[PATCH][gdb/exp] Fix internal error when printing C++ pointer-to-member

[PATCH][gdb/exp] Fix internal error when printing C++ pointer-to-member

[Tom de Vries]

Hi,

When running the test-case included with this patch, we run into:
...
(gdb) print ptm^M
$1 = gdb/gdbtypes.h:695: internal-error: loc_bitpos: \
  Assertion `m_loc_kind == FIELD_LOC_KIND_BITPOS' failed.^M
...
while printing a c++ pointer-to-member.

Fix this by skipping fields with m_loc_kind != FIELD_LOC_KIND_BITPOS in
cp_find_class_member, such that we have:
...
(gdb) print ptm^M
$1 = &A::i^M
...

Tested on x86_64-linux.

https://sourceware.org/bugzilla/show_bug.cgi?id=29294

Any comments?

Thanks,
- Tom

[gdb/exp] Fix internal error when printing C++ pointer-to-member

---
 gdb/cp-valprint.c                          |  5 ++++-
 gdb/testsuite/gdb.cp/pointer-to-member.cc  | 34 ++++++++++++++++++++++++++++++
 gdb/testsuite/gdb.cp/pointer-to-member.exp | 28 ++++++++++++++++++++++++
 3 files changed, 66 insertions(+), 1 deletion(-)

diff --git a/gdb/cp-valprint.c b/gdb/cp-valprint.c
index 43a52698b71..11d75ced602 100644
--- a/gdb/cp-valprint.c
+++ b/gdb/cp-valprint.c
@@ -636,7 +636,10 @@ cp_find_class_member (struct type **self_p, int *fieldno,
 
   for (i = TYPE_N_BASECLASSES (self); i < len; i++)
     {
-      LONGEST bitpos = self->field (i).loc_bitpos ();
+      field &f = self->field (i);
+      if (f.loc_kind () != FIELD_LOC_KIND_BITPOS)
+	continue;
+      LONGEST bitpos = f.loc_bitpos ();
 
       QUIT;
       if (offset == bitpos)
diff --git a/gdb/testsuite/gdb.cp/pointer-to-member.cc b/gdb/testsuite/gdb.cp/pointer-to-member.cc
new file mode 100644
index 00000000000..0570c9cc8ae
--- /dev/null
+++ b/gdb/testsuite/gdb.cp/pointer-to-member.cc
@@ -0,0 +1,34 @@
+/* This testcase is part of GDB, the GNU debugger.
+
+   Copyright 2022 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/>.  */
+
+struct A
+{
+  int i;
+};
+
+struct B : public A
+{
+  static const int j = 0;
+};
+
+int B::*ptm = &B::i;
+
+int
+main ()
+{
+  return 0;
+}
diff --git a/gdb/testsuite/gdb.cp/pointer-to-member.exp b/gdb/testsuite/gdb.cp/pointer-to-member.exp
new file mode 100644
index 00000000000..d946a60af7b
--- /dev/null
+++ b/gdb/testsuite/gdb.cp/pointer-to-member.exp
@@ -0,0 +1,28 @@
+# Copyright (C) 2022 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.
+
+# Test printing c++ pointer-to-member.
+
+if { [skip_cplus_tests] } { continue }
+
+standard_testfile .cc
+
+if {[prepare_for_testing "failed to prepare" $testfile $srcfile {debug c++}]} {
+    return -1
+}
+
+gdb_test "print ptm" " = &A::i"

Re: [PATCH][gdb/exp] Fix internal error when printing C++ pointer-to-member

[Tom Tromey]

>>>>> "Tom" == Tom de Vries via Gdb-patches <gdb-patches@sourceware.org> writes:

Tom> (gdb) print ptm^M
Tom> $1 = gdb/gdbtypes.h:695: internal-error: loc_bitpos: \
Tom>   Assertion `m_loc_kind == FIELD_LOC_KIND_BITPOS' failed.^M
Tom> ...
Tom> while printing a c++ pointer-to-member.

Is this crashing due to the static field?

Tom> Fix this by skipping fields with m_loc_kind != FIELD_LOC_KIND_BITPOS in
Tom> cp_find_class_member, such that we have:
Tom> ...
Tom> (gdb) print ptm^M
Tom> $1 = &A::i^M
Tom> ...

If so then I think it's better to skip when the field is static, not
based on the loc_kind.

Tom

Re: [PATCH][gdb/exp] Fix internal error when printing C++ pointer-to-member

[Tom de Vries]

On 7/5/22 20:47, Tom Tromey wrote:
>>>>>> "Tom" == Tom de Vries via Gdb-patches <gdb-patches@sourceware.org> writes:
> 
> Tom> (gdb) print ptm^M
> Tom> $1 = gdb/gdbtypes.h:695: internal-error: loc_bitpos: \
> Tom>   Assertion `m_loc_kind == FIELD_LOC_KIND_BITPOS' failed.^M
> Tom> ...
> Tom> while printing a c++ pointer-to-member.
> 
> Is this crashing due to the static field?
> 

Yes.

> Tom> Fix this by skipping fields with m_loc_kind != FIELD_LOC_KIND_BITPOS in
> Tom> cp_find_class_member, such that we have:
> Tom> ...
> Tom> (gdb) print ptm^M
> Tom> $1 = &A::i^M
> Tom> ...
> 
> If so then I think it's better to skip when the field is static, not
> based on the loc_kind.

Ack, committed using field_is_static instead.

Thanks,
- Tom