = TYPE_CONST (type);
> - is_volatile = TYPE_VOLATILE (type);
> -
> /* If this is a struct/class/union with no fields, then check
> whether a full definition exists somewhere else. This is for
> systems where a type definition with no fields is issued for such
This part doesn't feel right.
For example, when a stub is later filled in the accumulated cv
qualifiers will get filled in too, but that's wrong.
[Right?]
Plus there's more than just const/volatile to preserve.
[Right?]
I like this patch better.
2010-10-13 Doug Evans <dje@google.com>
Jan Kratochvil <jan.kratochvil@redhat.com>
PR exp/12117
* gdbtypes.c (check_typedef): Clean up function comment.
Keep track of instance flags as we strip typedefs and create a new
type to preserve them if necessary.
testsuite/
* gdb.cp/ptype-cv-cp.cc: New file.
* gdb.cp/ptype-cv-cp.exp: New file.
[-- Attachment #2: gdb-101013-12117-3.patch.txt --]
[-- Type: text/plain, Size: 9425 bytes --]
2010-10-13 Doug Evans <dje@google.com>
Jan Kratochvil <jan.kratochvil@redhat.com>
PR exp/12117
* gdbtypes.c (check_typedef): Clean up function comment.
Keep track of instance flags as we strip typedefs and create a new
type to preserve them if necessary.
testsuite/
* gdb.cp/ptype-cv-cp.cc: New file.
* gdb.cp/ptype-cv-cp.exp: New file.
Index: gdbtypes.c
===================================================================
RCS file: /cvs/src/src/gdb/gdbtypes.c,v
retrieving revision 1.201
diff -u -p -r1.201 gdbtypes.c
--- gdbtypes.c 12 Oct 2010 20:58:17 -0000 1.201
+++ gdbtypes.c 14 Oct 2010 01:44:11 -0000
@@ -1347,7 +1347,21 @@ stub_noname_complaint (void)
complaint (&symfile_complaints, _("stub type has NULL name"));
}
-/* Added by Bryan Boreham, Kewill, Sun Sep 17 18:07:17 1989.
+/* Find the real type of TYPE. This function returns the real type,
+ after removing all layers of typedefs, and completing opaque or stub
+ types. Completion changes the TYPE argument, but stripping of
+ typedefs does not.
+
+ Instance flags (e.g. const/volatile) are preserved as typedefs are
+ stripped. If necessary a new qualified form of the underlying type
+ is created.
+
+ NOTE: This will return a typedef if TYPE_TARGET_TYPE for the typedef has
+ not been computed and we're either in the middle of reading symbols, or
+ there was no name for the typedef in the debug info.
+
+ If TYPE is a TYPE_CODE_TYPEDEF, its length is updated to the length of
+ the target type.
If this is a stubbed struct (i.e. declared as struct foo *), see if
we can find a full definition in some other file. If so, copy this
@@ -1355,26 +1369,15 @@ stub_noname_complaint (void)
(but not any code) that if we don't find a full definition, we'd
set a flag so we don't spend time in the future checking the same
type. That would be a mistake, though--we might load in more
- symbols which contain a full definition for the type.
-
- This used to be coded as a macro, but I don't think it is called
- often enough to merit such treatment.
-
- Find the real type of TYPE. This function returns the real type,
- after removing all layers of typedefs and completing opaque or stub
- types. Completion changes the TYPE argument, but stripping of
- typedefs does not.
-
- If TYPE is a TYPE_CODE_TYPEDEF, its length is (also) set to the length of
- the target type instead of zero. However, in the case of TYPE_CODE_TYPEDEF
- check_typedef can still return different type than the original TYPE
- pointer. */
+ symbols which contain a full definition for the type. */
struct type *
check_typedef (struct type *type)
{
struct type *orig_type = type;
- int is_const, is_volatile;
+ /* While we're removing typedefs, we don't want to lose qualifiers.
+ E.g., const/volatile. */
+ int instance_flags = TYPE_INSTANCE_FLAGS (type);
gdb_assert (type);
@@ -1388,7 +1391,7 @@ check_typedef (struct type *type)
/* It is dangerous to call lookup_symbol if we are currently
reading a symtab. Infinite recursion is one danger. */
if (currently_reading_symtab)
- return type;
+ return make_qualified_type (type, instance_flags, NULL);
name = type_name_no_tag (type);
/* FIXME: shouldn't we separately check the TYPE_NAME and
@@ -1398,7 +1401,7 @@ check_typedef (struct type *type)
if (name == NULL)
{
stub_noname_complaint ();
- return type;
+ return make_qualified_type (type, instance_flags, NULL);
}
sym = lookup_symbol (name, 0, STRUCT_DOMAIN, 0);
if (sym)
@@ -1407,10 +1410,33 @@ check_typedef (struct type *type)
TYPE_TARGET_TYPE (type) = alloc_type_arch (get_type_arch (type));
}
type = TYPE_TARGET_TYPE (type);
- }
- is_const = TYPE_CONST (type);
- is_volatile = TYPE_VOLATILE (type);
+ /* Preserve the instance flags as we traverse down the typedef chain.
+
+ Handling address spaces/classes is nasty, what do we do if there's a
+ conflict?
+ E.g., what if an outer typedef marks the type as class_1 and an inner
+ typedef marks the type as class_2?
+ This is the wrong place to do such error checking. We leave it to
+ the code that created the typedef in the first place to flag the
+ error. We just pick the outer address space (akin to letting the
+ outer cast in a chain of casting win), instead of assuming
+ "it can't happen". */
+ {
+ const int ALL_SPACES = (TYPE_INSTANCE_FLAG_CODE_SPACE
+ | TYPE_INSTANCE_FLAG_DATA_SPACE);
+ const int ALL_CLASSES = TYPE_INSTANCE_FLAG_ADDRESS_CLASS_ALL;
+ int new_instance_flags = TYPE_INSTANCE_FLAGS (type);
+
+ /* Treat code vs data spaces and address classes separately. */
+ if ((instance_flags & ALL_SPACES) != 0)
+ new_instance_flags &= ~ALL_SPACES;
+ if ((instance_flags & ALL_CLASSES) != 0)
+ new_instance_flags &= ~ALL_CLASSES;
+
+ instance_flags |= new_instance_flags;
+ }
+ }
/* If this is a struct/class/union with no fields, then check
whether a full definition exists somewhere else. This is for
@@ -1428,7 +1454,7 @@ check_typedef (struct type *type)
if (name == NULL)
{
stub_noname_complaint ();
- return type;
+ return make_qualified_type (type, instance_flags, NULL);
}
newtype = lookup_transparent_type (name);
@@ -1445,7 +1471,9 @@ check_typedef (struct type *type)
move over any other types NEWTYPE refers to, which could
be an unbounded amount of stuff. */
if (TYPE_OBJFILE (newtype) == TYPE_OBJFILE (type))
- make_cv_type (is_const, is_volatile, newtype, &type);
+ type = make_qualified_type (newtype,
+ TYPE_INSTANCE_FLAGS (type),
+ type);
else
type = newtype;
}
@@ -1464,17 +1492,18 @@ check_typedef (struct type *type)
if (name == NULL)
{
stub_noname_complaint ();
- return type;
+ return make_qualified_type (type, instance_flags, NULL);
}
sym = lookup_symbol (name, 0, STRUCT_DOMAIN, 0);
if (sym)
{
/* Same as above for opaque types, we can replace the stub
- with the complete type only if they are int the same
+ with the complete type only if they are in the same
objfile. */
if (TYPE_OBJFILE (SYMBOL_TYPE(sym)) == TYPE_OBJFILE (type))
- make_cv_type (is_const, is_volatile,
- SYMBOL_TYPE (sym), &type);
+ type = make_qualified_type (SYMBOL_TYPE (sym),
+ TYPE_INSTANCE_FLAGS (type),
+ type);
else
type = SYMBOL_TYPE (sym);
}
@@ -1534,8 +1563,12 @@ check_typedef (struct type *type)
TYPE_TARGET_STUB (type) = 0;
}
}
+
+ type = make_qualified_type (type, instance_flags, NULL);
+
/* Cache TYPE_LENGTH for future use. */
TYPE_LENGTH (orig_type) = TYPE_LENGTH (type);
+
return type;
}
Index: testsuite/gdb.cp/ptype-cv-cp.cc
===================================================================
RCS file: testsuite/gdb.cp/ptype-cv-cp.cc
diff -N testsuite/gdb.cp/ptype-cv-cp.cc
--- /dev/null 1 Jan 1970 00:00:00 -0000
+++ testsuite/gdb.cp/ptype-cv-cp.cc 14 Oct 2010 00:33:32 -0000
@@ -0,0 +1,17 @@
+typedef int my_int;
+typedef const my_int const_my_int;
+typedef volatile my_int volatile_my_int;
+typedef volatile const_my_int volatile_const_my_int;
+typedef const volatile_my_int const_volatile_my_int;
+
+my_int v_my_int (0);
+const_my_int v_const_my_int (1);
+volatile_my_int v_volatile_my_int (2);
+const_volatile_my_int v_const_volatile_my_int (3);
+volatile_const_my_int v_volatile_const_my_int (4);
+
+int
+main ()
+{
+ return 0;
+}
Index: testsuite/gdb.cp/ptype-cv-cp.exp
===================================================================
RCS file: testsuite/gdb.cp/ptype-cv-cp.exp
diff -N testsuite/gdb.cp/ptype-cv-cp.exp
--- /dev/null 1 Jan 1970 00:00:00 -0000
+++ testsuite/gdb.cp/ptype-cv-cp.exp 14 Oct 2010 00:33:32 -0000
@@ -0,0 +1,41 @@
+# Copyright 2010 Free Software Foundation, Inc.
+
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 3 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program. If not, see <http://www.gnu.org/licenses/>.
+
+# This file is part of the gdb testsuite.
+
+if { [skip_cplus_tests] } { continue }
+
+set testfile "ptype-cv-cp"
+set srcfile ${testfile}.cc
+
+if { [prepare_for_testing ${testfile}.exp ${testfile} ${srcfile} {debug c++}] } {
+ return -1
+}
+
+gdb_test "whatis v_my_int" "my_int"
+gdb_test "ptype v_my_int" "int"
+
+gdb_test "whatis v_const_my_int" "const_my_int"
+gdb_test "ptype v_const_my_int" "const int"
+
+gdb_test "whatis v_volatile_my_int" "volatile_my_int"
+gdb_test "ptype v_volatile_my_int" "volatile int"
+
+gdb_test "whatis v_const_volatile_my_int" "const_volatile_my_int"
+gdb_test "ptype v_const_volatile_my_int" "const volatile int"
+
+gdb_test "whatis v_volatile_const_my_int" "volatile_const_my_int"
+setup_kfail "gcc/45997" *-*-*
+gdb_test "ptype v_volatile_const_my_int" "const volatile int"
next prev parent reply other threads:[~2010-10-14 1:55 UTC|newest]
Thread overview: 12+ messages / expand[flat|nested] mbox.gz Atom feed top
2010-10-13 17:53 Doug Evans
2010-10-13 18:03 ` Doug Evans
2010-10-13 18:44 ` Tom Tromey
2010-10-13 18:55 ` Doug Evans
2010-10-13 19:04 ` Tom Tromey
2010-10-13 20:55 ` Doug Evans
2010-10-13 21:10 ` Jan Kratochvil
2010-10-14 1:55 ` Doug Evans [this message]
2010-10-14 9:35 ` Jan Kratochvil
2010-10-15 17:42 ` Tom Tromey
2010-10-15 17:49 ` Jan Kratochvil
2010-10-15 17:41 ` Tom Tromey
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to='AANLkTinksJaBz98tN3duEyzrT=GTi-PezsgeYEbi1ngq@mail.gmail.com' \
--to=dje@google.com \
--cc=gdb-patches@sourceware.org \
--cc=jan.kratochvil@redhat.com \
--cc=tromey@redhat.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox