Community
Participate
Working Groups
The user-visible effect is that the stepping control actions do not become disabled when stepping over a long-running operation. We found a bug in the SteppingController in cases where the whole process/container is resumed in response to a step request. The occurs when the IResumedEvent.getReason() != StateChangeReason.STEP. There's probably a few different methods to fix this behavior. My fix changes the handling of the timeout timer to start when the step command is executed successfully instead of when the resume event is received (I'll submit a patch in a bit).
Created attachment 204485 [details] Patch with fix.
Toni, could you please review my fix attempt. I tested it and it works but I'm not sure about potential side effects.
(In reply to comment #2) > Toni, could you please review my fix attempt. I tested it and it works but I'm > not sure about potential side effects. Looks good. There is just one subtle difference in public void eventDispatched(final ISuspendedDMEvent e) When the context has already timed out, the step queue was not processed immediately. I guess this makes sense to give a user the chance to stop stepping early enough. To avoid any surprises, I think this should be kept the same.
CQ:WIND00307786
Created attachment 204563 [details] Updated fix. I've been testing and debugging the patch for most of the day. I found a couple of issues and fixed them in this updated patch.
I pushed the fix to master and cdt8_0. (In reply to comment #3) > ... > public void eventDispatched(final ISuspendedDMEvent e) I almost forgot, made an additional commit.
*** cdt git genie on behalf of Pawel Piech *** Bug 359783 - [run control] SteppingController step timeout doesn't work when container is resumed [*] http://git.eclipse.org/c/cdt/org.eclipse.cdt.git/commit/?id=d643d3b26edd27abb0bd80c4c3621f00c95277b7
*** cdt git genie on behalf of Pawel Piech *** Bug 359783 - [run control] SteppingController step timeout doesn't work when container is resumed [*] http://git.eclipse.org/c/cdt/org.eclipse.cdt.git/commit/?id=53debd58aeaa953aa4fb7c4cdc98e4161b9fbfec
*** cdt git genie on behalf of Pawel Piech *** Bug 359783 - (comment #3: Fixed handling of suspended event to match previous behavior) [*] http://git.eclipse.org/c/cdt/org.eclipse.cdt.git/commit/?id=93d76c18d7b1444d206bb6311120a3a34d783275
*** cdt git genie on behalf of Pawel Piech *** Bug 359783 - (comment #3: Fixed handling of suspended event to match previous behavior) [*] http://git.eclipse.org/c/cdt/org.eclipse.cdt.git/commit/?id=2c9c30c92e08f928e6fb849265e4a2694aae12f3
I spent another day of testing and debugging the SteppingController, and found some more issues (this module is turning into a black hole for me... sucking up all available time-space). I pushed out a couple more fixes.
*** cdt git genie on behalf of Pawel Piech *** Bug 359783 - (Fix matching of timed-out event contexts and tracking timed-out flags) [*] http://git.eclipse.org/c/cdt/org.eclipse.cdt.git/commit/?id=d722e5a8f615e0368a3aa35c2d4fcbefb8bbebeb
*** cdt git genie on behalf of Pawel Piech *** Bug 359783 - (Avoid assertion error in case of inconsistent model events) [*] http://git.eclipse.org/c/cdt/org.eclipse.cdt.git/commit/?id=bdd74fd79e98c593e9e4eeba96843d278bb60ddc
*** cdt git genie on behalf of Pawel Piech *** Bug 359783 - (Fix matching of timed-out event contexts and tracking timed-out flags) [*] http://git.eclipse.org/c/cdt/org.eclipse.cdt.git/commit/?id=8433a8acd97cc783c9de2d794df00e7f01722716
*** cdt git genie on behalf of Pawel Piech *** Bug 359783 - (Avoid assertion error in case of inconsistent model events) [*] http://git.eclipse.org/c/cdt/org.eclipse.cdt.git/commit/?id=732b4c86abe03349b540f27fbddbf4e8405032a8
