[Aptitude-devel] Bug#607118: Bug#607118: Bug#607118: aptitude: after 'q' is cancelled with ctrl-q, it doesn't work again

[Manuel A. Fernandez Montecelo]

Control: forcemerge 607118 658271 681023
Control: tags 607118 + pending


(reorganising the content of previous messages...)

2015-09-07 22:50 To Axel Beckert:
>2015-09-07 10:02 Axel Beckert:
>>>I cannot reproduce this with 0.7.1-1,
>>
>>I can.
>>
>>>Unless it is a difference between i386 and amd64 systems, I wonder if it
>>>was just a temporary glitch, or if it was fixed in some way.
>>
>>I can reproduce it on an uptodate Sid amd64.
>
>That's odd, I cannot reproduce it (also unstable and amd64).
>
>I press 'q', dialog appears, with Ctrl-g disappears, with 'q' appears
>again, repeatedly without problems; and I don't get missing key presses
>when moving to Yes/No in the dialog buttons or any other observable
>misbehaviour.

I think that the difference was that you have this option set to true:

  Aptitude::UI::Minibuf-Prompts "true";

I thought that it was strange that the string was reported as "Really
quit Aptitude? [n]y", with the replies in that order instead of "[Yes]
[No]", but it didn't occured to me that the minibuf option was at play
at first, because I haven't used it recently.

Anyway, when I realised and enabled this option, I was able to reproduce
the problem in #607118 and then fix it.

At least the original reporter of #658271 (Matthew) and Axel also had
this option enabled, I don't know about Jö.


>>Control: tag -1 = confirmed
>>Control: found -1 0.7.1-1
>>Control: retitle -1 aptitude: after 'q' is cancelled with ctrl-g once, it doesn't work again
>>
>>Hi,
>>
>>Manuel A. Fernandez Montecelo wrote:
>>>2010-12-14 20:22 Andrew Pimlott:
>>>> Package: aptitude
>>>> Version: 0.6.3-3.2
>>>> Severity: minor
>>>>
>>>> Start aptitude.  Hit 'q', and the dialog "Really quit Aptitude? [n]y"
>>>> appears.  Hit <ctrl-g> to cancel, and the dialog is dismissed.  Hit 'q'
>>>> again, and nothing happens.
>>
>>Thanks a lot for that hint! I wondered for years what's the condition
>>under which "q" stops working. You found it!

I think that I found what the problem is, read above, bug #607118.

There are also the bugs #658271 and #681023, which are themselves
merged.

#658271 is not specific about how to reproduce it, but the description
of the problem in the original message of the bug report matches the
discussion in #607118.  The steps of Axel to try to reproduce it are
more circuitous, but I think that the fix should work for these cases,
so I merged those, and marked all three as 'pending'.

It woud be great if you can try to confirm if this is fixed once the
next version is released.


Cheers.
-- 
Manuel A. Fernandez Montecelo <manuel.montezelo at gmail.com>