Apps Team Meeting 26 October 2006

Agenda

• Question du Jour : Do you have users/dogfooders with who you communicate directly? (5 min)
• Alpha4 status : (15 min) We're down to 8 bugs for the Apps team, down from 12 bugs last week. There are only 7 blocking bugs left. As we did in the past once we reach less than 10 blocking bugs, we are going to create an svn branch and let alpha5 development start again on the trunk. Thanks a lot for helping getting there! We have still to see a 2 days period without the emergence of a blocker but the stream of stability issues is calming down since Monday. This is good. We fixed a ton of really nasty bugs in the past few weeks and this is very positive. Comments, feeling on the status of alpha4?
• Alpha5 and beyond : (5 min) I've been too drowned into reproducing alpha4 bugs to have spent much time on that though I did cleaned up some. This will be my focus this week. Any question entering this new milestone?
• Issues and Announcements : (10 min) Have something to announce? Please take the stage.
• Misc: (5 min)
  • Adam (QA) will stay till end of January.
  • Jeffrey to be working from Atlanta for some weeks
  • Last fortnight for Darshana and Markku. Check before they leave if you've stuff to finish with them.

Summary

• Progress
  • Alpha4 done modulo 7 blocking bugs. Likely to branch today and open the trunk to alpha5.
• Plan
  • 0.7alpha4 : get to zero blockers! and build an RC...
  • 0.7alpha5 : start the new milestone for good
  • 0.7 : Philippe to finish assign bugs and tasks to relevant milestones

Minutes

Question du Jour

Do you have users/dogfooders with who you communicate directly?

• John: no but we should have some!
• Markku, Darshana: no
• Brian: has contact with quite a bit of people for localization, some people logging bugs on email as well
• Reid: getting his family to dogfood
• Philippe: have 5 so far

Alpha4 status

• John: was great to fix all those bugs
• Reid: felt kind of long, may be we should have more, shorter releases (with shorter debug cycles)
• Bryan: liked the fact that we focused on quality, concerned by the slippage
• Philippe: the longer debug period helped gage the real amount of work we have in front of us before Preview. As a consequence, we already punted out of the sticky plan a bunch of features that were not critical.

Alpha5 and beyond

• Everybody ready to attack alpha5
• Philippe concerned by the perf and the UI responsivness

Issues and Announcement

• Robin: the preview of the transparency in wx is up! Check it out!

Status

Brian

• Fixed Bug#7162 Memory leak when displaying a mail error dialog
• Fixed Bug#7159 MailService ERROR: Exception raised in Twisted Framework Layer
• Investigated Thoroughly [Bug 7192] repository ERROR: Value <clip> of type <type 'str'> in attribute 'uid' on <IMAPDelivery: <clip> is not recognized by type //Schema/Core/Long. Was unable to reproduce.
• As part of Bug#7192 Signed up for Yahoo mail Plus service with POP enabled, Google Mail with Pop Enabled, and Fastmail with IMAP enabled and pulling mail from Yahoo via POP. These new accounts have been added to the QA Mail Testing Accounts page for future use.
• Have a patch ready for Bug#7163 New: Mail marked "read" on IMAP server when Chandler syncs
  

Bryan

• Done with my alpha4 bugs, just in time for
• PTO Thursday & Friday

Darshana

• Worked on a new release of Parsedatetime (0.8.0). Solved a couple of important issues in the release.
• Worked on Bug#7198 (Parse items for date/time when it is stamped as a Calendar Event by the user)
• Worked on Bug#6470 (Use search bar for Quick item entry). Converted the stamping to stamping using annotations

Jeffrey

• Your status here...

John

• Finished work on Bug#7177: "Position out of range" when deleting the Welcome message and Bug#6436: Implement a selected state for sidebar collection icons and sharing status icons.
• Fixed Bug#5243: removing events from the preview area confuses sidebar painting, it also fixes Bug#6410: preview area doesn't refresh properly after adding. This turned out to be much more work than estimated, but I suspect it may fix other redraw bugs on Mac.
• Fixed Bug#7160: Creating a New Message/Task/Event should not switch you out of the All App Area. But, unfortunately the functional tests revealed a tricky bug in which the markup bar does not update properly. I decided to postpone work on this bug until after I'm finished with my blocking bugs.
• Worked on Bug#6559: 12% startup performance regression with r11443.
• Spent most of the week working on blocking bugs:
  • Bug#6977: Guard byte incorrect during shared memory free. I've been running functional tests all night long for several weeks now and have not been able to reproduce this bug. I suspect I won't be able to reproduce it, or maybe it's gone.
  • Bug#6732: Intermittent crash during functional tests. This bug has been particularly difficult. After spending a very long time on it I am still not able to suceed running the functional tests even once. Just as I get close to narrowing down one problem a completely different problem pops up. Since the failures I'm getting don't happen on tinderboxes, I'm beginning to suspect either the problems aren't seen on the tinderboxes because they are running an out of date version of the OS, or a build/configuration problem. I'm further hampered by my Mac, which seems an order of magnitude slower than my Linux and Windows environment.

Markku

• Closed:
  • Bug#6932 Feeds tutorial needs to be updated.
• Commit pending opening of alpha5:
  • Bug#6729 The save settings feature does not work with Unicode.
• Code-review (and commit pending opening of alpha5):
  • Bug#6923 Include term vectors to Chandler Lucene indexing.
  • Bug#6926 Make sure that search results from Lucene are accessible in descending scoring order.
• Open (and commit pending opening of alpha5).
  • Bug#6805 Automatic tagging of new items: this is just a collection of bugs (Bug#6923, Bug#6926, and Bug#6927).
  • Bug#6927 Design and implement a simple example parcel that does automatic tagging using Chandler's search capabilities and K nearest neighbor algoritm: This still requires finetuning with Philippe's data set (see Bug#6966) but hopefully I am able to close it this week.
• My last day here at OSAF is going to be Tuesday Oct 31th, 2006 but I am going to stop by at next Thursday's staff meeting. Thank you everyone for making my learning experience so productive and fun!

Philippe

• Mostly busy doing alpha4 monitoring
• Reviewed sticky plan for alpha5 and alpha6
• Investigated bugs with dogfooders (Bug#7072, Bug#7133, Bug#7190 among others)

Reid

• Fixed Bug#6572 (Mark-up bar icons still displaying white background on Mousedown state)
• Have a patch (for alpha 5) that prevents the markup bar from getting focus
  • I think this may help with other markup bar issues
• looked into Bug#7153 (Stamping button is stuck in Mousedown state and won't go away)
• looked into Bug#6745 (Reply/Reply All/Forward buttons are always enabled)
  • this may be the easier of my last two alpha 4 bugs, so I'm concentrating on it

Robin

• Fixed Bug#7167, removing a log comment that was triggering the feedback dialog.
• Rolled a new wx tarball with fixes for bugs 5822, 6572 and 7167.
• Helped John, Reid, and Jeffrey with questions, and suggestions.
• Worked a bit more with getting familiar with WebKit, discussed with Kevin O. about how to do the rendering with wx and how to share code while we are getting it going.
• Worked on helping to smooth out some rough edges in the wxGraphicsContext work. Also some debugging and experimenting with it. I've added a sample to the wxPython demo, and put a screenshot in the wiki at WxPythonGraphicsContext.
• Made both a 2.7.1.2 and a 2.7.1.3 release of wxPython, to flatten some bugs not found by my testing group. I'm thinking that maybe they were not as agressive of testers as they've been in the past. After 3 releases in 8 days I am a bit brain-dead.
• Various community support activities via email and IRC.

-- PhilippeBossut - 26 Oct 2006