bug#19102: 24.4; outline-move-subtree-up/down error at last and second-last subtree

[Stephen Berman]

On Thu, 20 Nov 2014 16:46:22 +1000 Paul Rankin <paul@tilk.co> wrote:

> Stephen Berman <stephen.berman@gmx.net> writes:
>
>> On Wed, 19 Nov 2014 22:32:06 +0200 Eli Zaretskii <eliz@gnu.org> wrote:
>>
>>>> From: Stephen Berman <stephen.berman@gmx.net>
>>>> Cc: paul@tilk.co,  19102@debbugs.gnu.org
>>>> Date: Wed, 19 Nov 2014 21:14:42 +0100
>>>> 
>>>> > Wouldn't using eolp instead of the comparison solve that problem more
>>>> > easily?
>>>> 
>>>> Well, that eliminates the wrong-type-argument error in the current code,
>>>> but it instead signals "End of buffer"
>>>
>>> From forward-char?  If so, you could avoid the call if eobp.  Or wrap
>>> the call in condition-case and ignore errors.
>>
>> Those solutions work only if there's an empty line after the last
>> subtree.  If there isn't, I can't see any way other than my patch.
>>
>>>> Do you see any other problems with the patch or more room for
>>>> improvement?
>>>
>>> I was worried by the complexity of maybe-forward-char, but maybe now
>>> it is much simpler.
>>
>> I simplified it just a bit more, but that's all I can think of.  I also
>> had overlooked a side-effect of the error signaled when trying to move
>> the last subtree down: it leaves a dangling newline; this is avoided
>> using with-demoted-errors.  The revised patch is appended.
>>
>> Steve Berman
>
> Sorry if there is some very obvious reason I'm missing, but can't we
> just change
>
>     (= (char-after) ?\n)
>
> to
>
>     (eq (char-after) ?\n)
>
> ...?

This has the same effect as Eli's suggestion of not calling forward-char
at eob or wrapping it in condition-case: outline-move-subtree-up/down
works if there's an empty line after the last subtree, but if not, it
puts the last two subtree headers on the same line.

Steve Berman