public inbox for libstdc++-cvs@sourceware.org
help / color / mirror / Atom feed
From: Jonathan Wakely <redi@gcc.gnu.org>
To: gcc-cvs@gcc.gnu.org, libstdc++-cvs@gcc.gnu.org
Subject: [gcc r13-4091] libstdc++: Fix std::any pretty printer
Date: Wed, 16 Nov 2022 13:37:56 +0000 (GMT)	[thread overview]
Message-ID: <20221116133756.CC86C395A07B@sourceware.org> (raw)

https://gcc.gnu.org/g:3c54805d03ac1bcc3d8547ffb5e6c4e1f301a7a2

commit r13-4091-g3c54805d03ac1bcc3d8547ffb5e6c4e1f301a7a2
Author: Jonathan Wakely <jwakely@redhat.com>
Date:   Wed Nov 16 12:22:04 2022 +0000

    libstdc++: Fix std::any pretty printer
    
    The recent changes to FilteringTypePrinter affect the result of
    gdb.lookup_type('std::string') in StdExpAnyPrinter, causing it to always
    return the std::__cxx11::basic_string specialization. This then causes a
    gdb.error exception when trying to lookup the std::any manager type for
    a specliaization using that string, but that manager was never
    instantiated in the program. This causes FAILs when running the tests
    with -D_GLIBCXX_USE_CXX11_ABI=0:
    
    FAIL: libstdc++-prettyprinters/libfundts.cc print as
    FAIL: libstdc++-prettyprinters/libfundts.cc print as
    
    The ugly solution used in this patch is to repeat the lookup for every
    type that std::string could be a typedef for, and hope it only works for
    one of them.
    
    libstdc++-v3/ChangeLog:
    
            * python/libstdcxx/v6/printers.py (StdExpAnyPrinter): Make
            expansion of std::string in manager name more robust.

Diff:
---
 libstdc++-v3/python/libstdcxx/v6/printers.py | 31 +++++++++++++++++++++++++---
 1 file changed, 28 insertions(+), 3 deletions(-)

diff --git a/libstdc++-v3/python/libstdcxx/v6/printers.py b/libstdc++-v3/python/libstdcxx/v6/printers.py
index 19c70d12035..1abf0a4bce3 100644
--- a/libstdc++-v3/python/libstdcxx/v6/printers.py
+++ b/libstdc++-v3/python/libstdcxx/v6/printers.py
@@ -1272,9 +1272,34 @@ class StdExpAnyPrinter(SingleObjContainerPrinter):
             mgrname = m.group(1)
             # FIXME need to expand 'std::string' so that gdb.lookup_type works
             if 'std::string' in mgrname:
-                mgrname = re.sub("std::string(?!\w)", str(gdb.lookup_type('std::string').strip_typedefs()), m.group(1))
-
-            mgrtype = gdb.lookup_type(mgrname)
+                # This lookup for std::string might return the __cxx11 version,
+                # but that's not necessarily the one used by the std::any
+                # manager function we're trying to find.
+                strings = {str(gdb.lookup_type('std::string').strip_typedefs())}
+                # So also consider all the other possible std::string types!
+                s = 'basic_string<char, std::char_traits<char>, std::allocator<char> >'
+                quals = ['std::', 'std::__cxx11::', 'std::' + _versioned_namespace]
+                strings |= {q+s for q in quals} # set of unique strings
+                mgrtypes = []
+                for s in strings:
+                    try:
+                        x = re.sub("std::string(?!\w)", s, m.group(1))
+                        # The following lookup might raise gdb.error if the
+                        # manager function was never instantiated for 's' in the
+                        # program, because there will be no such type.
+                        mgrtypes.append(gdb.lookup_type(x))
+                    except gdb.error:
+                        pass
+                if len(mgrtypes) != 1:
+                    # FIXME: this is unlikely in practice, but possible for
+                    # programs that use both old and new string types with
+                    # std::any in a single program. Can we do better?
+                    # Maybe find the address of each type's _S_manage and
+                    # compare to the address stored in _M_manager?
+                    raise ValueError('Cannot uniquely determine std::string type used in std::any')
+                mgrtype = mgrtypes[0]
+            else:
+                mgrtype = gdb.lookup_type(mgrname)
             self.contained_type = mgrtype.template_argument(0)
             valptr = None
             if '::_Manager_internal' in mgrname:

                 reply	other threads:[~2022-11-16 13:37 UTC|newest]

Thread overview: [no followups] expand[flat|nested]  mbox.gz  Atom feed

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=20221116133756.CC86C395A07B@sourceware.org \
    --to=redi@gcc.gnu.org \
    --cc=gcc-cvs@gcc.gnu.org \
    --cc=libstdc++-cvs@gcc.gnu.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).