Mirror of the gdb-patches mailing list
 help / color / mirror / Atom feed
From: Abdul Basit Ijaz <abdul.b.ijaz@intel.com>
To: gdb-patches@sourceware.org
Cc: abdul.b.ijaz@intel.com, tom@tromey.com, kevinb@redhat.com
Subject: [PATCH v2 1/1] gdb, fortran: Fix local variable lookup in Fortran parser
Date: Thu, 30 Apr 2026 18:29:55 +0200	[thread overview]
Message-ID: <20260430162955.295647-2-abdul.b.ijaz@intel.com> (raw)
In-Reply-To: <20260430162955.295647-1-abdul.b.ijaz@intel.com>

This change fixes https://sourceware.org/bugzilla/show_bug.cgi?id=34059.

The Fortran expression parser in GDB ("gdb/f-exp.y") previously performed
symbol lookup in this order when parsing identifiers:

- SEARCH_STRUCT_DOMAIN (types/structs)
- SEARCH_VFT ("C-like" name lookups for variables/types/functions)
- SEARCH_MODULE_DOMAIN (modules)

It searched for "types before variables", causing type names from shared
libraries to shadow local variable names.

For a reproducer like the one given below, the Fortran parser fails to
look up local variable name "array" and treats it as a type value instead
of a variable, because there's a conflicting "array" type from system
libraries.

1 program test
2
3  ! Declare variables used in this test.
4  integer, dimension (-2:2) :: array
5
6  array = 1
7
8  print *, ""           ! Break here
9  print *, array
10
11 end program test

Before the change, GDB shows:

'''
./gdb --data-directory=./data-directory --args /tmp/a.out
GNU gdb (GDB) 18.0.50.20260408-git
Copyright (C) 2026 Free Software Foundation, Inc.
...
Reading symbols from /tmp/a.out...
(gdb) break 8
Breakpoint 1 at 0x11b3: file test.f90, line 8.
(gdb) run
Starting program: /tmp/a.out
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".

Breakpoint 1, test () at test.f90:8
8         print *, ""           ! Break here
(gdb) info locals
array = (1, 1, 1, 1, 1)
(gdb) print array
Attempt to use a type name as an expression
(gdb) ptype array
type = Type array
    Type, C_Union :: :: u
    PTR TO -> ( char :: scratch(0:15 ))
End Type array
'''

This issue is fixed in the Fortran expression parser in GDB ("gdb/f-exp.y")
by prioritizing SEARCH_VFT over other search domains during symbol lookup.
After the change, the 'array' variable is resolved to the correct value.

'''
(gdb) print array
$1 = (1, 1, 1, 1, 1)
(gdb) ptype array
type = integer(kind=4) (-2:2)
'''
---
 gdb/f-exp.y                                   |  2 +-
 .../gdb.fortran/var-type-precedence-lib.c     | 30 ++++++++++
 .../gdb.fortran/var-type-precedence.exp       | 60 +++++++++++++++++++
 .../gdb.fortran/var-type-precedence.f90       | 40 +++++++++++++
 4 files changed, 131 insertions(+), 1 deletion(-)
 create mode 100644 gdb/testsuite/gdb.fortran/var-type-precedence-lib.c
 create mode 100644 gdb/testsuite/gdb.fortran/var-type-precedence.exp
 create mode 100644 gdb/testsuite/gdb.fortran/var-type-precedence.f90

diff --git a/gdb/f-exp.y b/gdb/f-exp.y
index 278e2091403..4216112c10b 100644
--- a/gdb/f-exp.y
+++ b/gdb/f-exp.y
@@ -1651,8 +1651,8 @@ yylex (void)
     struct block_symbol result;
     const domain_search_flags lookup_domains[] =
     {
-      SEARCH_STRUCT_DOMAIN,
       SEARCH_VFT,
+      SEARCH_STRUCT_DOMAIN,
       SEARCH_MODULE_DOMAIN
     };
     int hextype;
diff --git a/gdb/testsuite/gdb.fortran/var-type-precedence-lib.c b/gdb/testsuite/gdb.fortran/var-type-precedence-lib.c
new file mode 100644
index 00000000000..87bd19fad84
--- /dev/null
+++ b/gdb/testsuite/gdb.fortran/var-type-precedence-lib.c
@@ -0,0 +1,30 @@
+/* Copyright 2026 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/>.  */
+
+/* Define a struct type that will conflict with Fortran variable name.  */
+struct type_shadowing_var
+{
+  int value;
+};
+
+/* Global variables to ensure types are in debug info.  */
+static struct type_shadowing_var global_conflicting_var = {42};
+
+/* Function to ensure library is linked and types are used.  */
+void
+fortran_var_type_order_test (void)
+{
+  global_conflicting_var.value = 42;
+}
diff --git a/gdb/testsuite/gdb.fortran/var-type-precedence.exp b/gdb/testsuite/gdb.fortran/var-type-precedence.exp
new file mode 100644
index 00000000000..8c9f80d43d9
--- /dev/null
+++ b/gdb/testsuite/gdb.fortran/var-type-precedence.exp
@@ -0,0 +1,60 @@
+# Copyright 2026 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/>.
+
+# Test variable lookup when local Fortran variable names conflict with
+# type names from a C shared library.
+
+require allow_fortran_tests
+
+standard_testfile var-type-precedence.f90 var-type-precedence-lib.c
+load_lib fortran.exp
+
+set libname "${testfile}-lib"
+set srcfile_lib ${srcdir}/${subdir}/${srcfile2}
+set binfile_lib [standard_output_file ${libname}.so]
+
+# Compile the C shared library.
+if {[gdb_compile_shlib ${srcfile_lib} ${binfile_lib} \
+	{debug}] != ""} {
+  untested "failed to compile C shared library"
+  return
+}
+
+# Compile the Fortran main program with the C shared library.
+if {[prepare_for_testing "failed to prepare" $testfile \
+	${srcfile} [list debug f90 shlib=${binfile_lib}]]} {
+  return
+}
+
+if {![fortran_runto_main]} {
+  return
+}
+
+set bp_loc [gdb_get_line_number "break-here"]
+gdb_breakpoint "$srcfile:$bp_loc"
+gdb_continue_to_breakpoint "stop-at-bp" ".*$srcfile:$bp_loc.*"
+
+set integer4 [fortran_int4]
+
+# Test that local variable 'type_shadowing_var' is resolved correctly.
+gdb_test "print type_shadowing_var" "= \\(1, 1, 1, 1, 1\\)" \
+    "print local type_shadowing_var variable"
+gdb_test "ptype type_shadowing_var" "type = $integer4 \\(-2:2\\)" \
+    "ptype local type_shadowing_var"
+
+# Verify the C types exist in debug info by checking info types.
+gdb_test "info types type_shadowing_var" \
+    ".*File.*var-type-precedence-lib\\.c:.*struct type_shadowing_var.*" \
+    "C type type_shadowing_var exists in debug info"
diff --git a/gdb/testsuite/gdb.fortran/var-type-precedence.f90 b/gdb/testsuite/gdb.fortran/var-type-precedence.f90
new file mode 100644
index 00000000000..d692018faba
--- /dev/null
+++ b/gdb/testsuite/gdb.fortran/var-type-precedence.f90
@@ -0,0 +1,40 @@
+! Copyright 2026 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/> .
+
+! Source code for var-type-precedence test to verify that local Fortran
+! variables take precedence over conflicting C type names from shared
+! libraries.
+
+program test
+  use iso_c_binding
+  implicit none
+
+  interface
+    subroutine fortran_var_type_order_test () bind (C)
+    end subroutine fortran_var_type_order_test
+  end interface
+
+  ! Declare variables with names that conflict with types in C library.
+  integer, dimension (-2:2) :: type_shadowing_var
+
+  ! Call C library function to ensure it's linked.
+  call fortran_var_type_order_test ()
+
+  type_shadowing_var = 1
+
+  print *, "" ! break-here
+  print *, type_shadowing_var
+
+end program test
-- 
2.34.1

Intel Deutschland GmbH
Registered Address: Dornacher Strasse 1, 85622 Feldkirchen, Germany
Tel: +49 89 991 430, www.intel.de
Managing Directors: Harry Demas, Jeffrey Schneiderman, Yin Chong Sorrell
Chairperson of the Supervisory Board: Nicole Lau
Registered Seat: Munich
Commercial Register: Amtsgericht Muenchen HRB 186928


  reply	other threads:[~2026-04-30 16:30 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-04-30 16:29 [PATCH v2 0/1] " Abdul Basit Ijaz
2026-04-30 16:29 ` Abdul Basit Ijaz [this message]
2026-05-01 17:47   ` [PATCH v2 1/1] " Kevin Buettner
2026-05-04  7:05     ` Ijaz, Abdul B

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=20260430162955.295647-2-abdul.b.ijaz@intel.com \
    --to=abdul.b.ijaz@intel.com \
    --cc=gdb-patches@sourceware.org \
    --cc=kevinb@redhat.com \
    --cc=tom@tromey.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