bug#31274: 27.0.50; xdisp.c:7575: Emacs fatal error: assertion failed: IT_BYTEPOS (*it) == CHAR_TO_BYTE (IT_CHARPOS (*it))

bug#31274: 27.0.50; xdisp.c:7575: Emacs fatal error: assertion failed: IT_BYTEPOS (*it) == CHAR_TO_BYTE (IT_CHARPOS (*it))

[Stefan Monnier]

Package: Emacs
Version: 27.0.50


I can reproduce the above assertion failure when I do the following:

    % src/emacs -Q /usr/bin/perl -l .../elpa/packages/nhexl-mode/nhexl-mode.el -f nhexl-mode --eval '(setq word-wrap t)'
    y
    M->

The `y` is because nhexl-mode prompts the user to convert the buffer
to unibyte and is not directly relevant: you can use find-file-literally
instead and the result is the same.  I used /usr/bin/perl in the above
example, but I could reproduce the same crash with "any" binary
executable such as /bin/gzip and src/emacs (tho with some it doesn't
crash immediately).

Obviously, word-wrap in a binary buffer is not very meaningful, but
that doesn't justify a discrepancy between charpos and bytepos.


        Stefan

bug#31274: 27.0.50; xdisp.c:7575: Emacs fatal error: assertion failed: IT_BYTEPOS (*it) == CHAR_TO_BYTE (IT_CHARPOS (*it))

[Eli Zaretskii]

> From: Stefan Monnier <monnier@IRO.UMontreal.CA>
> Date: Thu, 26 Apr 2018 15:54:55 -0400
> 
> Package: Emacs
> Version: 27.0.50
> 
> I can reproduce the above assertion failure when I do the following:
> 
>     % src/emacs -Q /usr/bin/perl -l .../elpa/packages/nhexl-mode/nhexl-mode.el -f nhexl-mode --eval '(setq word-wrap t)'
>     y
>     M->

I cannot reproduce this, with today's master of Emacs and of ELPA.  I
tried on GNU/Linux (in -nw) and on MS-Windows (in a GUI session), and
both didn't trigger the assertion (tried with 3 executables, including
those you mentioned).  I'm guessing some particular byte sequence
present in your binaries triggers this.  So please provide more
details.

P.S.  I frequently wonder why veteran experienced users omit crucial
information from their bug reports, such as the C backtrace in this
case, that is very easy for them to collect.  Please don't ever assume
that something which is 100% reproducible on your system can be
similarly easily reproduced on any other system, and that therefore
those details are an unnecessary luxury.

bug#31274: 27.0.50; xdisp.c:7575: Emacs fatal error: assertion failed: IT_BYTEPOS (*it) == CHAR_TO_BYTE (IT_CHARPOS (*it))

[Eli Zaretskii]

> Date: Sat, 28 Apr 2018 12:29:47 +0300
> From: Eli Zaretskii <eliz@gnu.org>
> Cc: 31274@debbugs.gnu.org
> 
> I'm guessing some particular byte sequence present in your binaries
> triggers this.

Or maybe you have local changes that are responsible.

> P.S.  I frequently wonder why veteran experienced users omit crucial
> information                   ^^^^^^^^^^^^^^^^^^^^^^^^^

Let alone veteran experienced developers.

bug#31274: 27.0.50; xdisp.c:7575: Emacs fatal error: assertion failed: IT_BYTEPOS (*it) == CHAR_TO_BYTE (IT_CHARPOS (*it))

[Noam Postavsky]

Stefan Monnier <monnier@IRO.UMontreal.CA> writes:

> I can reproduce the above assertion failure when I do the following:
>
>     % src/emacs -Q /usr/bin/perl -l .../elpa/packages/nhexl-mode/nhexl-mode.el -f nhexl-mode --eval '(setq word-wrap t)'
>     y
>     M->
>
> The `y` is because nhexl-mode prompts the user to convert the buffer
> to unibyte and is not directly relevant: you can use find-file-literally
> instead and the result is the same.

I couldn't reproduce the assert failure, and I didn't get the unibyte
prompt either.

bug#31274: 27.0.50; xdisp.c:7575: Emacs fatal error: assertion failed: IT_BYTEPOS (*it) == CHAR_TO_BYTE (IT_CHARPOS (*it))

[Eli Zaretskii]

> From: Noam Postavsky <npostavs@gmail.com>
> Date: Fri, 04 May 2018 06:35:22 -0400
> Cc: 31274@debbugs.gnu.org
> 
> >     % src/emacs -Q /usr/bin/perl -l .../elpa/packages/nhexl-mode/nhexl-mode.el -f nhexl-mode --eval '(setq word-wrap t)'
> >     y
> >     M->
> >
> > The `y` is because nhexl-mode prompts the user to convert the buffer
> > to unibyte and is not directly relevant: you can use find-file-literally
> > instead and the result is the same.
> 
> I couldn't reproduce the assert failure, and I didn't get the unibyte
> prompt either.

I think the unibyte prompt may or may not be seen, it depends on
whether Emacs succeeds to decide the file is a binary file, and uses
unibyte to begin with.  And whether it decides that depends on the
contents of the file, so it could well be that you don't see it
(neither did I, FWIW).

Or maybe this is the sign of something special that happens on
Stefan's system, like local changes perhaps?

In any case, we need either a reproducer or a detailed backtrace to
make any progress with this bug.

Thanks.

bug#31274: 27.0.50; xdisp.c:7575: Emacs fatal error: assertion failed: IT_BYTEPOS (*it) == CHAR_TO_BYTE (IT_CHARPOS (*it))

[Lars Ingebrigtsen]

Eli Zaretskii <eliz@gnu.org> writes:

> In any case, we need either a reproducer or a detailed backtrace to
> make any progress with this bug.

More information was requested, but no response was given within a
year, so I'm closing this bug report.  If the problem still exists,
please reopen this bug report.

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no