Apps Team Meeting 13 July 2006

Agenda

• Question of the week: What projects are you working on for α4? What open questions are you facing?
• α3
  • Remaining bugs
  • Code reviews to schedule?
  • Unit or functional tests that should have been written but didn't make it?
• α4
  • DesktopZeroDotSeven, DetailedAlpha4Plan
  • Geek talks to schedule?
• Philippe back next week!

Notes

• What projects are you working on for α4? What open questions are you facing?
  • John: No idea! Sidebar? Documentation? There's a list that he has -- styles and other things on the list.
  • Darshana: Datetime parsing (NLP), attribute editor documentation would be useful, where to put the datetime code?
  • Markku: hoping to close current i18n workload -- tasks almost ready but waiting for input. Heikki and Philippe will discuss plans for another big project. John has some bugs that he could hand Markku. John mentioned David's header as a possible project.
  • Bryan: Dashboard. Domain model side is open -- list of questions to the dev list and Grant -- need to have a long talk. 2 weeks of not being blocked at least.
  • Jeffrey: Calendar ui tweaks. Lozenge design (won't be that complicated). A bit of freebusy work no longer blocked by cosmo. May end up doing recurrence work. What are the larger projects that he should be taking on?
  • Reid: more wx work. Printing -- nice big task. IE printing UI on the mac is pretty good. What it can print -- will have questions. Would be nice to print a nice calendar. What does it look like to print the dashboard. Brian notes: keep an eye out for how the localization is going to work -- from wx? from os?
  • Brian K: wrap up i18n. email. Figure out what the goals/spec is. A few schema changes for Ted. Getting functional tests to pass has been a bit of a pain.
  • Robin: ui linux bugs. others not targetted to alpha4. finalize some localization issues. A few more wx merges. Script is ready for the mirror.
    • What's the plan for running tinderboxes on wx? Robin knows there is a plan, but doesn't know the details. John: that would help quite a bit to get that going.

• α3
  • John thinks there is a one line fix in wx for the flicker bug. He needs someone to get back to him to be confident in his fix. Do we take it out? Worth the effort to take it out and roll another wx? Also, we're constantly tearing down the toolbar all the time. Reid: we're about to roll another wx?
  • Tests?
    • Printing?
    • If its in the menu item, there should be a functional test.
    • Toolbar functional tests?
    • Autocompleting functional tests
    • Bugs in tests make it hard. Emulation is hard to debug -- a lot of voodu. Emulate key press generates lots of strange events. Throwing in idles.
    • ProcessEvent^? none -- we know its a bug. Should run asserts in release binaries. Could be way simplified and a lot safer.
    • Bryan observed that Aparna had specific goals or reasons to make some decisions.
  • Need some brainstorming to come up with a set of choices and come up with a plan. Makes of sense. Everything on the list would be awkward.
    • First layer set of examples would be useful to help prevent wheels from spinning.
• α4
  • need some clarity on what's really in α4
• Code reviews
  • Reid's buttons
• Markku project ideas
  • test framework improvements?
  • column header?
  • sizing blocks for their content (widgets to size themselves for their content) -- text box is fixed box and adds a scroller -- Jeffrey thinks this would aid the ui the most
  • reminders when chandler isn't running?

• Bryan: I can really do source level debugging -- instructions are great! helps a lot and is really terrific!
• Jeffrey: I love the taskbar icon thing!

Status

John

• Fix Bug#5247: We see two Help menus and a new Window menu in Chandler on mac. Robin and Grant did all the heavy lifting on this bug fix.
• After postponing helping Dan on Bug#6166 (TestNewEvent throws unhandled exception in new test framework) for over 3 weeks, I finally tracked it down. It was probably the most difficult bug I've tracked down in a year. There are two other duplicates of this bug: Bug#6110: functional tests (debug) crash Python on occasion is a duplicate and Bug#5737:FunctionalTest crash - wxSetFocusToChild. I suspect it may come up in other situations as well.
• I have two remaining alpha 3 buga I've been working on: Bug#6127: detail view too big on startup and Bug#6139: menus not causing actions to happen. Both are Mac specific.
• Fixed Bug#6127: detail view too big on startup
• Fixed Bug#6139: menus not causing actions to happen. Both are Mac specific.
• Fixed Bug#6127 also fixed Bug#6164: Resizing application to minimal size and then resizing bigger messes up layout
• Finally tracked down Bug#6163: Toolbar icons flash when resizing window, also during functional tests, which has plagued us for so many years. The final bug fix depends upon some feedback from Julien Smart.
• Finally figured out, with the help of the wing guys, how to have project files for each of the platforms that just work.
• Fixed a calendar redraw bug I encountered while tracking down Bug#6163

Brian Kirsch

• Continued work architecting egg i18n and Chandler integration.
  
  • Sent out email containing information as well as demo samples using egg i18n.
    
• Tested Linux and Windows snapshots of Chandler which contain new code that allows wx to discover the location of its translation wxstd.mo files.
  • TODO: get a OS X build tested and then check in wx changes and i18nmanager changes to remove wx catalog configuration
• Helped resolve [Bug 6224] logger doesn't like %s'ing items with i18n displaynames.
• Working heavily on [Bug 1745] Use displayName for Item titles?
  • Ported more code to use title.
  • Opened two new bugs related to Feeds egg example and Flickr sample.
  • Testing and fixing as needed code while running unit / functional tests on Windows XP and Linux.
  • Opened [Bug 6237] The displayName on kindInfo is not needed.
• Resolving issues related to the loss of my G4 Powerbook laptop
  • Received and started configuring new Mac Pro laptop
  • Opened old powerbook laptop case and performed diagnostics.
  • Removed hard drive from old powerbook laptop and purchased an enclosure for drive to access important OSAF files.
  • Installed hard drive in enclosure.
  • Mailing old laptop back to KEI IT deparment.
  • Requested to Dave Cowen 512mb additional RAM and an addition battery for new Mac Pro to match configuration of old powerbook G4.
    

Reid

• Checked in wx change to fix Bug#6123 "Toolbar buttons are small and have dark selection outline"
  • removed code that defaulted to small icons
  • turned on native toolbar
  • it's been difficult to tell for sure if native toolbars are on or not. Many times I thought they were native and they were obviously not.
• rewriting Mac toolbar code to use proper callback APIs
  • this will allow embedding of arbitrary controls
  • means by which search bar will be enabled
• reading up more on printing in wx

Robin

• Helped John and Reid with various alpha 3 bugs and other issues.
• Fixed where the .mo files are installed on Windows so wx.Locale can find them by default.
• Added project files and made makefile changes so the wxrc utility is built on Windows. Combined with Bear's make file changes to include wxrc in the tarballs.
• Updated wxPython build scripts to be able to build Universal binaries on OS X 10.4.
• Updates to wxPython to be able to build with Python 2.5. This entailed switching some int typess to Py_ssize_t.
• Added transparent dragging and hints for wxAUI on the Mac.
• Added wxAUI (Advanced User Interface) wrappers to wxPython. It's a nice docking frame library.
• Added a contrib from Paul Lanier calledPseudoDC which provides a way to record operations on a DC-like object and then replay them later to a real DC.
• Built some 2.6.3.3 and 2.7.0.0 preview releases.
• Vaious wx and wxPython bugs fixed amd enahancments added based on reports from the community.
• Various community support activities.

Markku

• Bug#5972 is almost ready and is waiting input from Bear (Bug#6092).
• Bug#6007 is waiting Brian to finalize the integration of the egg based resource management system (RMS) to Chandler. Based on current knowledge, this should be ready very soon.
• Bug#6092 is almost ready and is waiting input from Brian (Bug#6226).
• Bug#6171 needs some finalizing and more testing. The work on this is currently halted as this is considered a low priority task and ironing out the last bug(s) requires consultation from Robin. Hence, the plan is to wait until alpha3 is finished and until we have more time (Robin's time) to invest in this.
• Bug#6227 has been implemented on the eggdemo and the next step, which is integrating the tests to Chandler, is pending Brian's work on RMS-Chandler integration.
• Bug#3824 is still punted (at least until alpha3 is finalized).
• As the previous tasks have not provided enough work to keep me busy, I have started to look for secondary work from bugzilla (tasks to do when my primary work is halted). Current list of secondary work include Bug#6090 and Bug#6212.
• Proposed improvements to Chandler: Bug#6212, Bug#6215, Bug#6216 and a discussion.
• Investigated an unicode related problem in MSVC debugging with John.
• Finished my CIEE midterm evaluation (and reminded Philippe to do his).

Bryan

• Dashboard: Working on fixing sectioning bugs as I find them; at the moment, clicking on section rows crashes.
• Got a big boost from getting wx source-level debugging to work (thanks to John for help, and Robin for making things simpler!)
• Working with the ever-helpful Wing folks to figure out the stepping problem that John and I see when debugging using python_d

Jeffrey

• Caught up on a mountain of email (still have 200 messages, mainly design threads, but that's nothing )
• Fixed several alpha3 bugs involving all day events in the calendar, and recurrence interactions with all day events
• Worked with Darshana to get her set up and working on integrating Bear's date parsing into Chandler
• Worked with the Calconnect folks on getting their freebusy demo for Boeing to talk successfully to Cosmo (we're doing freebusy the way Lisa and Bernard say we should, but it's not the way other folks expect)
• Worked on getting the old patch in bug 5953 to work with the new background sync sharing

Philippe

• Went to EuroPython in Geneva with Andi:
  • My talk on PyICU was well received, nice feedback and interest from the community
  • See my daily logs for more details
    • EuroPython Day 1
    • EuroPython Day 2
    • EuroPython Day 3
• PTO after that till Friday 14th