public inbox for gdb-patches@sourceware.org
 help / color / mirror / Atom feed
* [PATCH] Improve GDB's baseclass detection with typedefs
@ 2022-09-29  7:14 Bruno Larsen
  2022-09-29 17:54 ` Tom Tromey
  0 siblings, 1 reply; 3+ messages in thread
From: Bruno Larsen @ 2022-09-29  7:14 UTC (permalink / raw)
  To: gdb-patches

When a class inherits from a typedef'd baseclass, GDB may be unable to
find the baseclass if the user is not using the typedef'd name, as is
tested on gdb.cp/virtbase2.exp; the reason that test case is working
under gcc is that the dwarf generated by gcc links the class to the
original definition of the baseclass, not to the typedef.  If the
inheritance is linked to the typedef, such as how clang does it,
gdb.cp/virtbase2.exp starts failing.

This can also be seen in gdb.cp/impl-this.exp, when attempting to print
D::Bint::i, and GDB not being able to find the baseclass Bint.

This happens because searching for baseclasses only uses the macro
TYPE_BASECLASS_NAME, which returns the typedef'd name. However, we can't
switch that macro to checking for typedefs, otherwise we wouldn't be
able to find the typedef'd name anymore. This is fixed by searching for
members or baseclasses by name, we check both the saved name and the
name after checking for typedefs.

This also fixes said long-standing bug in gdb.cp/impl-this.exp when the
compiler adds information about typedefs in the debuginfo.
---
 gdb/cp-namespace.c                 |  5 ++--
 gdb/testsuite/gdb.cp/impl-this.exp | 45 ++++++++++++++++++++++--------
 gdb/valops.c                       |  4 +--
 3 files changed, 37 insertions(+), 17 deletions(-)

diff --git a/gdb/cp-namespace.c b/gdb/cp-namespace.c
index 837eafc52f1..1d7759a7669 100644
--- a/gdb/cp-namespace.c
+++ b/gdb/cp-namespace.c
@@ -780,12 +780,13 @@ cp_find_type_baseclass_by_name (struct type *parent_type, const char *name)
   for (i = 0; i < TYPE_N_BASECLASSES (parent_type); ++i)
     {
       struct type *type = check_typedef (TYPE_BASECLASS (parent_type, i));
-      const char *base_name = TYPE_BASECLASS_NAME (parent_type, i);
+      const char *tdef_name = TYPE_BASECLASS_NAME (parent_type, i);
+      const char *base_name = type->name ();
 
       if (base_name == NULL)
 	continue;
 
-      if (streq (base_name, name))
+      if (streq (tdef_name, name) || streq (base_name, name))
 	return type;
 
       type = cp_find_type_baseclass_by_name (type, name);
diff --git a/gdb/testsuite/gdb.cp/impl-this.exp b/gdb/testsuite/gdb.cp/impl-this.exp
index 2910bf7d79d..4a11fb21823 100644
--- a/gdb/testsuite/gdb.cp/impl-this.exp
+++ b/gdb/testsuite/gdb.cp/impl-this.exp
@@ -26,6 +26,8 @@ if {[prepare_for_testing "failed to prepare" $testfile $srcfile {debug c++}]} {
     return -1
 }
 
+set gcc_used [test_compiler_info gcc-*-*]
+
 # First test expressions when there is no context.
 with_test_prefix "before run" {
     gdb_test "print i" "No symbol \"i\" in current context."
@@ -63,15 +65,23 @@ with_test_prefix "at D::f (valid expressions)" {
     gdb_test "print D::i" "= 4"
     gdb_test "print D::B<int>::i" "= 2"
     gdb_test "print B<int>::i" "= 2"
-    gdb_test "print D::Bint::i" \
-	"No type \"Bint\" within class or namespace \"D\"."
+
+    if {$gcc_used} {
+	setup_xfail *-*-* gcc/60833
+    }
+    gdb_test "print D::Bint::i" "= 2"
+
     gdb_test "print Bint::i" "= 2"
     gdb_test "print D::C::i" "= 3"
     gdb_test "print C::i" "= 3"
     gdb_test "print D::B<int>::A<int>::i" "= 1"
     gdb_test "print B<int>::A<int>::i" "= 1"
-    gdb_test "print D::Bint::A<int>::i" \
-	"No type \"Bint\" within class or namespace \"D\"."
+
+    if {$gcc_used} {
+	setup_xfail *-*-* gcc/60833
+    }
+    gdb_test "print D::Bint::A<int>::i" "= 1"
+
     gdb_test "print Bint::A<int>::i" "= 1"
     gdb_test "print A<int>::i" "= 1"
     gdb_test "print D::C::A<int>::i" "= 1"
@@ -88,11 +98,16 @@ with_test_prefix "at D::f (valid expressions)" {
 with_test_prefix "at D::f (invalid expressions)" {
     gdb_test "print D::B<int>::c" "There is no field named c"
     gdb_test "print D::B<int>::A<int>::c" "There is no field named c"
-    gdb_test "print D::Bint::c" \
-	"No type \"Bint\" within class or namespace \"D\"."
 
-    gdb_test "print D::Bint::A<int>::c" \
-	"No type \"Bint\" within class or namespace \"D\"."
+    if {$gcc_used} {
+	setup_xfail *-*-* gcc/60833
+    }
+    gdb_test "print D::Bint::c" "There is no field named c"
+    if {$gcc_used} {
+	setup_xfail *-*-* gcc/60833
+    }
+    gdb_test "print D::Bint::A<int>::c" "There is no field named c"
+
     gdb_test "print D::C::A<int>::c" "There is no field named c"
     gdb_test "print B<int>::c" "There is no field named c"
     gdb_test "print B<int>::A<int>::c" "There is no field named c"
@@ -101,10 +116,16 @@ with_test_prefix "at D::f (invalid expressions)" {
     gdb_test "print C::A<int>::c" "There is no field named c"
     gdb_test "print D::B<int>::x" "There is no field named x"
     gdb_test "print D::B<int>::A<int>::x" "There is no field named x"
-    gdb_test "print D::Bint::x" \
-	"No type \"Bint\" within class or namespace \"D\"."
-    gdb_test "print D::Bint::A<int>::x" \
-	"No type \"Bint\" within class or namespace \"D\"."
+
+    if {$gcc_used} {
+	setup_xfail *-*-* gcc/60833
+    }
+    gdb_test "print D::Bint::x" "There is no field named x"
+    if {$gcc_used} {
+	setup_xfail *-*-* gcc/60833
+    }
+    gdb_test "print D::Bint::A<int>::x" "There is no field named x"
+
     gdb_test "print B<int>::x" "There is no field named x"
     gdb_test "print B<int>::A<int>::x" "There is no field named x"
     gdb_test "print Bint::x" "There is no field named x"
diff --git a/gdb/valops.c b/gdb/valops.c
index b4866d089ac..345be3ea5c9 100644
--- a/gdb/valops.c
+++ b/gdb/valops.c
@@ -2066,9 +2066,7 @@ struct_field_searcher::search (struct value *arg1, LONGEST offset,
 	 name is not yet filled in.  */
       int found_baseclass = (m_looking_for_baseclass
 			     && TYPE_BASECLASS_NAME (type, i) != NULL
-			     && (strcmp_iw (m_name,
-					    TYPE_BASECLASS_NAME (type,
-								 i)) == 0));
+			     && (strcmp_iw (m_name, basetype->name ()) == 0));
       LONGEST boffset = value_embedded_offset (arg1) + offset;
 
       if (BASETYPE_VIA_VIRTUAL (type, i))
-- 
2.37.3


^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [PATCH] Improve GDB's baseclass detection with typedefs
  2022-09-29  7:14 [PATCH] Improve GDB's baseclass detection with typedefs Bruno Larsen
@ 2022-09-29 17:54 ` Tom Tromey
  2022-10-03  8:10   ` Bruno Larsen
  0 siblings, 1 reply; 3+ messages in thread
From: Tom Tromey @ 2022-09-29 17:54 UTC (permalink / raw)
  To: Bruno Larsen via Gdb-patches

>>>>> "Bruno" == Bruno Larsen via Gdb-patches <gdb-patches@sourceware.org> writes:

Bruno> When a class inherits from a typedef'd baseclass, GDB may be unable to
Bruno> find the baseclass if the user is not using the typedef'd name, as is
Bruno> tested on gdb.cp/virtbase2.exp; the reason that test case is working
Bruno> under gcc is that the dwarf generated by gcc links the class to the
Bruno> original definition of the baseclass, not to the typedef.  If the
Bruno> inheritance is linked to the typedef, such as how clang does it,
Bruno> gdb.cp/virtbase2.exp starts failing.

Bruno> This can also be seen in gdb.cp/impl-this.exp, when attempting to print
Bruno> D::Bint::i, and GDB not being able to find the baseclass Bint.

Bruno> This happens because searching for baseclasses only uses the macro
Bruno> TYPE_BASECLASS_NAME, which returns the typedef'd name. However, we can't
Bruno> switch that macro to checking for typedefs, otherwise we wouldn't be
Bruno> able to find the typedef'd name anymore. This is fixed by searching for
Bruno> members or baseclasses by name, we check both the saved name and the
Bruno> name after checking for typedefs.

Bruno> This also fixes said long-standing bug in gdb.cp/impl-this.exp when the
Bruno> compiler adds information about typedefs in the debuginfo.

Thank you for the patch.  This is ok.

Tom

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [PATCH] Improve GDB's baseclass detection with typedefs
  2022-09-29 17:54 ` Tom Tromey
@ 2022-10-03  8:10   ` Bruno Larsen
  0 siblings, 0 replies; 3+ messages in thread
From: Bruno Larsen @ 2022-10-03  8:10 UTC (permalink / raw)
  To: Tom Tromey, Bruno Larsen via Gdb-patches

On 29/09/2022 19:54, Tom Tromey wrote:
>>>>>> "Bruno" == Bruno Larsen via Gdb-patches <gdb-patches@sourceware.org> writes:
> Bruno> When a class inherits from a typedef'd baseclass, GDB may be unable to
> Bruno> find the baseclass if the user is not using the typedef'd name, as is
> Bruno> tested on gdb.cp/virtbase2.exp; the reason that test case is working
> Bruno> under gcc is that the dwarf generated by gcc links the class to the
> Bruno> original definition of the baseclass, not to the typedef.  If the
> Bruno> inheritance is linked to the typedef, such as how clang does it,
> Bruno> gdb.cp/virtbase2.exp starts failing.
>
> Bruno> This can also be seen in gdb.cp/impl-this.exp, when attempting to print
> Bruno> D::Bint::i, and GDB not being able to find the baseclass Bint.
>
> Bruno> This happens because searching for baseclasses only uses the macro
> Bruno> TYPE_BASECLASS_NAME, which returns the typedef'd name. However, we can't
> Bruno> switch that macro to checking for typedefs, otherwise we wouldn't be
> Bruno> able to find the typedef'd name anymore. This is fixed by searching for
> Bruno> members or baseclasses by name, we check both the saved name and the
> Bruno> name after checking for typedefs.
>
> Bruno> This also fixes said long-standing bug in gdb.cp/impl-this.exp when the
> Bruno> compiler adds information about typedefs in the debuginfo.
>
> Thank you for the patch.  This is ok.
Thanks, pushed!


Cheers,
Bruno

>
> Tom
>


^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2022-10-03  8:10 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2022-09-29  7:14 [PATCH] Improve GDB's baseclass detection with typedefs Bruno Larsen
2022-09-29 17:54 ` Tom Tromey
2022-10-03  8:10   ` Bruno Larsen

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).