Community
Participate
Working Groups
Build 20020814 2.0.1 1. Open a New Resource Perspective. 2. Customize the perspective so that the External Tools coolItem and Search coolItem are the only cool items on the row (along with the default UI coolItem). 3. Move the Search coolItem and External Tools coolItem to their own row, so that the coolbar looks like this: UI Srch ExtTools 4. Customize the perspective and uncheck Search and check CVS, which will remove Srch from row 2 and add the CVS coolItem to the end of row 2. You will see noticeable flash/jumpiness as the coolbar is collapsed to 1 row and then re-expanded to 2 rows when the delete takes place - even though this action is not necessary since ExtTools still exists on the row. If you decide to use setRedraw(false) around the code that is causing this, can you please look into [Bug 17994]. setRedraw is still noticeably flashy due to this PR (i.e., you see all the grippers on the coolbar flash). Other than that, the painting is smooth (i.e., you don't see tool items flashing just the grippers). This NEEDS to be fixed for 2.0.1. It leads to way too much jumpiness on the coolbar (i.e., whenever the first item in a row is removed) and is much more apparent now that coolitem positions are being remembered. For example, if you have the *.java editor coolitem as the first item on a row and switch back and forth between a *.java file and a *.class file, you will see major flash/jumpiness.
Released temporary workaround for the 20020815 2.0.1 build (want to wsad to see the coolbar remember position behavior without this bug). Also released the same change into the 2.1 dev stream. You'll be able to reproduce the behavior I describe above in builds <= 20020814. Still want a fix for this on your end.
the fix for this bug is not related with using setRedraw. During the destroyItem the following step are perform: (assuming that: the coobar has R rows, deleting the item at index I) a) Detect if the item is a wrap b) Delete the item, if the item was a wrap it will cause WM_SIZE, what causes the workbench to layout for R - 1 rows. c) If the item was wrap AND now there is a new item in index 'I' make the new item be a wrap, what causes the workbench to layout back to R rows. The fix is stop the WM_SIZE, but this have to be done very carefully. I will study more this case and try to submit the fix as soom as I can.
Fixed > 20020815 and R2.0.1
