D7660: Fix a regression caused by changing backspace key behavior

[Dominik Haumann]

dhaumann added a comment.


  The concept with the 'Set' flag is not yet correct.

INLINE COMMENTS

> kateconfig.cpp:1241
>      m_autoBracketsSet(false),
> +    m_backspaceRemoveComposed(false),
>      m_view(nullptr)

Should be: m_backspaceRemoveComposed_Set_(False)

> kateconfig.cpp:1288
>      m_autoBracketsSet(false),
> +    m_backspaceRemoveComposed(false),
>      m_view(view)

This should be: m_backspaceRemoveComposed_Set_(false)

> kateconfig.cpp:2221
> +{
> +    if (m_backspaceRemoveComposed || isGlobal()) {
> +        return m_backspaceRemoveComposed;

Wrong, correct is: m_backspaceRemoveComposed_Set_

> kateconfig.cpp:2235
> +    configStart();
> +
> +    m_backspaceRemoveComposed = on;

missing: m_backstapceRemoveComposedSet = true;

> kateconfig.h:610
>      bool m_showWordCount;
>      bool m_autoBrackets;
>  

After this line you should add bool m_backspaceRemoveComposed;

> kateconfig.h:644
>      bool m_autoBracketsSet : 1;
> +    bool m_backspaceRemoveComposed : 1;
>  

You need two flags: one m_backspaceRemoveComposedSet flat that knows whether this option is set for an individual view, and another one called m_backspaceRemoveComposed for the actual config value.
Currently, you only have the one for the actual value.

Just look at how this is e.g. done for m_autoCenterlinesSet and m_autoCenterlines.

REPOSITORY
  R39 KTextEditor

REVISION DETAIL
  https://phabricator.kde.org/D7660

To: safaalfulaij, #ktexteditor, jgrulich, hein
Cc: jgrulich, dhaumann, hein, kwrite-devel, #frameworks
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/kde-frameworks-devel/attachments/20170910/15a857ef/attachment-0001.html>