Opened 5 years ago

Last modified 4 years ago

#322 new defect

undefined pt server breaks assert (crash in DEBUG)

Reported by: aboeckma Owned by: devel
Priority: major Milestone:
Component: global Version: SVN
Keywords: Cc:

Description

  • click Probes → Probe Design
  • select -undefined- from the pt server selection drop down
  • click GO
  • BOOOM

Stack trace:

assertion 'id >= 0' failed in adtcp.cxx #325
-------------------- ARB-backtrace 'ARB_SIGSEGV':
0   arb_ntree                           0x000000010021e369 _Z18GBK_dump_backtraceP7__sFILEPKc + 41
1   arb_ntree                           0x00000001001e1fcd _Z16GBS_ptserver_tagi + 109
2   arb_ntree                           0x0000000100127200 _ZL27PD_probe_pt_look_for_serverP7AW_rootRPKc + 64
3   arb_ntree                           0x000000010011f9f1 _ZL18probe_design_eventP9AW_windowl + 129
4   arb_ntree                           0x00000001002779d1 _ZN12AW_cb_struct12run_callbackEv + 817
5   arb_ntree                           0x0000000100295931 _ZN9AW_window13click_handlerEP10_GtkWidgetPv + 353
6   libgobject-2.0.0.dylib              0x0000000101336a35 _g_closure_invoke_va + 309
7   libgobject-2.0.0.dylib              0x00000001013470d8 g_signal_emit_valist + 1352
8   libgobject-2.0.0.dylib              0x0000000101347d24 g_signal_emit + 116
9   libgtk-x11-2.0.0.dylib              0x0000000100897a0f gtk_real_button_released + 127
10  libgobject-2.0.0.dylib              0x0000000101336a35 _g_closure_invoke_va + 309
11  libgobject-2.0.0.dylib              0x00000001013470d8 g_signal_emit_valist + 1352
12  libgobject-2.0.0.dylib              0x0000000101347d24 g_signal_emit + 116
13  libgtk-x11-2.0.0.dylib              0x0000000100897663 gtk_button_button_release + 99
14  libgtk-x11-2.0.0.dylib              0x00000001009423ef _gtk_marshal_BOOLEAN__BOXED + 159
15  libgobject-2.0.0.dylib              0x0000000101336bc4 g_closure_invoke + 308
16  libgobject-2.0.0.dylib              0x000000010134683a signal_emit_unlocked_R + 2218
17  libgobject-2.0.0.dylib              0x00000001013477f8 g_signal_emit_valist + 3176
18  libgobject-2.0.0.dylib              0x0000000101347d24 g_signal_emit + 116
19  libgtk-x11-2.0.0.dylib              0x0000000100a596ac gtk_widget_event_internal + 620
20  libgtk-x11-2.0.0.dylib              0x00000001009404df gtk_propagate_event + 383
21  libgtk-x11-2.0.0.dylib              0x000000010094013e gtk_main_do_event + 1294
22  libgdk-x11-2.0.0.dylib              0x0000000100d2a081 gdk_event_dispatch + 81
23  libglib-2.0.0.dylib                 0x00000001013b18f8 g_main_context_dispatch + 328
24  libglib-2.0.0.dylib                 0x00000001013b3a0e g_main_context_iterate + 510
25  libglib-2.0.0.dylib                 0x00000001013b4faf g_main_loop_run + 287
26  libgtk-x11-2.0.0.dylib              0x000000010093f94f gtk_main + 191
27  arb_ntree                           0x000000010028d831 _ZN7AW_root9main_loopEv + 17
28  arb_ntree                           0x000000010002cc66 _ZL11startup_guiR16NtreeCommandLineR9ARB_ERROR + 3062
29  arb_ntree                           0x000000010002bf52 _Z8ARB_mainiPPc + 290
30  arb_ntree                           0x0000000100001cf2 main + 34
31  libdyld.dylib                       0x00007fff922517e1 start + 0
-------------------- End of backtrace

Change History (2)

comment:1 follow-up: Changed 4 years ago by epruesse

  • Component changed from Library (GUI) to global
  • Summary changed from Design Probe with undefined pt server crashes ARB to undefined pt server breaks assert (crash in DEBUG)
  • Version changed from gtkport to SVN

Changing from gtkport to trunk, this happens always.

Should there even be "-undefined-" as possibility in the PT server choice? It's probably easier to ensure that there is at least one defined in arb_tcp.dat at startup and remove "-undefined-" from the list (confusing to users anyway) than to add error handling everywhere (not just probe design affected).

comment:2 in reply to: ↑ 1 Changed 4 years ago by westram

Should there even be "-undefined-" as possibility in the PT server choice? It's probably easier to ensure that there is at least one defined in arb_tcp.dat at startup and remove "-undefined-" from the list (confusing to users anyway) than to add error handling everywhere (not just probe design affected).

TODO

  • remove "-undefined-" from pt-server selection lists
  • use first user-specific ptserver as default (e.g. "ralf1")
    • doing so will move that entry to bottom in motif version (accept, will be fixed by swap to gtk).
  • make sure awar-settings pointing to "-undefined-" and saved to properties get superseeded properly
Note: See TracTickets for help on using tickets.