solitaire: Add README.md

README.md

@@ -30,8 +30,12 @@ found [on my site](http://martinfitzpatrick.name/tag/pyqt).
 1. [Currency converter](currency/) - "Doughnut" (PyQtGraph)
 1. [Solitaire](solitaire/) - "Ronery" (QGraphicsScene)
 
-# License
+## License
 
 All code is **licensed under an MIT license**. This allows you to re-use the code freely,
 remixed in both commercial and non-commercial projects. The only requirement is to
 include the same license when distributing.
+
+## Other licenses
+
+Icons used in the applications are by [Yusuke Kamiyaman](http://p.yusukekamiyamane.com/).

paint/README.md

@@ -6,16 +6,16 @@ ready made pictures of pie.
 Piecasso is a clone of the Paint programme from Windows 95 (ish) with a
 few additions (and subtractions). The programme features standard
 tools including pen, brush, fill, spray can, eraser, text and a number of
-shape drawing widgets. 
+shapes. 
 
-![Piecasso](screenshot-paint.png)
+![Piecasso](screenshot-paint2.png)
 
 You can copy from the image, with a custom shape,
 although pasting + floating is not supported. The canvas is a fixed size
 and loaded images are adjusted to fit. A stamp tool is also included
 which is pre-loaded with pictures of delicious pie.
 
-![Piecasso](screenshot-paint2.png)
+![Piecasso](screenshot-paint.png)
 
 > If you think this example app is neat and want to learn more about
 PyQt in general, [take a look at my ebook & online course

solitaire/README.md

@@ -0,0 +1,60 @@
+# Ronery — A Solitaire game in PyQt
+
+The classic card Solitaire (Klondike) game with a North Korean / 
+'Team America: World Police' twist. The game is otherwise identical, 
+ offering the same options. Built with PyQt, it uses QGraphicsScene
+ to handle the play area.
+
+![Ronery](screenshot-solitaire.png)
+
+The classic Solitaire win-screen is also implemented, although since
+the play canvas is not a bitmap it doesn't "fill up". It's still
+sufficiently satisfying. 
+
+![Ronery](screenshot-solitaire2.png)
+
+The configuration options allow for 3-draw and 1 draw styles, with 3, 5
+or infinite rounds.
+
+## Code notes
+
+### Card handling
+
+The approach I used here was to treat each place where cards can be placed
+as a 'stack' object, which handles the cards it currently holds. Each stack
+is responsible for the layout of the cards on it's pile.
+
+A stack is also responsibly for accepting/rejecting an attempted drop of a
+card on it. Some stacks, e.g. the deck + deal pile do not accept any drop.
+Others, such as the finish piles, have specific rules.
+
+In most cases there is no relationship between any cards on a stack, with
+the exception of the 'in play' stacks along the bottom. Here cards have
+a parent-child relationship, allowing multiple cards to be dragged at once.
+
+### The end animation
+
+The end-game was a bit weird to implement. Since it happens 'outside' of
+game time the first thought was to have a self contained loop, using
+`QApplication.processEvents` to tick over. But that's nasty.
+
+Instead I used a timer, idle during play, which constantly pings an
+animation step-forward endpoint. This moves cards (faking 'gravity', bouncing)
+as well as re-stacking them once they're out of the play area. 
+Restacking uses the normal stacking code.
+
+The final step was to block event handling, otherwise the cards could
+still be grabbed while they bounced (and dropped). The easiest way to 
+achieve this was to simply place an event-capturing object over the entire
+window.
+
+## Other licenses
+
+The card images were made from a [freely available set of PNGs](https://github.com/hayeah/playing-cards-assets)
+which were themselves auto-generated from SVGs traced from actual cards(!).
+The original card designs are out of copyright.
+
+Note that the linked repository contains some custom art on ace cards which may still be copyrighted and so 
+was not used.
+
+Icons used in the application are by [Yusuke Kamiyaman](http://p.yusukekamiyamane.com/).

solitaire/solitaire.py

@@ -651,8 +651,8 @@ class MainWindow(QMainWindow):
 
                 if card.pos().x() < - CARD_DIMENSIONS.width():
                     card.vector = None
-                    drop = random.choice(self.drops)
-                    drop.add_card(card)
+                    # Put the card back where it started.
+                    card.stack.add_card(card)