fsanitize=bounds-strict caught potential error

fsanitize=bounds-strict caught potential error

[Po Lu]

I don't know if this message is correct or not, but I got

xdisp.c:30394:37: runtime error: index -1 out of bounds for type 'iterator_stack_entry [5]'

After displaying the Emacs manual and building with
`-fsanitize=undefined,bounds-strict'.

Thanks.

Re: fsanitize=bounds-strict caught potential error

[Eli Zaretskii]

> From: Po Lu <luangruo@yahoo.com>
> Date: Sun, 05 Dec 2021 21:04:47 +0800
> 
> I don't know if this message is correct or not, but I got
> 
> xdisp.c:30394:37: runtime error: index -1 out of bounds for type 'iterator_stack_entry [5]'

The compiler doesn't know enough to realize that this code can never
be executed with it->sp < 1.  This code is there since 2003, and never
gave us any trouble.

Re: fsanitize=bounds-strict caught potential error

[Andreas Schwab]

On Dez 05 2021, Eli Zaretskii wrote:

>> From: Po Lu <luangruo@yahoo.com>
>> Date: Sun, 05 Dec 2021 21:04:47 +0800
>> 
>> I don't know if this message is correct or not, but I got
>> 
>> xdisp.c:30394:37: runtime error: index -1 out of bounds for type 'iterator_stack_entry [5]'
>
> The compiler doesn't know enough to realize that this code can never
> be executed with it->sp < 1.  This code is there since 2003, and never
> gave us any trouble.

It is a runtime error, thus the erroneous access is actually being
executed.

Andreas.

-- 
Andreas Schwab, schwab@linux-m68k.org
GPG Key fingerprint = 7578 EB47 D4E5 4D69 2510  2552 DF73 E780 A9DA AEC1
"And now for something completely different."

Re: fsanitize=bounds-strict caught potential error

[Eli Zaretskii]

> From: Andreas Schwab <schwab@linux-m68k.org>
> Cc: Po Lu <luangruo@yahoo.com>,  emacs-devel@gnu.org
> Date: Sun, 05 Dec 2021 16:15:15 +0100
> 
> >> xdisp.c:30394:37: runtime error: index -1 out of bounds for type 'iterator_stack_entry [5]'
> >
> > The compiler doesn't know enough to realize that this code can never
> > be executed with it->sp < 1.  This code is there since 2003, and never
> > gave us any trouble.
> 
> It is a runtime error, thus the erroneous access is actually being
> executed.

Then let's put an assertion there and find the culprit.  Maybe it's
the min-width feature, which is new on master.

Re: fsanitize=bounds-strict caught potential error

[Andreas Schwab]

On Dez 05 2021, Eli Zaretskii wrote:

>> From: Po Lu <luangruo@yahoo.com>
>> Date: Sun, 05 Dec 2021 21:04:47 +0800
>> 
>> I don't know if this message is correct or not, but I got
>> 
>> xdisp.c:30394:37: runtime error: index -1 out of bounds for type 'iterator_stack_entry [5]'
>
> The compiler doesn't know enough to realize that this code can never
> be executed with it->sp < 1.  This code is there since 2003, and never
> gave us any trouble.

If you put a breakpoint there, you will see it.sp == 0 most of the time
actually.

Andreas.

-- 
Andreas Schwab, schwab@linux-m68k.org
GPG Key fingerprint = 7578 EB47 D4E5 4D69 2510  2552 DF73 E780 A9DA AEC1
"And now for something completely different."

Re: fsanitize=bounds-strict caught potential error

[Andreas Schwab]

On Dez 05 2021, Po Lu wrote:

> I don't know if this message is correct or not, but I got
>
> xdisp.c:30394:37: runtime error: index -1 out of bounds for type 'iterator_stack_entry [5]'

This happens when called by display_min_width.

Andreas.

-- 
Andreas Schwab, schwab@linux-m68k.org
GPG Key fingerprint = 7578 EB47 D4E5 4D69 2510  2552 DF73 E780 A9DA AEC1
"And now for something completely different."

Re: fsanitize=bounds-strict caught potential error

[Eli Zaretskii]

> Date: Sun, 05 Dec 2021 17:23:30 +0200
> From: Eli Zaretskii <eliz@gnu.org>
> Cc: luangruo@yahoo.com, emacs-devel@gnu.org
> 
> > It is a runtime error, thus the erroneous access is actually being
> > executed.
> 
> Then let's put an assertion there and find the culprit.  Maybe it's
> the min-width feature, which is new on master.

Yep, that it is.  Should be fixed now.