Bug #76082 | Trigger changes are not retained | ||
---|---|---|---|
Submitted: | 1 Mar 2015 0:37 | Modified: | 18 Mar 2015 18:35 |
Reporter: | Sherzod Ruzmetov | Email Updates: | |
Status: | Closed | Impact on me: | |
Category: | MySQL Workbench: Modeling | Severity: | S1 (Critical) |
Version: | 6.2.5 | OS: | Linux (Linux 3.18.7-200.fc21.x86_64 #1 SMP Wed Feb 11 21:53:17 UTC 2015 x86_64 x86_64) |
Assigned to: | CPU Architecture: | Any |
[1 Mar 2015 0:37]
Sherzod Ruzmetov
[1 Mar 2015 0:49]
Sherzod Ruzmetov
Now I received an updated error message during the crash: (mysql-workbench-bin:19736): Gtk-CRITICAL **: IA__gtk_progress_set_percentage: assertion 'percentage >= 0 && percentage <= 1.0' failed ** GLib-GObject:ERROR:gsignal.c:756:emission_pop: code should not be reached
[4 Mar 2015 17:35]
Sherzod Ruzmetov
trigger_BEFORE_UPDATE.sql that still keeps crashing mysql-workbench
Attachment: invoice_content_BEFORE_UPDATE.sql (application/sql, text), 1.53 KiB.
[4 Mar 2015 17:39]
Sherzod Ruzmetov
I just upgraded to 6.2.5. In most tables I tested triggers can be edited just fine. However, on this one particular table that I have it crashes after I make chanage to trigger content, and navigate away. I uploaded the trigger's source (see attachment "trigger_BEFORE_UPDATE.sql")
[4 Mar 2015 17:46]
Sherzod Ruzmetov
I confirm that it's the content of the trigger that's causing the crash. To test it i removed all the triggers from that particular table. Create a "hello world" trigger. It works, saves, updates, and does not crash. But once I insert the actual trigger content (attached above) it crashes again. I'll start re-typing the trigger code a new and see which statement is the culprit. Btw, the error during the crash is the same
[5 Mar 2015 13:49]
Sherzod Ruzmetov
I finally narrowed down the cause of this crash: it was the begin block label!!! If I remove the "update_fifo_layer:" label before the BEGIN it stops crashing! I double-checked this by adding labels for other BEGIN blocks in other triggers, mysql-workbench crashes furiously!!! Now the workaround for me is to re-write the triggers without having to label the blocks. However it's fine with me labeling the loops. If someone could tell me where in the source code I should be looking for this perhaps I could give a hand to locate this bug?
[5 Mar 2015 17:31]
MySQL Verification Team
Thank you for the bug report. (mysql-workbench-bin:30466): GLib-GObject-WARNING **: attempting to add an interface (GtkTreeDragDest) to class (gtkmm__CustomObject_16ListModelWrapper) after class_init (mysql-workbench-bin:30466): GLib-GObject-WARNING **: attempting to add an interface (GtkTreeDragSource) to class (gtkmm__CustomObject_16ListModelWrapper) after class_init (mysql-workbench-bin:30466): glibmm-ERROR **: unhandled exception (type std::exception) in signal handler: what: basic_string::substr: __pos (which is 18446744073701655208) > this->size() (which is 0) [miguel@vbfed21 ~]$
[18 Mar 2015 18:35]
Philip Olson
Posted by developer: Fixed as of the upcoming MySQL Workbench 6.3.3 release, and here's the changelog entry: The trigger editor would fail to save data for triggers including the "BEGIN ... END" construct. It would also generate a "GLib-GObject:ERROR:gsignal.c:756:emission_pop: code should not be reached" error. Thank you for the bug report.