[mythtv-commits] Ticket #13069: Fix QObject related compile warnings in trunk

[MythTV]

#13069: Fix QObject related compile warnings in trunk
------------------------------------------+-------------------------
     Reporter:  David Hampton <mythtv@…>  |      Owner:
         Type:  Developer Task            |     Status:  new
     Priority:  minor                     |  Milestone:  29.0
    Component:  MythTV - General          |    Version:  Master Head
     Severity:  medium                    |   Keywords:
Ticket locked:  0                         |
------------------------------------------+-------------------------
 These fixes originated from compiler warnings generated from adding
 the -Wextra flag to the build.  This pull request only addresses
 changes related to fixing this error:

   In copy constructor ‘blah::blah(const blah&)’:
     warning: base class ‘class blah’ should be explicitly initialized
     in the copy constructor [-Wextra]

 Implementing explicit initializers for these objects results in an
 error instead of a warning:

    In copy constructor ‘blah::blah(const blah&)’:
      error: ‘QObject::QObject(const QObject&)’ is private within
      this context : QObject(other)
      /usr/include/qt5/QtCore/qobject.h: note: declared private here
      Q_DISABLE_COPY(QObject)

 Fix these errors by doing the following:

 1) FileSystemInfo and MetaGrabberScript don't use any of the features of
 QObject, so make them standalone classes.

 2) Remove public copy constructors from MythSystemLegacy.  This object
 uses Qt signals, so it has to be based on QObject.

 3) Remove the copy constructor from MythCookieJar.  Do this by
 splitting it into its two components; creating a new jar and
 copying cookies from one jar to another.  This is probably not the
 ideal solution, but it has the smallest impact on the existing code
 base.  (An better solution would allow cookiejars to be created
 once, and then copy cookies back and forth.  This would eliminate
 the need to destroy and recreate the jar on each copy.)

 4) Clean up the RecStatus object.  This object is essentially an enum
 with some related functions. It doesn't ever appear to actually be
 instantiated as an object.  This object is based on QObject and
 therefore the object itself cannot be used with the Qt metatype
 system.  The pointer, however, is automatically registered with the
 metatype system (as are pointers to all QObject derived classes).
 Remove the copy constructor and the unnecessary/duplicate metatype
 declaration/registration code.

 5) Clean up all the data/service contract objects that are based on
 QObject.  Like RecStatus, the objects themselves cannot be used
 with the Qt MetaType system, and the pointers are automatically
 registered.  This allows all of the declaration/registration code
 to be removed from the files.  Removing the registration code
 allows all of the InitializeCustomTypes functions go away, which
 allows the service object constructors to be cleaned up.

 The CopyListContents code also needs to be simplified to not convert
 from pointer to object to reference, because this causes an implicit
 copy of the object. That means that all the object copy functions need
 to be updated to use pointers instead of references.

 6) Create a test framework for the data contracts.  This currently
 builds five different types of objects, converts them to a QVariant
 and back, and then compares the results to the original.  It should
 be expanded to all the data objects, and other tests added.

 These changes have been tested by working through most of the menu
 items in mythfrontend, and by working through most of the items in the
 new web frontend.

--
Ticket URL: <https://code.mythtv.org/trac/ticket/13069>
MythTV <http://www.mythtv.org>
MythTV Media Center