[Interest] QTreeView::resizeColumnToContents not accurate enough

[Elvis Stansvik]

2017-02-02 12:51 GMT+01:00 Elvis Stansvik <elvstone at gmail.com>:
> 2017-02-02 11:39 GMT+01:00 Elvis Stansvik <elvstone at gmail.com>:
>> 2017-02-02 11:23 GMT+01:00 Elvis Stansvik <elvstone at gmail.com>:
>>> WIth the following minimal example:
>>>
>>> #include <QApplication>
>>> #include <QTreeView>
>>> #include <QStandardItemModel>
>>> #include <QStandardItem>
>>>
>>> int main(int argc, char *argv[])
>>> {
>>>     QApplication app(argc, argv);
>>>
>>>     QStandardItemModel model(3, 2);
>>>     model.setHorizontalHeaderLabels({ "Header 1", "Header 2" });
>>>
>>>     for (int row = 0; row < 3; ++row) {
>>>         for (int column = 0; column < 2; ++column) {
>>>             model.setItem(row, column, new QStandardItem("Foo"));
>>>         }
>>>     }
>>>
>>>     QTreeView view;
>>>     view.setModel(&model);
>>>     view.show();
>>>     view.resizeColumnToContents(0);
>>>
>>>     return app.exec();
>>> }
>>>
>>> I get the attached result. Note how the first column does not snugly
>>> fit the contents. The limiting factor should be the width of the
>>> header ("Header 1"), but there's plenty of room left to shrink the
>>> column.
>>>
>>> Anyone else had this problem? Is there a way to improve the accuracy
>>> of QTreeView::resizeColumnToContents ?
>>>
>>> This is with Qt 5.5.1 (Ubuntu 16.04 package).
>>
>> I wonder, considering the amount of extra space, could it be that the
>> calculation is taking the width of a sorting indicator (the ^ and v
>> arrows) into account, even if no such handle is visible (sorting is
>> not enabled)? If so, should that be considered a bug that I should
>> report?
>
> Nevermind, I realized now that QHeaderView delegates to the style to
> draw the header section, and that this is a bug in my style (Breeze).
> With the Fusion style the result is as expected.
>
> Breeze should take the sort indicator into account somewhere around
> here: https://github.com/KDE/breeze/blob/master/kstyle/breezestyle.cpp#L2901-L2940
>
> I'll report it to the KDE folks.

(For those interested, I submitted this patch to fix it:
https://phabricator.kde.org/D4406)

>
> Elvis
>
>>
>> Elvis
>>
>>>
>>> Thanks in advance,
>>> Elvis