GSOC 14 PDE 3.0

Rough Schedule

• May 19 - June 1: PDE work in 2.x, 3.0 work in PDE X
• June 5: shallow fork Github repo into processing3?
• June 1 6: split to 3.0, all work happens in 3.0 from now on
• June 23: GSOC Midterm
• June 27: Release 3.0 beta, default download
• Aug 11: GSOC Final
• Aug ???: 3.0 release?

Manindra / Dan

• PDE X Robustness
  • Code completion doesn't always work for local variables exp#68
  • Memory management exp#1
  • Breakpoints in classes exp#47
  • Matching offsets between PDE Code and Java code
    • Initial implementation complete. Now on the look out for related bugs.
  • build instructions wiki for PDE X. no longer relevant when PDE X takes over?
• PDE Menu items prior to Alpha Release
• Transition PDE X to 3.0 Alpha Release (June 1 target date)

Joel / Florian

• Existing PDE Problems (Do these get fixed in 2.x or in PDE X at first?)
  • list of open sketches in the menu (a la photoshop)
  • code syntax highlighting (all goes to gray), also 1681
  • undo 707
  • auto-format 364, 2271
  • add present mode (fullscreen) background color to preferences window 69,
  • auto-save 2286
• Contribution Manager
  • fail more gracefully if no internet connection 2426 or a proxy server is in use
  • Show installed/available versions of modes in modes list 2561
  • go through existing tool contributions with Elie (are there outdated tools?)
  • manage example packages through contribution manager
  • Include dates for last update? 2498
  • Guidelines for modes 1656
  • help / reference for contrib libraries 943
• Pattern Sketchbook after Examples window
  • The new Open menu is:
    • Open
    • Examples
    • Sketchbook
• Recent (as a list of most recent after a line)
• Language stuff 2173
• Internationalization 632

Gal / Dan

• Implement PShape in Java2D
• Tweak mode integration exp#42. Let's explore how much work this is, if it proves to be overly complicated or a distraction from core PDE 3.0 development, we would maybe table it for later. But would be great to happen if we can.

More PDE

• return keyword not treated as such when followed by a bracket 2099
• IllegalArgumentException when clicking between editor windows 2530
• "String index out of range" error 1940

Other To do for Later in GSOC?

Git Repo

• Clean up and make smaller repos (web, processing , android), 1898

Preprocessor

• using a processing keyword as a variable name gives unhelpful error message 93
• Functions reference for set() need to go to PImage or main set() reference by context

Fantasy PDE X features (we likely won't get to these)

• Auto brackets/quotes closing exp#60
• Tracing tool? exp#56
• REPL in console exp#55
• Line numbers

June Meeting Discussion

PApplet

• keep PApplet but no longer subclass applet?
• Make new PSketch?
• Break out things in PApplet
• this would move static utility functions to PUtil (sort() and others). It would be transparent to users of the PDE, but Eclipse users might notice.
• Ben: three alternatives for the PApplet transition:
• 1. Just remove 'extends Applet' from PApplet. Pros: fewest changes to libraries and other code. Cons: Anachronistic 'applet' misnomer of a name. Doesn't provide a clear way for people to use old PApplet features inherited from Component (ability to embed a PApplet into a Java app, any Java AWT things done by libraries). Not sure how we handle embedding into other Java apps.
• 2. Introduce PSketch, a new base class for most sketches. Keep PApplet as a shell that still extends Applet. Pros: Avoids misnomer. Makes the separation clear. Provides one location for people to maintain the embeddable version (i.e. I'm not really interested in supporting this Component thing, but others could take that over). Cons: More disruptive for libraries. The 'parent' object name has changed.
• 3. Err... forgot the third one.

PShape

• Java2D PShape should work like P2D 1623

PVector?

• return this
• need to know if this breaks method signatures, and therefore, backwards compatibility (Dan to check?)
• new Vector classes?
• Ben: recommended way forward if we want to make material changes to PVector. PVector was designed for speed/efficiency constraints that aren't as relevant nowadays, and I'd rather do a more significant design change and deprecate the old PVector.

Touch

• touchX, touchY, touch array?
• high level events like swipe

Size of PDE download

• video library trimming?
• Bring back multi-platform export via tool?