public inbox for gdb-prs@sourceware.org
help / color / mirror / Atom feed
From: "dje at google dot com" <sourceware-bugzilla@sourceware.org>
To: gdb-prs@sourceware.org
Subject: [Bug symtab/17591] New: dwarf2read.c:find_slot_in_mapped_hash needs to handle "(anonymous namespace)"
Date: Wed, 12 Nov 2014 22:31:00 -0000	[thread overview]
Message-ID: <bug-17591-4717@http.sourceware.org/bugzilla/> (raw)

https://sourceware.org/bugzilla/show_bug.cgi?id=17591

            Bug ID: 17591
           Summary: dwarf2read.c:find_slot_in_mapped_hash needs to handle
                    "(anonymous namespace)"
           Product: gdb
           Version: 7.8
            Status: NEW
          Severity: normal
          Priority: P2
         Component: symtab
          Assignee: unassigned at sourceware dot org
          Reporter: dje at google dot com

I was debugging a perf issue and found this in find_slot_in_mapped_hash:

  if (current_language->la_language == language_cplus
      || current_language->la_language == language_java
      || current_language->la_language == language_fortran)
    {
      /* NAME is already canonical.  Drop any qualifiers as .gdb_index does     
         not contain any.  */
      const char *paren = strchr (name, '(');

      if (paren)
        {
          char *dup;

          dup = xmalloc (paren - name + 1);
          memcpy (dup, name, paren - name);
          dup[paren - name] = 0;

          make_cleanup (xfree, dup);
          name = dup;
        }
    }

What if name is "(anonymous namespace)" ?

For the program at hand, this results in symbol table entry 0 which has 1085
entries.

[  0] :
        4 [static, function]
        5 [static, function]
        6 [static, function]
... some global functions
        2737 [static, function]
        2739 [static, function]
        2743 [static, function]


None of the CUs for these entries define anonymous namespace, but since we're
looking up in VAR_DOMAIN the index returns a match and we expand the CU.

While we should fix find_slot_in_mapped_hash, there is a side issue of doing a
general purpose lookup (VAR_DOMAIN) when we have (or should have) enough
knowledge to do a more specific lookup (we're looking for a namespace(/type)
not a function).

-- 
You are receiving this mail because:
You are on the CC list for the bug.


             reply	other threads:[~2014-11-12 22:31 UTC|newest]

Thread overview: 13+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-11-12 22:31 dje at google dot com [this message]
2014-11-12 22:38 ` [Bug symtab/17591] " dje at google dot com
2014-11-12 22:49 ` dje at google dot com
2014-11-12 22:50 ` dje at google dot com
2014-11-12 22:57 ` dje at google dot com
2014-11-13 20:02 ` dje at google dot com
2014-11-14  0:52 ` cvs-commit at gcc dot gnu.org
2014-11-14  0:59 ` dje at google dot com
2014-11-14 19:30 ` dje at google dot com
2014-12-03  0:25 ` cvs-commit at gcc dot gnu.org
2014-12-25  0:45 ` cvs-commit at gcc dot gnu.org
2015-02-13 19:15 ` dje at google dot com
2015-02-13 19:30 ` dje at google dot com

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=bug-17591-4717@http.sourceware.org/bugzilla/ \
    --to=sourceware-bugzilla@sourceware.org \
    --cc=gdb-prs@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;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).