Another todo-mode bug fix, ok for emacs-25?

Another todo-mode bug fix, ok for emacs-25?

[Stephen Berman]

In bug#23447 I show how to reproduce a todo-mode bug I encountered
involving archived items and provide a fix for it.  The bug does not
result in data loss or file corruption, but changes the behavior of
todo-show in a way that can only be repaired by resetting internal
variables or unloading and reloading todo-mode.el.  The patch has no
effect outside of todo-mode, and I've tested all uses of the function
involved within todo-mode and encountered no problems, so is it all
right for me to commit the fix to emacs-25?

Steve Berman

Re: Another todo-mode bug fix, ok for emacs-25?

[Eli Zaretskii]

> From: Stephen Berman <stephen.berman@gmx.net>
> Date: Wed, 04 May 2016 15:59:11 +0200
> 
> In bug#23447 I show how to reproduce a todo-mode bug I encountered
> involving archived items and provide a fix for it.  The bug does not
> result in data loss or file corruption, but changes the behavior of
> todo-show in a way that can only be repaired by resetting internal
> variables or unloading and reloading todo-mode.el.  The patch has no
> effect outside of todo-mode, and I've tested all uses of the function
> involved within todo-mode and encountered no problems, so is it all
> right for me to commit the fix to emacs-25?

We don't want to release Emacs 25.1 with a buggy todo-mode, so the
fact the change is local helps, but not enough.  Please try to
convince us that the change itself is safe.  For example, is the
affected code called in different use cases than the one you anted to
fix, and if so, did you make sure those other use cases still work OK?

Thanks.

Re: Another todo-mode bug fix, ok for emacs-25?

[Stephen Berman]

On Wed, 04 May 2016 18:18:39 +0300 Eli Zaretskii <eliz@gnu.org> wrote:

>> From: Stephen Berman <stephen.berman@gmx.net>
>> Date: Wed, 04 May 2016 15:59:11 +0200
>> 
>> In bug#23447 I show how to reproduce a todo-mode bug I encountered
>> involving archived items and provide a fix for it.  The bug does not
>> result in data loss or file corruption, but changes the behavior of
>> todo-show in a way that can only be repaired by resetting internal
>> variables or unloading and reloading todo-mode.el.  The patch has no
>> effect outside of todo-mode, and I've tested all uses of the function
>> involved within todo-mode and encountered no problems, so is it all
>> right for me to commit the fix to emacs-25?
>
> We don't want to release Emacs 25.1 with a buggy todo-mode, so the
> fact the change is local helps, but not enough.  Please try to
> convince us that the change itself is safe.  For example, is the
> affected code called in different use cases than the one you anted to
> fix, and if so, did you make sure those other use cases still work OK?

The code changes are confined to the function todo-jump-to-category,
which can be invoked interactively and is also called noninteractively
from two other functions, todo-jump-to-archive-category and
todo-insert-category-line.  The last of these uses is what give rise to
the bug, under the conditions given in the bug report (archived items,
non-nil todo-skip-archived-categories).  With the fix applied, I've also
tested this use when these conditions are not met, and I've tested the
other noninteractive use and the interactive uses (the command can be
called not only within Todo mode but also from a non-Todo mode buffer).
In all of my tests with the patched code, the behavior was as expected.
Unfortunately, I don't have a test suite for todo-mode (I hope to
develop one, but that's a long-term goal).

Steve Berman

Re: Another todo-mode bug fix, ok for emacs-25?

[Eli Zaretskii]

> From: Stephen Berman <stephen.berman@gmx.net>
> Cc: emacs-devel@gnu.org
> Date: Wed, 04 May 2016 19:46:27 +0200
> 
> The code changes are confined to the function todo-jump-to-category,
> which can be invoked interactively and is also called noninteractively
> from two other functions, todo-jump-to-archive-category and
> todo-insert-category-line.  The last of these uses is what give rise to
> the bug, under the conditions given in the bug report (archived items,
> non-nil todo-skip-archived-categories).  With the fix applied, I've also
> tested this use when these conditions are not met, and I've tested the
> other noninteractive use and the interactive uses (the command can be
> called not only within Todo mode but also from a non-Todo mode buffer).
> In all of my tests with the patched code, the behavior was as expected.

Thanks, I'm sold.

Re: Another todo-mode bug fix, ok for emacs-25?

[Stephen Berman]

On Wed, 04 May 2016 22:06:10 +0300 Eli Zaretskii <eliz@gnu.org> wrote:

>> From: Stephen Berman <stephen.berman@gmx.net>
>> Cc: emacs-devel@gnu.org
>> Date: Wed, 04 May 2016 19:46:27 +0200
>> 
>> The code changes are confined to the function todo-jump-to-category,
>> which can be invoked interactively and is also called noninteractively
>> from two other functions, todo-jump-to-archive-category and
>> todo-insert-category-line.  The last of these uses is what give rise to
>> the bug, under the conditions given in the bug report (archived items,
>> non-nil todo-skip-archived-categories).  With the fix applied, I've also
>> tested this use when these conditions are not met, and I've tested the
>> other noninteractive use and the interactive uses (the command can be
>> called not only within Todo mode but also from a non-Todo mode buffer).
>> In all of my tests with the patched code, the behavior was as expected.
>
> Thanks, I'm sold.

Thanks, committed as emacs-25 0932b94.

Steve Berman