Community
Participate
Working Groups
Compile a list of all bugs fixed in September 2003 by Chris McLaren and Douglas Pollock, and write test cases that highlight the problems. Where test cases are not possible, explain why. This bug should contain all the documentation for bugs fixed. These fixes should mostly deal with the new key binding architecture.
Bug 30273 - Problems with Help Keys - Fixed by SWT on September 9th Bug 30936 - Missing action definition IDs - WorkbenchActionBuilder, ActionHandler, plugin.xml, plugin.properties - Changes to ActionHandler lead to Bug 43321 - No direct connection to key binding - Only realized as an internal bug - Fixed by us September 12th - To test we'd need to compare the actions added by WorkbenchActionBuilder with all of the actions defined in plugin.xml Bug 31700 - Limitations of previous key binding architecture fixed by new architecture - New architecture tweaked to handle this case - Fixed by us September 12th - Need OS-level events and key binding configuration API to test this bug Bug 32375 - Mouse wheel inserts 'z' on Motif - Fixed itself - Need OS-level events to test Bug 34397 - Couldn't walk backwards with Ctrl+Shift+6 through editors - Fixed by migration to new architecture - New related problems introduced, which are filed as separate bugs - Need OS-level events to test; should be tested along with Bug 31700 Bug 36420 - Importing key bindings have no effect - Can probably testing by importing a generated set of preferences, and then dumping out the key bindings for comparison - Fixed by us on September 25th Bug 36537 - Duplicate/unnecessary key bindings - Fixed by us September 22nd - Can probably test by dumping key bindings and looking for duplicate/unnecessary key bindings Bug 37280 - Provide an option to turn off the rocker effect on CyclePartAction - Fixed by us September 3rd - Can't really test without OS-level events; should be tested along with Bug 31700 Bug 38757 - Provide more information in Event - Fixed by SWT on September 22nd - Can't really test without OS-level events Bug 40023 - Update the checked state using key bindings - Can probably test using the "press(...)" method - Fixed by us September 12th - Led to Bug 43321 Bug 40388 - Using F3 in HTML file leads to error - Fixed itself - Not sure how to test Bug 40583 - Broken Workbench extension point doc links - Waiting to be applied by Chris McLaren - Not sure how to test. ********** Build M2 occurred ********** Bug 41452 - Can't bind to "command+o" - Invalidated by changes to keys preference page Bug 41676 - Next/previous for search not working anymore - Caused by changes made by JDT - Still OPEN (assigned to Chris, slated for M4) ********** Key Binding Architecture Release to HEAD ********** Bug 41926 - "Ctrl+V" gets wrong focus in Find/Replace dialog - Fixed by us on August 26th - Probably can't test without OS-level events Bug 41931 - Debugger breakpoint activation does not update editor list order (for CyclePartAction) - This is a rather obscure bug - Fixed by us on September 5th - Should be able to test, but the test might be very complicated Bug 41962 - "Ctrl+Tab" cannot be entered on KeysPreferencePage - Fixed by us on September 15th - Can't test without OS-level events Bug 42024 - API to limit key strokes in KeySequenceText - Fixed by us on August 26th - Should be able to test using only KeySequenceText Bug 42035 - Handle shifted/unshifted characters - Can partially test with public API, but a real test would require OS-level events (and different locales) - Fixed by SWT on September 22nd, and our work was done on September 24th Bug 42049 - Status line flashes (excessive updating) - Fixed by us on September 5th - Difficult to test as the effects are visual Bug 42151 - Some accelerators do not appear when workbench first opened - Fixed by us on September 3rd - Should be able to test by querying getAcceleratorText() Bug 42225 - Key events aren't consistent across platforms - Fixed by SWT on September 23rd - Can't test without OS-level events ********** Build M3 occurred ********** Bug 42465 - StackOverflow on KeysPreferencePage - Very similiar to Bug 43168, but the problem was in SWT code - Fixed by SWT on September 4th - Can't be tested without OS-level events Bug 42627 - Actions not displayed if command ID not defined - Test that the log is updated if commands aren't defined - Fixed by us on September 25th Bug 42763 - Some keys stop working properly - Couldn't reproduce; bug was closed Bug 42789 - "assert" keyword uses in "org.eclispe.ui.tests" - Renamed uses on September 12th - No useful test Bug 43011 - "Alt+" not always received by "zh_CN.GB18030" - Problem in SWT - Fixed by SWT on September 12th Bug 43118 - Formatted text for key strokes is non-standard - Fixed by us on September 25th Bug 43161 - "Esc" should be processed as a cancel action before being trapped by key bindings - Partial fix is ready for Chris McLaren to commit; complete fix needs more thought - Needs OS-level events to test Bug 43168 - StackOverflow on KeysPreferencePage - Fixed by us on September 22nd - Should be able to test without OS-level events - Note: this is very similar to Bug 42465, but the problem is in our code - A separate bug, Bug 43177, has been filed against GTK Bug 43321 - All menu items get check marks when activated by keyboard - Caused by the fix for Bug 40023 - Fix by us on September 19th - Should be able to test Bug 43538 - Ctrl+Space gives two key down events - Problem with SWT - Fixed by SWT on September 25th - Can't be testing without OS-level events Bug 43597 - setText("") resets font on carbon - Problem with SWT - Fixed by SWT on September 25th - Can easily be tested (simply snippet attach to bug) Bug 43610 - "Alt+Shift+" gives 0x00 as keyCode and character - Problem with SWT - Fixed by SWT on September 24th Bug 43613 - Meta-bug (encapsulating changes and refactoring) - No testing required - Fix by us on September 24th Bug 43800 - >16 bit characters not displayed properly - Fix by us on September 29th - Can be tested by using KeySupport directly Haven't checked for bugs fixed by Chris that I haven't been involved in.
Created attachment 6296 [details] version 1 Provides test coverage for three bugs: Bug 43321, Bug 43597 and Bug 43800.
Still working... SUMMARY ------- testable - Bug 36420, Bug 36537, Bug 40023, Bug 41676, Bug 41931, Bug 42024, Bug 42035, Bug 42627, Bug 43168, Bug 43321, Bug 43597, Bug 43800 need OS - Bug 30273, Bug 31700, Bug 32375, Bug 38757, Bug 34397, Bug 37280, events Bug 41926, Bug 41962, Bug 42225, Bug 42465, Bug 43011, Bug 43161, Bug 43538, Bug 43610 not worth - Bug 40388, Bug 40583, Bug 42789, Bug 43118 testing no test - Bug 41452, Bug 42763, Bug 43613 required cannot - Bug 42049 test
Also test Bug 42418 if possible.
Bug 42418, Bug 41931, and Bug 41676 still need test cases. Bug 42627 still needs a comment from Chris, but a stub test case is in CVS.
bug 42627 is now reopened for review in m5 (regarding logging only - the main problem is fixed)
Bug 41676 will not be fixed for M4. Bug 42418 is not testable. Bug 41931 is the only one left from M4.
FINAL LIST FOR M4 Summary: ------- Total Fixed 36 Testable 11 (31%) Waiting for Fixes* 02 (6%) Need OS Events 15 (42%) Not Worth Testing 04 (11%) No Test Required 03 (8%) Can't Test 01 (3%) * means the bug is fixed, but the test case requires a bug to be fixed. :) List of Bugs: ------------ testable - Bug 36420, Bug 41931, Bug 42024, Bug 42035, Bug 43168, Bug 43321, Bug 43597, Bug 43800 waiting - Bug 36537, Bug 40023 for fixes need OS - Bug 30273, Bug 31700, Bug 32375, Bug 38757, Bug 34397, Bug 37280, events Bug 41926, Bug 41962, Bug 42225, Bug 42418, Bug 42465, Bug 43011, Bug 43161, Bug 43538, Bug 43610 not worth - Bug 40388, Bug 40583, Bug 42789, Bug 43118 testing no test - Bug 41452, Bug 42763, Bug 43613 required cannot - Bug 42049 test
What do you mean by "waiting for fixes" on Bug 36537 and Bug 40023? These are both resolved as fixed.
To quote from comment #8: "means the bug is fixed, but the test case requires a bug to be fixed"