Save on save

[racoon]

On 2020-03-28 20:16, Richard Kimberly Heck wrote:
> On 3/28/20 3:00 PM, Richard Kimberly Heck wrote:
>> On 3/28/20 2:55 PM, Daniel wrote:
>>> On 2020-03-28 19:47, Richard Kimberly Heck wrote:
>>>> On 3/28/20 5:17 AM, Daniel wrote:
>>>>> On 2020-03-28 04:27, Richard Kimberly Heck wrote:
>>>>>> On 3/27/20 2:48 PM, racoon wrote:
>>>>>>> On 2020-03-27 19:25, Richard Kimberly Heck wrote:
>>>>>>>> On 3/27/20 2:21 PM, racoon wrote:
>>>>>>>>> On 2020-03-27 17:52, racoon wrote:
>>>>>>>>>> On 2020-03-27 17:41, Richard Kimberly Heck wrote:
>>>>>>>>>>> On 3/27/20 4:21 AM, Daniel wrote:
>>>>>>>>>>>> On 2020-03-19 15:03, racoon wrote:
>>>>>>>>>>>>> On 2020-03-19 14:53, Richard Kimberly Heck wrote:
>>>>>>>>>>>>>> Yes, you could assign something like "command-sequence
>>>>>>>>>>>>>> self-insert s;
>>>>>>>>>>>>>> char-delete-backward; buffer-write" to Ctrl-S. Undo won't
>>>>>>>>>>>>>> work,
>>>>>>>>>>>>>> because
>>>>>>>>>>>>>> then the document isn't dirty again.
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> Riki
>>>>>>>>>>>>>>
>>>>>>>>>>>>>>
>>>>>>>>>>>>> Unfortunately, when the settings dialog is closed and re-opened
>>>>>>>>>>>>> everything that comes after the first semi-colon is chopped
>>>>>>>>>>>>> off. Is
>>>>>>>>>>>>> that
>>>>>>>>>>>>> a bug?
>>>>>>>>>>>> Oddly enough, it is possible to use similar command sequences.
>>>>>>>>>>>> So, the
>>>>>>>>>>>> problem isn't general. For example,
>>>>>>>>>>>>
>>>>>>>>>>>> command-sequence self-insert .; space-insert normal
>>>>>>>>>>>>
>>>>>>>>>>>> just works fine (see
>>>>>>>>>>>> https://www.lyx.org/trac/ticket/11798#comment:6).
>>>>>>>>>>>> I don't know what the difference might be.
>>>>>>>>>>> Don't see it here.
>>>>>>>>>>>
>>>>>>>>>>> You can put it manually into your user.bind file if need be.
>>>>>>>>>>>
>>>>>>>>>>> Riki
>>>>>>>>>> Thanks, that helped. And I figured out what the problem was. I
>>>>>>>>>> copied
>>>>>>>>>> the command from your email not knowing that my email program had
>>>>>>>>>> inserted a line-break after the first command in the sequence.
>>>>>>>>>> Unfortunately, the input box in the shortcut editor masked this. I
>>>>>>>>>> guess
>>>>>>>>>> it would be better if text pasted into the input box was
>>>>>>>>>> cleaned of
>>>>>>>>>> characters it does not support rather than just hiding them.
>>>>>>>>>>
>>>>>>>>>> Daniel
>>>>>>>>>
>>>>>>>>> Seems like at least this particular command sequence is not such a
>>>>>>>>> good
>>>>>>>>> idea. I just realized that it wrecks havoc if there is an active
>>>>>>>>> selection of text because the text gets removed.
>>>>>>>> Try adding "escape" first. That clears the selection. Of course, you
>>>>>>>> lose the selection.
>>>>>>>>
>>>>>>>> Riki
>>>>>>> Thanks. That works better. Yes, losing the selection isn't perfect.
>>>>>>>
>>>>>>> Btw. the action input field suffers from the same problem as the
>>>>>>> shortcut editor, e.g. it masks line-breaks which make a command fail.
>>>>>> I don't know if there is any easy solution to this. I think I'd regard
>>>>>> it as a Qt bug.
>>>>>>
>>>>>> Riki
>>>>> Yes, might be a bug. I have checked the input fields in Scribus and
>>>>> they
>>>>> exhibit the same problem.
>>>>>
>>>>> However, as far as I understood, there is a function to determine which
>>>>> characters are valid input. Maybe that helps? Here is an example:
>>>>>
>>>>> https://doc.qt.io/qt-5/qregexpvalidator.html#details
>>>>>
>>>>> Maybe one could set the validator to a regex that does not match any
>>>>> newline (\n) or return (\r)? But I don't know enough about Qt or regex.
>>>> It wouldn't really help. That would prevent you from actually saving the
>>>> shortcut, but you'd be totally stumped why. (Well, not you, but most
>>>> users.)
>>>>
>>>> Riki
>>> Okay, then I seem to have misunderstood what the validator is doing.
>>>
>>> "Sets this line edit to only accept input that the validator, v, will
>>> accept." (https://doc.qt.io/qt-5/qlineedit.html#setValidator)
>>>
>>> That sounded to me as if the *input* would be limited. And I'd hoped
>>> that on paste the characters would be stripped (or the string cut off at
>>> the first occurrence).
>> Yes, sorry. In that case, it wouldn't let you paste it at all. Maybe
>> that's better, though.
>
> Hmm. In some cases, it won't let you paste, but using our
> NoNewLineValidator, it does, and it just removes the CR.
>
> Riki

Sounds promising, or? I hope there is a way to apply it across the board
to all qLineEdits. Seems like an oversight that Qt hasn't set it as default.

Daniel