public inbox for libabigail@sourceware.org
 help / color / mirror / Atom feed
From: Dodji Seketeli <dodji@redhat.com>
To: libabigail@sourceware.org
Subject: [PATCH 0/3, applied] Bug 29857 - Fix comparing binary with decl-only unions & cycles
Date: Mon, 19 Dec 2022 18:07:13 +0100	[thread overview]
Message-ID: <87wn6ne30e.fsf@redhat.com> (raw)

Hello,

Fixing the problem reported at
https://sourceware.org/bugzilla/show_bug.cgi?id=29857 involves
addressing several issues.

The first patch of this series addresses the fact that in some cases,
when comparing decl-only types, the comparison engine can pop
comparison operands that haven't been pushed to the comparison stack,
leading to some assert violation.

Then, I realized that several occurences of a given type (due to
DWARF-related type duplication) are considered different because of
comparing decl-only unions against their definition; this is because
decl-only unions are not yet resolved to their definition, unlike what
is done already for classes and enums.  The second patch adds support
for resolving decl-only unions to their definition.  This drastically
reduces the amount of spurious type changes that were causing lots of
issues down the road.

Fixing that second issue uncovered a third one which is basically
related to undetected cycles in the type graph, while comparing types
for canonicalization purposes.  Those undected cycles are leading to
infinite loops during structural comparison of types.  This is fixed
in the third patch.

Together, these three patches should fix bugs 29857 and 29901.

Dodji Seketeli (3):
  Bug 29857 - Don't pop comparison operands that haven't been pushed
  Bug 29857 - dwarf-reader: Resolve decl-only unions
  Bug 29857 - Better detect comparison cycles in type graph

 include/abg-fwd.h                             |    9 +
 src/abg-dwarf-reader.cc                       |   60 +-
 src/abg-ir-priv.h                             |   94 +-
 src/abg-ir.cc                                 |   59 +-
 ...19-pr19023-libtcmalloc_and_profiler.so.abi | 9003 +++++++++--------
 ...19-pr19023-libtcmalloc_and_profiler.so.abi | 8969 ++++++++--------
 6 files changed, 9100 insertions(+), 9094 deletions(-)

-- 
2.31.1


-- 
		Dodji


             reply	other threads:[~2022-12-19 17:07 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-12-19 17:07 Dodji Seketeli [this message]
2022-12-19 17:11 ` [PATCH 1/3, applied] Bug 29857 - Don't pop comparison operands that haven't been pushed Dodji Seketeli
2022-12-19 17:12 ` [PATCH 2/3, applied] Bug 29857 - dwarf-reader: Resolve decl-only unions Dodji Seketeli
2022-12-19 17:14 ` [PATCH 3/3, applied] Bug 29857 - Better detect comparison cycles in type graph Dodji Seketeli

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=87wn6ne30e.fsf@redhat.com \
    --to=dodji@redhat.com \
    --cc=libabigail@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).