peter9477

Cascades bugs et al

Nov 27th, 2012
209
0
Never
Not a member of Pastebin yet? Sign Up, it unlocks many cool features!
text 6.15 KB | None | 0 0
  1. Bugs
  2. ---------
  3. 1. Tab button text doesn't update properly when you do this:
  4.  
  5. onClicked: {
  6. var tab = root.activeTab;
  7. root.remove(tab);
  8. root.activeTab = defaultTab; // show first tab
  9. }
  10.  
  11. ... but it works when you do this:
  12.  
  13. onClicked: {
  14. var tab = root.activeTab;
  15. root.activeTab = defaultTab; // show first tab
  16. root.remove(tab);
  17. }
  18.  
  19. In both cases, it does properly change the tab.
  20.  
  21. (NOT A BUG) 2. Can't store a null in any QML property, even if in an object attached to variant property.
  22.  
  23. 3. SubmitKey.None doesn't make submit key vanish; SubmitKey.Search shows an "enter" symbol
  24. with down-the-left arrow
  25.  
  26. With default TextField, and submit Search, doesn't show as "Search" but doesn't onSubmit either
  27. With Pin TextField, and Search, shows as "Search" but no onSubmit.
  28. Lots of other combinations seem screwed up too...
  29.  
  30. 4. Pane with Sheet opened, can peek back and sometimes rather than bouncing back into position
  31. after you release, it will actually be closed yet the page is completely blank and you
  32. have no choice but to exit the app (though top-swipe menu will open, at least the dummy
  33. one).
  34.  
  35. 5. Doc bug: https://developer.blackberry.com/cascades/reference/bb__cascades__arraydatamodel.html#value
  36. says "If index i is out of bounds (that is, if i is less than 0 or greater than size()),"
  37. but that should be "size() - 1", or "greater than or equal to size()".
  38.  
  39. 6. Adding item which sorts to end of ListView with GroupDataModel is not resulting in ListView
  40. hearing about it so the list size isn't updated and I can't scroll down to the last item,
  41. or at least when i scroll there the list will "snap" back to the previous item.
  42.  
  43. 7. Can't load .amd files for nine-slice etc from anywhere other than asset: URL, and
  44. it doesn't even allow traversal up out of there.
  45. Basically asset: is deliberately crippled...
  46.  
  47. 8. TabbedPane, with NavigationPane and secondary Page, if you use Peek to make the
  48. page close, the darkening animation leaves it darkened. If you're re-using the
  49. page later, the opening animation doesn't lighten it so it's still dark. You
  50. can peek/unpeek it and it will lighten again but if you merely open the page
  51. and tap on Back (not slide) then it stays dark indefinitely but is otherwise normal.
  52.  
  53. 9. Problems with ArrayDataModel and move() downwards, and to end. Moving down requires
  54. index one greater than it should be. A side effect is that you cannot move
  55. to the final position as there's a restriction on that of size() - 1.
  56. Also the ListView does not preserve selections when items move.
  57.  
  58. 10. TitleBar doesn't move when peeking at previous page.. only content slides over.
  59.  
  60. 11. "not unrealized" messages in log at app termination:
  61. 09:20 < vladest> ERROR: Context: Object name=" "Container" " [objectId= 34 ] not unrealized
  62. 09:20 < vladest> ERROR: Context: Object name=" "Label" " [objectId= 35 ] not unrealized
  63. 09:20 < vladest> ERROR: Context: Object name=" "EngineDocument" " [objectId= 36 ] not unrealized
  64. 09:20 < vladest> ERROR: Context: Object name=" "TextStyleData" " [objectId= 4 ] not unrealized
  65. 09:20 < vladest> ERROR: Context: Object name=" "Page" " [objectId= 33 ] not unrealized
  66.  
  67. Mine:
  68. 11:11 < peter9477> Anyone know the implications of a whack of these on app termination: qt-msg 0
  69. ERROR: Context: Object name=" "Tab" " [objectId= 88 ] not unrealized
  70.  
  71. 12. Entering a leading space in a TextField makes any following text vanish.
  72.  
  73. 13. TextField and others (?) don't respect right-alignment except for hintText
  74.  
  75. 14. FilePicker allowOverwrite in QML doesn't get respected... always treats it
  76. as though it were set to "false" (i.e. adding (1) to filename even if you
  77. try saying true)
  78.  
  79. 15. Example code in https://github.com/blackberry/Cascades-Samples/blob/master/filepicker/src/main.cpp is wrong since it tries using pickers::ViewMode rather than pickers::FilePickerViewMode
  80.  
  81. 16. leadingVisual in ListView snaps back to hidden if you change the
  82. sort order or sorting keys with a GroupDataModel. Need it to be able to
  83. stay open, in case the user is modifying both of those, or other parameters
  84. that may affect the main list, or they can only change one at a time and
  85. it "slams" shut unexpectedly.
  86.  
  87. 17. leadingVisual can be pulled down even when visible=false on the ListView
  88.  
  89. 18. context menu can slide out under your finger and it's too easy to activate
  90. it if you're not careful... need a sample
  91.  
  92. 19. privacy concern: regular wallpaper is shown on lock screen
  93.  
  94. 20. if bbm_connect permission not requested (or perhaps not granted),
  95. attemping to call bbmsp_request_events(0) results in a segfault:
  96. [BPS:INFO] BPS setting verbosity 2
  97. [BPS:INFO] channel_remove_fd: Removing fd 3 from chid 2
  98. [BP
  99. Process 20717570 (tart-launch) terminated SIGSEGV code=1 fltno=11 ip=7bee0426(/base/usr/lib/libbbmsp.so.1@_ZN3bbm18ListApiMultiplexerC2E14QSharedPointerINS_8JsonPipeEE+0x9d) mapaddr=00040426. ref=00000000
  100.  
  101. 21. calling bbmsp_get_domain() prior to calling bbmsp_request_events(0) returns
  102. a value different than when you call it after... that is, the function call
  103. is not idempotent, and in fact the value returned cannot be used unless
  104. you're calling it *after* you request the events. Is that correct behavior,
  105. and if so, is it documented?
  106.  
  107. 22. Nine-slice margins are off-by-one, and also don't work with 3x3 images.
  108. If you have a 3x3 image, with 1-pixel edges and a 1-pixel centre,
  109. specifying margins of 1 1 1 1 gives a shaded inner region, while 2 2 2 2
  110. gives only corner pixels.
  111. If you double everything for a 6x6 image with 2-pixel edges and 2-pixel center,
  112. specifying 2 2 2 2 as the docs would suggest gives a shaded inner region,
  113. but specifying 3 3 3 3 ("slicing off 3 pixels" supposedly) actually gives
  114. the desired results.
  115.  
  116. 23. After the modified bbutil_terminate() you can't reuse the window.
  117. It looks like EGL pollutes the window in some fashion so you need to
  118. destroy it rather than just reusing.
  119. Write up as an email referencing bbutil.c for seaner to pass on?
Advertisement
Add Comment
Please, Sign In to add comment