Bug 393168

Created attachment 112047 [details]
Working week patch

Hello developers,

First of all, thank you all for this great piece of software!

(I've been mostly a user of KMyMoney, and just started to learn about coding recently.  It's my first time submitting a bug report (and a proposed patch), so kindly please bear with me if there's any infelicities in my expression)

There's a feature in scheduled transactions which allow them to move backwards/forwards in case a scheduled entry falls on a non-processing day, e.g. a weekend.  However, I've noticed that that seems to have stopped working in 5.x/master.

That's working in 4.8.  Out of curiosity I tried to find out what's changed.

It turns out there seems to be an 'overlooked' word in the porting of 'working week' to kf5, namely that:

-  // TODO: port kf5
-  int weekStart = 1;//KLocale::global()->workingWeekStartDay();
-  int weekEnd = 7;//KLocale::global()->workingWeekEndDay();
+  QLocale locale;
+  int weekStart = locale.firstDayOfWeek();
+  int weekEnd = weekStart-1;
+  if (weekEnd < Qt::Monday) {
+    weekEnd = Qt::Sunday;
+  }

well in fact, weekStart and weekEnd would be referring to the workingWeek, so in most cultures that'll actually be 1 and 5 (i.e. Monday to Friday).

In searching for alternatives (in porting away from KLocale in its wake), I found from my Google searches QLocale::weekdays() (as opposed to weekends) which seems to be a good replacement -- from the QT docs:

Returns a list of days that are considered weekdays according to the current locale.

So in the proposed patch attached I've set m_processingDays accordingly from QLocale::weekdays().  I've also moved down the #ifdef in KMyMoneyApp::isProcessingDate so that it could always check against weekdays/ends even without KF5Holidays.

Sorry for the lengthy description  (And don't be mistaken, there's not a scant of criticism at all.  To the contrary I appreciate all the developers' efforts all along, and the KF5 port must have been a little bit tedious)

And actually I'm very green at coding.   Very much appreciated if interested developer(s) could help review the proposed patch --- and feel free to amend / enhance it in any way you like.    

* * *

Some related past commits below for your reference:

2017-09-13  Ported start of week handing to KF5
(https://cgit.kde.org/kmymoney.git/commit/?id=e3aafecfbd096f048dcd53ae80f43f10e935c40e)

2016-06-16  Remove kdelibs4support, disbaled some code which will need to be ported.
(https://cgit.kde.org/kmymoney.git/commit/?id=fcfe5beca1656092d6f45495816e1b604427279c)

The feature (and the m_processingDays bitArray) was first introduced some 8 years ago:
2010-01-17  Added patch provided by Ian Neal
(https://cgit.kde.org/kmymoney.git/commit/?id=406cd4845f3b4685194142746a93a2720293c0be)
For reference, the related discussion about the implementation could be found in FEATURE: 221800
(https://bugs.kde.org/show_bug.cgi?id=221800)