Re: [PATCH] Fix the processing of Meta-key commands in TUI

[Patrick Palka <patrick@parcs.ath.cx>]

On Thu, Aug 28, 2014 at 7:31 AM, Pedro Alves <palves@redhat.com> wrote:
> On 08/22/2014 09:44 PM, Patrick Palka wrote:
>> +
>> +  if (async_command_editing_p && key_is_start_sequence (ch))
>
> I think the key_is_start_sequence check means that we'll
> fail to compensate in case the sequence if longer than
> 2 bytes?  That is, we'll compensate for the second char,
> but fail to compensate for the third, because by then,
> ch will not be a start sequence key.

Yes I think so.  I only took into account common 2-byte sequences such
as Alt_F.  I suppose that the check could be removed to account for
3+-byte key sequences too, but I haven't thought out the consequences
of such change.  And I'm not sure that readline uses any of such
sequences.

>
>> +    {
>> +      int ch_pending;
>> +
>> +      nodelay (w, TRUE);
>> +      ch_pending = wgetch (w);
>> +      nodelay (w, FALSE);
>> +
>
> Thanks,
> Pedro Alves
>