Mirror of the gdb-patches mailing list
 help / color / mirror / Atom feed
From: Tom de Vries <tdevries@suse.de>
To: gdb-patches@sourceware.org
Subject: Re: [PING][PATCH] [gdb/exp] Limit workaround in using_direct::valid_line to broken GCC versions
Date: Mon, 13 Jul 2026 18:17:01 +0200	[thread overview]
Message-ID: <00c8dabe-0035-44d1-911c-1cb041bef589@suse.de> (raw)
In-Reply-To: <6e2bce3d-96f2-4957-bf5a-bf66219bd7dd@suse.de>

On 6/18/26 9:53 AM, Tom de Vries wrote:
> On 6/4/26 6:21 PM, Tom de Vries wrote:
>> Consider test.c:
>> ...
>>       1    namespace mod_a {
>>       2      int xxx = 10;
>>       3    }
>>       4
>>       5    static void
>>       6    foo ()
>>       7    {
>>       8    }
>>       9
>>      10    int
>>      11    main ()
>>      12    {
>>      13      {
>>      14        foo ();
>>      15        using namespace mod_a;
>>      16      }
>>      17
>>      18      return mod_a::xxx;
>>      19    }
>> ...
>> compiled with "g++ test.c -g".
>>
>> Attempting to print xxx at line 14 shouldn't find anything (because it's
>> before the "using namespace mod_a"), but it does:
>> ...
>> $ gdb -q -batch a.out -ex start -ex "p xxx"
>> ...
>> Temporary breakpoint 1, main () at test.c:14
>> 14        foo ();
>> $1 = 10
>> ...
>>
>> This happens because using_direct::valid_line returns true here:
>> ...
>>        return (decl_line <= curr_sal.line)
>>          || (decl_line >= boundary);
>> ...
>>
>> Since we have decl_line == 15 and curr_sal.line == 14,
>> "(decl_line <= curr_sal.line)" evaluates to false.
>>
>> But boundary == 14, so "(decl_line >= boundary)" evaluates to true.
>>
>> The "(decl_line >= boundary)" bit was added as a workaround for
>> GCC PR debug/108716.
>>
>> Since I'm using GCC 15, the workaround is not needed.
>>
>> Fix this by limiting the workaround to broken GCC versions.
>>
> 
> Ping.
> 

I've pushed this.

Thanks,
- Tom

> Thanks,
> - Tom
> 
>> Tested on x86_64-linux, using GCC 15.2, and 7.5.
>>
>> Bug: https://sourceware.org/bugzilla/show_bug.cgi?id=34203
>> ---
>>   gdb/cp-namespace.c                       | 26 +++++++++++---
>>   gdb/namespace.c                          |  9 +++--
>>   gdb/testsuite/gdb.cp/ns-before-using.c   | 37 ++++++++++++++++++++
>>   gdb/testsuite/gdb.cp/ns-before-using.exp | 44 ++++++++++++++++++++++++
>>   4 files changed, 110 insertions(+), 6 deletions(-)
>>   create mode 100644 gdb/testsuite/gdb.cp/ns-before-using.c
>>   create mode 100644 gdb/testsuite/gdb.cp/ns-before-using.exp
>>
>> diff --git a/gdb/cp-namespace.c b/gdb/cp-namespace.c
>> index c8cd5c245aa..16fd4e2ded8 100644
>> --- a/gdb/cp-namespace.c
>> +++ b/gdb/cp-namespace.c
>> @@ -33,6 +33,7 @@
>>   #include "namespace.h"
>>   #include "inferior.h"
>>   #include "gdbsupport/unordered_map.h"
>> +#include "producer.h"
>>   #include <string>
>>   #include <string.h>
>> @@ -410,9 +411,26 @@ cp_lookup_symbol_via_imports (const char *scope,
>>       found_symbols[sym.symbol->m_name] = sym;
>>       }
>> -  /* Due to a GCC bug, we need to know the boundaries of the current 
>> block
>> -     to know if a certain using directive is valid.  */
>> -  symtab_and_line boundary_sal = find_sal_for_pc (block->end () - 1, 0);
>> +  unsigned boundary_line = 0;
>> +  {
>> +    struct symbol *fn = block->containing_function ();
>> +    int major, minor;
>> +    if (fn != nullptr
>> +    && producer_is_gcc (fn->symtab ()->compunit ()->producer (),
>> +                &major, &minor)
>> +    && (major <= 9
>> +        || (major == 10 && minor < 5)
>> +        || (major == 11 && minor < 4)
>> +        || (major == 12 && minor < 3)
>> +        || (major == 13 && minor < 1)))
>> +      {
>> +    /* Due to a GCC bug (PR debug/108716, fixed in 10.5, 11.4, 12.3, 
>> 13.1),
>> +       we need to know the boundaries of the current block to know if a
>> +       certain using directive is valid.  */
>> +    symtab_and_line boundary_sal = find_sal_for_pc (block->end () - 
>> 1, 0);
>> +    boundary_line = boundary_sal.line;
>> +      }
>> +  }
>>     /* Go through the using directives.  If any of them add new names to
>>        the namespace we're searching in, see if we can find a match by
>> @@ -423,7 +441,7 @@ cp_lookup_symbol_via_imports (const char *scope,
>>         /* If the using directive was below the place we are stopped at,
>>        do not use this directive.  */
>> -      if (!current->valid_line (boundary_sal.line))
>> +      if (!current->valid_line (boundary_line))
>>       continue;
>>         len = strlen (current->import_dest);
>>         directive_match = (search_parents
>> diff --git a/gdb/namespace.c b/gdb/namespace.c
>> index 5b5749369ee..d81365a284a 100644
>> --- a/gdb/namespace.c
>> +++ b/gdb/namespace.c
>> @@ -113,8 +113,13 @@ using_direct::valid_line (unsigned int boundary) 
>> const
>>       {
>>         CORE_ADDR curr_pc = get_frame_pc (get_selected_frame (nullptr));
>>         symtab_and_line curr_sal = find_sal_for_pc (curr_pc, 0);
>> -      return (decl_line <= curr_sal.line)
>> -         || (decl_line >= boundary);
>> +
>> +      /* Apply GCC PR debug/108716 workaround.  */
>> +      if (boundary != 0
>> +      && decl_line >= boundary)
>> +    return true;
>> +
>> +      return decl_line <= curr_sal.line;
>>       }
>>     catch (const gdb_exception &ex)
>>       {
>> diff --git a/gdb/testsuite/gdb.cp/ns-before-using.c b/gdb/testsuite/ 
>> gdb.cp/ns-before-using.c
>> new file mode 100644
>> index 00000000000..f74c35b0d3f
>> --- /dev/null
>> +++ b/gdb/testsuite/gdb.cp/ns-before-using.c
>> @@ -0,0 +1,37 @@
>> +/* This testcase is part of GDB, the GNU debugger.
>> +
>> +   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/>.  */
>> +
>> +namespace mod_a
>> +{
>> +  int xxx = 10;
>> +}
>> +
>> +static void
>> +foo ()
>> +{
>> +}
>> +
>> +int
>> +main ()
>> +{
>> +  {
>> +    foo ();
>> +    using namespace mod_a;
>> +  }
>> +
>> +  return mod_a::xxx;
>> +}
>> diff --git a/gdb/testsuite/gdb.cp/ns-before-using.exp b/gdb/testsuite/ 
>> gdb.cp/ns-before-using.exp
>> new file mode 100644
>> index 00000000000..f891cf56c49
>> --- /dev/null
>> +++ b/gdb/testsuite/gdb.cp/ns-before-using.exp
>> @@ -0,0 +1,44 @@
>> +# 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/>.
>> +
>> +standard_testfile .c
>> +
>> +if {[prepare_for_testing "failed to prepare" $testfile $srcfile \
>> +     {debug c++}]} {
>> +    return
>> +}
>> +
>> +if {![runto_main]} {
>> +    continue
>> +}
>> +
>> +# Xfail for incorrect decl_line on DW_TAG_imported_module,
>> +# GCC PR debug/108716.
>> +set have_gcc108716_xfail \
>> +    [expr {[test_compiler_info gcc-*] && [gcc_major_version] < 13}]
>> +
>> +# Regression test for PR exp/34203.
>> +set re_pass [string_to_regexp {No symbol "xxx" in current context.}]
>> +gdb_test_multiple "print xxx" "" {
>> +    -re -wrap $re_pass {
>> +    pass $gdb_test_name
>> +    }
>> +    -re -wrap "$valnum_re = 10" {
>> +    if {$have_gcc108716_xfail} {
>> +        setup_xfail *-*-* gcc/108716
>> +    }
>> +    fail $gdb_test_name
>> +    }
>> +}
>>
>> base-commit: bd64797371d27c766d551d0bf115d9090f1d0594
> 


      reply	other threads:[~2026-07-13 16:17 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-06-04 16:21 [PATCH] " Tom de Vries
2026-06-18  7:53 ` [PING][PATCH] " Tom de Vries
2026-07-13 16:17   ` Tom de Vries [this message]

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=00c8dabe-0035-44d1-911c-1cb041bef589@suse.de \
    --to=tdevries@suse.de \
    --cc=gdb-patches@sourceware.org \
    /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