From: "dodji at redhat dot com" <sourceware-bugzilla@sourceware.org>
To: libabigail@sourceware.org
Subject: [Bug default/26684] abidiff says that some bit field elements missing in version of binary with dwarf5 debuginfo.
Date: Fri, 23 Oct 2020 11:29:01 +0000 [thread overview]
Message-ID: <bug-26684-9487-gvscRtjrDY@http.sourceware.org/bugzilla/> (raw)
In-Reply-To: <bug-26684-9487@http.sourceware.org/bugzilla/>
https://sourceware.org/bugzilla/show_bug.cgi?id=26684
--- Comment #16 from dodji at redhat dot com ---
"fche at redhat dot com" <sourceware-bugzilla@sourceware.org> writes:
> --- Comment #15 from Frank Ch. Eigler <fche at redhat dot com> ---
Thank you Frank for taking the time to reply to this!
> Just some vague opinion from the backward compatibility peanut galary:
>
> Was the misinterpretation ambiguous / information-losing?
Sadly, I am afraid it was :-( (unless I am missing something).
For little endian machines, libabigail was wrongly interpreting
DW_AT_bit_offset as being the same value as for big endian machines. So
the resulting data member offset we were getting is wrong. And that
wrong number is what's saved in the abixml file.
At the abixml level, we don't save the other low level details of the
bitfield layout that would help us convert that buggy data member offset
into the correct one. We just save the offset of the data member and
its type. Now in hindsight, I am thinking maybe I should have save some
of that information ...
> If not, then it does not seem necessary to do a major flag-day version
> bump, because abidiff could read a v1 xml file, convert from its buggy
> bit numbering scheme to the correct one. It could treat the data (or
> even rewrite it) as though it were a v2 correct one.
> Heck, if the buggy data is straightforward to correct in xml, could
> put the correct bit# into a new-name xml element/attribute nearby,
> rather than requiring a top level version bump?
Yes, it would have been super cool to be able to do that ...
--
You are receiving this mail because:
You are on the CC list for the bug.
next prev parent reply other threads:[~2020-10-23 11:29 UTC|newest]
Thread overview: 19+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-09-30 18:18 [Bug default/26684] New: " wcohen at redhat dot com
2020-09-30 18:19 ` [Bug default/26684] " wcohen at redhat dot com
2020-09-30 18:20 ` wcohen at redhat dot com
2020-09-30 18:21 ` wcohen at redhat dot com
2020-09-30 18:22 ` wcohen at redhat dot com
2020-10-01 16:01 ` mark at klomp dot org
2020-10-01 17:00 ` wcohen at redhat dot com
2020-10-01 20:50 ` mark at klomp dot org
2020-10-01 20:54 ` mark at klomp dot org
2020-10-01 21:55 ` mark at klomp dot org
2020-10-02 10:07 ` dodji at redhat dot com
2020-10-02 14:42 ` mark at klomp dot org
2020-10-02 14:46 ` mark at klomp dot org
2020-10-02 14:52 ` mark at klomp dot org
2020-10-20 10:26 ` dodji at redhat dot com
2020-10-23 8:24 ` dodji at redhat dot com
2020-10-23 10:35 ` fche at redhat dot com
2020-10-23 11:29 ` dodji at redhat dot com [this message]
2023-01-01 18:09 ` dodji at redhat 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-26684-9487-gvscRtjrDY@http.sourceware.org/bugzilla/ \
--to=sourceware-bugzilla@sourceware.org \
--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).