Code Log

Commit Date  
[r12423] by iangardner

Adding an "unlink" operation for changing a linked segment into a deep copy. In the UI I've followed the pattern laid down by the "Turn Repeats into Copies" menu option, adding a "Turn Links into Copies" option.

To implement the undo for this unlink operation, I need to guarantee that the SegmentLinker (the object responsible for coordinating a set of linked segments) is still alive at the time the undo is actioned. This is why the SegmentLinkToCopyCommand holds a copy of the original segment before unlinking it. It's a blunt but effective solution, at the cost of holding a copy of the segment in memory (the segments may be big, there may be a lot of them). Should be OK for the time being is my assumption for most real life use-cases.

The lighter alternative would be to have the SegmentLinker as a ref counted object and just hold a reference to it in the SegmentLinkToCopyCommand.

2011-09-11 08:41:58 Tree
[r12422] by dmmcintyr

Patches from Tom Breton (Tehom)

2011-09-11 01:48:43 Tree
[r12421] by tedfelix

SegmentJoinCommand will now select the old segments on undo. This leaves
the selection as it was prior to the join.

2011-09-07 22:29:47 Tree
[r12420] by tedfelix

Changed the behavior of SegmentSplitCommand to ensure that if a segment was
selected before being split, the two new split segments would also be
selected. Selection state of the original segment will be preserved on
undo. This required exposing
CompositionModelImpl::isSelected(const Segment* )
as public.

2011-09-07 22:19:14 Tree
[r12419] by tedfelix

After joining segments, the newly created segment is now selected. This
required exposing a previously protected version of setSelected() in
CompositionModel (and Impl) as public.

Also added some more comments for doxygen.

2011-09-06 23:58:31 Tree
[r12418] by tedfelix

Added comments to improve doxygen output. May not be quite correct but
represent my best understanding at this point. Hopefully better than

2011-09-06 00:07:17 Tree
[r12417] by dmmcintyr

I don't think this helped a bit, but I'm committing it anyway.

2011-09-05 02:10:05 Tree
[r12416] by dmmcintyr

The QStringList::split rewrite needed a QString::skipEmptyParts added to
reproduce the original behavior. This was probably the most satisfying
breakthrough I'm going to have on this whole subproject. With this simple
change, pitch pickers, the configuration dialog, and the entire notation editor
are all working again; or at least no longer crashing instantly.

2011-09-04 17:28:23 Tree
[r12415] by dmmcintyr

This particular setBackgroundPixmap is an internal method that still needs to be
called. Putting this line back in gives us texture on the tiny nugget of
segment canvas that is drawn, but the bulk of the canvas is still a black void.

2011-09-04 16:56:09 Tree
[r12414] by dmmcintyr

The slog ends!

2011-09-03 21:27:32 Tree
[r12413] by yguillemot

Third fix for the same crash !
(When moving toward the right twice the first segment of the last track of
bogus_surf_jam while notation is opened).

2011-09-03 20:40:54 Tree
[r12412] by dmmcintyr


2011-09-03 09:44:38 Tree
[r12411] by dmmcintyr


2011-09-02 23:43:20 Tree
[r12410] by dmmcintyr


2011-09-02 23:43:15 Tree
[r12409] by dmmcintyr


2011-09-02 23:43:10 Tree
[r12408] by dmmcintyr


2011-09-02 23:43:06 Tree
[r12407] by dmmcintyr


2011-09-02 23:43:02 Tree
[r12406] by dmmcintyr


2011-09-02 23:42:57 Tree
[r12405] by dmmcintyr


2011-09-02 23:42:52 Tree
[r12404] by dmmcintyr


2011-09-02 23:12:48 Tree
[r12403] by dmmcintyr


2011-09-02 23:11:18 Tree
[r12402] by dmmcintyr


2011-09-02 23:11:14 Tree
[r12401] by dmmcintyr


2011-09-02 23:11:08 Tree
[r12400] by dmmcintyr


2011-09-02 23:11:04 Tree
[r12399] by dmmcintyr


2011-09-02 23:10:59 Tree
Older >