Lazarus Release Candidate 1 of 3.0

[dsiders]

Is anyone else seeing an EAccessViolation under a Linux x86_64 build with the Gtk2 widget set when debugging?

I'm unable to reproduce the issue as described. I got the issue that the assembler window keeps popping up instead.

I did came across the following though:

Code: [Select]

lazarus stuff
WARNING: TGtk2WidgetSet.InvalidateRect refused invalidating during paint message: TLazVirtualStringTree
TIssueManager.ReadFileIssues failed: In 'file:///media/ramdisk/work/lazarus/3.0rc1/lcl/interfaces/gtk3/issues.xml' (line 41 pos 5): Unmatching element end tag (expected "</short>")

during debugging
(lazarus:160656): GLib-CRITICAL **: 02:48:00.443: Source ID 11341 was not found when attempting to remove it
(lazarus:163024): GLib-GObject-WARNING **: 02:52:38.624: ../../../gobject/gsignal.c:1285: no emission of signal "button-release-event" to stop for instance '0x32fad90'

And to answer: yes, gtk2 for both Lazarus and Project 

The issues.xml problem has been fixed in https://gitlab.com/freepascal.org/lazarus/lazarus/-/commit/6717565e9b4b700d79ff3dc89f5f33180200dbb7.

[TRon]

@dsiders: Thank you !

[Martin_fr]

If I create a breakpoint in the OnClick() handler and step through it once, and then continue running the app, when I exit the application I am seeing an EAccessViolation from Lazarus on the console and Lazarus freezes. I was using FpDebugger with dwarf3, same also occurred using dwarf2 w/ sets, and when using Gdb instead of FpDebugger.

If it happens again, can you copy the printout from the console and paste it here (or as part of a bug report)?

There usually is a stack trace being printed. If it does not have unit/line info, then go to menu Tools>Congifure build lazarus and add the custom options
-gw3 -gl
and rebuild.
Then see if the issue happens again.

Have you installed any packages?

[Martin_fr]

I got the issue that the assembler window keeps popping up instead.

With breakpoint? normal stepping? Caught Exceptions?

Details?
How to reproduce?


This is normal if the debugger pauses at code that has no debug info => eg in the RTL.

So when you got an access violation in the RTL....

[TRon]

Sorry Martin_fr,  i thought the information detailed by user alink (post) was enough.

With breakpoint? normal stepping? Caught Exceptions?

Details?
How to reproduce?

Writing a small program as instructed, setting a breakpoint at the onclick event (begin), build the program, then run it with the debugger, press the button, then the debugger pops-up with the assembler window. No errors, exceptions or anything else. Just stepping through till just after the increase of the variable then let the program continue to run.

[T-Khan]

Unfortunately, https://wiki.lazarus.freepascal.org/Lazarus_3.0_release_notes doesn't say anything about changed default behavior for TToggleBox and the like and about introduction of TCustomCheckBox.VCL_OnClick_Emulation. Lost about 20 minutes trying to understand if I'm going crazy and why clicks don't count or, as it turned out, now count twice.

[Martin_fr]

@TRon: https://forum.lazarus.freepascal.org/index.php?topic=63931 for follow up please.

[T-Khan]

Found one more thing. When I try to load an old project with ComboBoxEx with ReadOnly property set to True, which was deleted in later versions for some reason, I'm getting an error, and then being presented with an option to delete unsupported properties. I accept, but nothing happens seemingly, the program compiles fine, and when I run the program I get "Error reading ComboBoxFonts.ReadOnly: Unknown property: "ReadOnly". When I try to close Lazarus after this, I get a bunch of access violations from it also.
Eventually, only editing form file by hand does the trick.

[alank]

If I create a breakpoint in the OnClick() handler and step through it once, and then continue running the app, when I exit the application I am seeing an EAccessViolation from Lazarus on the console and Lazarus freezes. I was using FpDebugger with dwarf3, same also occurred using dwarf2 w/ sets, and when using Gdb instead of FpDebugger.

If it happens again, can you copy the printout from the console and paste it here (or as part of a bug report)?

There usually is a stack trace being printed. If it does not have unit/line info, then go to menu Tools>Congifure build lazarus and add the custom options
-gw3 -gl
and rebuild.
Then see if the issue happens again.

Have you installed any packages?

Martin,

I recompiled Lazarus with the following Options: -g -gl -gw3 -O1

Here is the stack trace that is generated (no file names or line numbers displayed, despite the specified options):

Code: Bash  [Select][+][-]

1. TApplication.HandleException: EAccessViolation
2. Access violation
3.   Stack trace:
4.   $00007F0D4DE94352
5.   $00007F0D4E0A7A46
6.   $00007F0D4DE7BA53
7.   $00007F0D4E296F2E
8.   $00007F0D4DE7BA53
9.   $00007F0D4E28FD66
10.   $00007F0D4E276CD9
11.   $00007F0D4DE6D44C
12.   $00007F0D4DE9C2A8
13.   $00007F0D4DE8CCBD
14.   $00007F0D4DE8CF33
15.   $00007F0D4E18E931
16.   $00007F0D4DE7BAAA
17.   $00007F0D4E1D348D
18.   $00007F0D4E0CC084
19.   $00007F0D4DE6D4EA
20.   $00007F0D4DE9C2A8
21. lazarus.pp - unhandled exception
22.  

This access violation occurs any time I set a breakpoint in the project and exit the running project, even if that breakpoint is not triggered. Removing the breakpoint allows lazarus to "recover" from running the project correctly. With a breakpoint set, Lazarus freezes on application exit.

Packages I install with each Lazarus build beyond those provided in bigide are: anchordockingdsgn 1.0, sparta_dockedformeditor 0.0, lazdaemon 0.9.9, and weblaz 1.0.

I'm running this on Fedora 38 x86_64 with Fedora 38 repository installations of fpc and fpc-src.

Alan

[Martin_fr]

sparta_dockedformeditor

There is a package without "sparta", just "dockedformeditor". The sparta package isn't recommended. (Most of the sparta package have some issues).

Strange there should have been unit/line names. Though it may be in some OS library...


could you try 
   gdb path/to/lazarus

If your lazarus is installed into a global dir (/usr/.....) and you use startlazarus, you need to find the real lazarus to use it directly. Usually it is somewhere in ~/.lazarus.

then
    r

and when the crash happens
   bt

[alank]

Strange there should have been unit/line names. Though it may be in some OS library...


could you try 
   gdb path/to/lazarus

If your lazarus is installed into a global dir (/usr/.....) and you use startlazarus, you need to find the real lazarus to use it directly. Usually it is somewhere in ~/.lazarus.

then
    r

and when the crash happens
   bt

Here's the lengthy backtrace from gdb:

Code: Pascal  [Select][+][-]

1. Thread 1 "lazarus" received signal SIGSEGV, Segmentation fault.
2. 0x00007ffff7894352 in g_type_check_instance_is_fundamentally_a (type_instance=0x2dbd160,
3.     fundamental_type=0x50) at ../gobject/gtype.c:4167
4. Downloading source file /usr/src/debug/glib2-2.76.3-1.fc38.x86_64/redhat-linux-build/../gobject/gtype.c
5. 4167      return node && (NODE_FUNDAMENTAL_TYPE(node) == fundamental_type);                  
6. (gdb) bt
7. #0  0x00007ffff7894352 in g_type_check_instance_is_fundamentally_a
8.     (type_instance=0x2dbd160, fundamental_type=0x50) at ../gobject/gtype.c:4167
9. #1  0x00007ffff787b822 in g_object_unref (_object=0x2dbd160) at ../gobject/gobject.c:3807
10. #2  0x00007ffff7aa7a46 in gtk_cell_renderer_pixbuf_finalize (object=0x2f5e040)
11.     at /usr/src/debug/gtk2-2.24.33-13.fc38.x86_64/gtk/gtkcellrendererpixbuf.c:224
12. #3  0x00007ffff787ba53 in g_object_unref (_object=0x2f5e040) at ../gobject/gobject.c:3938
13. #4  g_object_unref (_object=0x2f5e040) at ../gobject/gobject.c:3802
14. #5  0x00007ffff7c96f2e in gtk_tree_view_column_finalize (object=0x2f5df40)
15.     at /usr/src/debug/gtk2-2.24.33-13.fc38.x86_64/gtk/gtktreeviewcolumn.c:418
16. #6  0x00007ffff787ba53 in g_object_unref (_object=0x2f5df40) at ../gobject/gobject.c:3938
17. #7  g_object_unref (_object=0x2f5df40) at ../gobject/gobject.c:3802
18. #8  0x00007ffff7c8fd66 in IA__gtk_tree_view_remove_column
19.     (tree_view=0x2f5b840, column=0x2f5df40)
20.     at /usr/src/debug/gtk2-2.24.33-13.fc38.x86_64/gtk/gtktreeview.c:11274
21. #9  0x00007ffff7c76cd9 in gtk_tree_view_destroy (object=0x2f5b840)
22.     at /usr/src/debug/gtk2-2.24.33-13.fc38.x86_64/gtk/gtktreeview.c:1616
23. #10 0x00007ffff786d44c in g_closure_invoke
24.     (closure=0x2be2a20, return_value=0x0, n_param_values=1, param_values=0x7fffffffba50, invocation_hint=0x7fffffffb9d0) at ../gobject/gclosure.c:832
25. #11 0x00007ffff789c2a8 in signal_emit_unlocked_R.isra.0
26.     (node=node@entry=0x2bb8150, detail=detail@entry=0, instance=instance@entry=0x2f5b840, emission_return=emission_return@entry=0x0, instance_and_params=instance_and_params@entry=0x7fffffffba50) at ../gobject/gsignal.c:3930
27. #12 0x00007ffff788ccbd in g_signal_emit_valist
28.     (instance=<optimized out>, signal_id=<optimized out>, detail=<optimized out>, var_args=var_args@entry=0x7fffffffbbf0) at ../gobject/gsignal.c:3565
29. #13 0x00007ffff788cf33 in g_signal_emit
30.     (instance=instance@entry=0x2f5b840, signal_id=<optimized out>, detail=detail@entry=0)
31.     at ../gobject/gsignal.c:3622
32. #14 0x00007ffff7b8e931 in gtk_object_dispose (gobject=0x2f5b840)
33.     at /usr/src/debug/gtk2-2.24.33-13.fc38.x86_64/gtk/gtkobject.c:421
34. #15 0x00007ffff787baaa in g_object_run_dispose (object=0x2f5b840)
35.     at ../gobject/gobject.c:1448
36. #16 g_object_run_dispose (object=0x2f5b840) at ../gobject/gobject.c:1441
37. #17 0x00007ffff7bd348d in gtk_scrolled_window_forall
38.     (container=0x2f5ab70, include_internals=0, callback=0x7ffff7ca51e0 <IA__gtk_widget_destroy>, callback_data=0x0)
39.     at /usr/src/debug/gtk2-2.24.33-13.fc38.x86_64/gtk/gtkscrolledwindow.c:1082
40. #18 0x00007ffff7acc084 in gtk_container_destroy (object=0x2f5ab70)
41.     at /usr/src/debug/gtk2-2.24.33-13.fc38.x86_64/gtk/gtkcontainer.c:1073
42. #19 0x00007ffff786d4ea in g_closure_invoke
43.     (closure=0x2be2a20, return_value=0x0, n_param_values=1, param_values=0x7fffffffbf20, invocation_hint=0x7fffffffbea0) at ../gobject/gclosure.c:832
44. #20 0x00007ffff789c2a8 in signal_emit_unlocked_R.isra.0
45.     (node=node@entry=0x2bb8150, detail=detail@entry=0, instance=instance@entry=0x2f5ab70, emission_return=emission_return@entry=0x0, instance_and_params=instance_and_params@entry=0x7fffffffbf20) at ../gobject/gsignal.c:3930
46. #21 0x00007ffff788ccbd in g_signal_emit_valist
47.     (instance=<optimized out>, signal_id=<optimized out>, detail=<optimized out>, var_args=var_args@entry=0x7fffffffc0c0) at ../gobject/gsignal.c:3565
48. #22 0x00007ffff788cf33 in g_signal_emit
49.     (instance=instance@entry=0x2f5ab70, signal_id=<optimized out>, detail=detail@entry=0)
50.     at ../gobject/gsignal.c:3622
51. #23 0x00007ffff7b8e931 in gtk_object_dispose (gobject=0x2f5ab70)
52.     at /usr/src/debug/gtk2-2.24.33-13.fc38.x86_64/gtk/gtkobject.c:421
53. #24 0x00007ffff787baaa in g_object_run_dispose (object=0x2f5ab70)
54.     at ../gobject/gobject.c:1448
55. #25 g_object_run_dispose (object=0x2f5ab70) at ../gobject/gobject.c:1441
56. #26 0x000000000064c2be in DestroyWidget (Widget=0x2f5ab70) at gtk2/gtk2proc.inc:4247
57. #27 0x00000000004fadf2 in DestroyConnectedWidget
58.     (this=0x7ffff6856250, Widget=0x2f5ab70, CheckIfDestroying=false)
59.     at gtk2/gtk2widgetset.inc:4960
60. #28 0x00000000004fab29 in DestroyLCLComponent (this=0x7ffff6856250, Sender=0x7fffc7ca17d0)
61.     at gtk2/gtk2widgetset.inc:4871
62. #29 0x000000000078eb76 in DestroyHandle (self=0x1dbbe80, AWinControl=0x7fffc7ca17d0)
63.     at gtk2/gtk2wscontrols.pp:565
64. #30 0x00000000007f57d3 in DestroyHandle (self=0x7fffc7e38750, AWinControl=0x7fffc7ca17d0)
65.     at gtk2/gtk2wscustomlistview.inc:1832
66. #31 0x00000000005935e6 in DestroyWnd (this=0x7fffc7ca17d0) at include/wincontrol.inc:7853
67. #32 0x00000000005fbb82 in DestroyWnd (this=0x7fffc7ca17d0)
68. --Type <RET> for more, q to quit, c to continue without paging--
69.     at include/customlistview.inc:1201
70. #33 0x000000000058ddd2 in DestroyHandle (this=0x7fffc7ca17d0) at include/wincontrol.inc:5333
71. #34 0x000000000058ddbc in DestroyHandle (this=0x7fffc67111f0) at include/wincontrol.inc:5329
72. #35 0x000000000058ddbc in DestroyHandle (this=0x7fffc6712b10) at include/wincontrol.inc:5329
73. #36 0x000000000058ddbc in DestroyHandle (this=0x7fffc7ca9a90) at include/wincontrol.inc:5329
74. #37 0x00000000005ea282 in DestroyHandle (this=0x7fffc7ca9a90) at include/custompage.inc:206
75. #38 0x00000000005ec3d7 in AddRemovePageHandle (this=0x7fffc7ca8e30, APage=0x7fffc7ca9a90)
76.     at include/customnotebook.inc:738
77. #39 0x00000000005ec496 in RemovePage (this=0x7fffc7ca8e30, Index=0)
78.     at include/customnotebook.inc:760
79. #40 0x00000000005e9edc in SetParent (this=0x7fffc7ca9a90, AParent=0x0)
80.     at include/custompage.inc:119
81. #41 0x00000000005eabbc in Delete (this=0x7fffc67558c0, Index=0)
82.     at include/customnotebook.inc:136
83. #42 0x00000000005eab69 in Clear (this=0x7fffc67558c0) at include/customnotebook.inc:116
84. #43 0x00000000005eb379 in Destroy (this=0x7fffc7ca8e30, $vmt=0x1)
85.     at include/customnotebook.inc:334
86. #44 0x000000000043323b in FREE (this=0xffff5d5dff00007c) at ../inc/objpas.inc:336
87. #45 0x0000000000000001 in  ()
88. #46 0x00000000004738a7 in FREEANDNIL (OBJ=0) at ../objpas/sysutils/sysutils.inc:160
89. #47 0x00007fffffffcff0 in  ()
90. #48 0x000000000134702c in FreePages (this=0x7fffc7ca76b0) at anchordocking.pas:4795
91. #49 0x000000000134c81b in Destroy (this=0x7fffc7ca76b0, $vmt=0x1) at anchordocking.pas:6454
92. #50 0x000000000043323b in FREE (this=0xffff5d5dff00007c) at ../inc/objpas.inc:336
93. #51 0x00007fffffffd070 in  ()
94. #52 0x000000000133d0f7 in CloseUnneededAndWronglyLocatedControls
95.     (this=0x7fffe3f98040, Tree=0x7fffe3fab8c0) at anchordocking.pas:1922
96. #53 0x000000000133fc85 in FullRestoreLayout
97.     (this=0x7fffe3f98040, Tree=0x7fffe3fab8c0, Scale=true) at anchordocking.pas:2627
98. #54 0x000000000142ca75 in RestoreDesktop (this=0x7fffe3fab880)
99.     at anchordesktopoptions.pas:297
100. #55 0x0000000000dad0c0 in RESTOREDESKTOP (this=0x7ffff6821e20) at envguioptions.pas:701
101. #56 0x0000000000db0ef5 in USEDESKTOP (this=0x7fffe3f5aac0, ADESKTOP=0x7fffc6b3f9a0)
102.     at envguioptions.pas:1278
103. #57 0x0000000000db1135 in DISABLEDEBUGDESKTOP (this=0x7fffe3f5aac0)
104.     at envguioptions.pas:1317
105. #58 0x00000000004b5e67 in SETTOOLSTATUS (this=0x7fffe3edc110, AVALUE=itNone) at main.pp:3906
106. #59 0x0000000000b290b9 in UPDATETOOLSTATUS (this=0x7fffe3fb8e40) at debugmanager.pas:2311
107. #60 0x0000000000b25672 in DEBUGGERCHANGESTATE
108.     (this=0x7fffe3fb8e40, ADEBUGGER=0x7fffc7c64f90, OLDSTATE=dsRun) at debugmanager.pas:1483
109. #61 0x0000000000b5cb76 in DoState (this=0x7fffc7c64f90, OldState=dsRun)
110.     at dbgintfdebuggerbase.pp:5493
111. #62 0x00000000012f5195 in DoState (this=0x7fffc7c64f90, OldState=dsRun)
112.     at fpdebugdebugger.pas:4328
113. #63 0x0000000000b5e862 in SetState (this=0x7fffc7c64f90, AValue=dsStop)
114.     at dbgintfdebuggerbase.pp:5908
115. #64 0x00000000012f153f in FDbgControllerProcessExitEvent (this=0x7fffc7c64f90, AExitCode=0)
116.     at fpdebugdebugger.pas:3437
117. #65 0x00000000010c64f8 in SendEvents (this=0x7fffc72bd2a0, continue=true)
118.     at fpdbgcontroller.pas:2052
119. #66 0x00000000012f4a4d in DebugLoopFinished (this=0x7fffc7c64f90, Data=0)
120.     at fpdebugdebugger.pas:4210
121. #67 0x00000000012e7e49 in LoopFinished_DecRef (this=0x7fffc75da1c0, Data=0)
122.     at fpdebugdebugger.pas:707
123. #68 0x0000000000491342 in ProcessAsyncCallQueue (this=0x7ffff6855bf0)
124.     at include/application.inc:1073
125. #69 0x000000000048f993 in Idle (this=0x7ffff6855bf0, Wait=true)
126.     at include/application.inc:440
127. #70 0x0000000000491b66 in HandleMessage (this=0x7ffff6855bf0)
128.     at include/application.inc:1269
129. #71 0x000000000049214f in RunLoop (this=0x7ffff6855bf0) at include/application.inc:1408
130. #72 0x0000000000636fc3 in AppRun (this=0x7ffff6856250, ALoop=...)
131.     at include/interfacebase.inc:54
132. #73 0x00000000004920f7 in Run (this=0x7ffff6855bf0) at include/application.inc:1396
133. #74 0x0000000000421846 in $main () at lazarus.pp:157
134. (gdb)
135.  

Thanks for the tip on the sparta docked form editor - I'll try removing it and installing the regular dockedformeditor alone.

Alan

[darthvader]

Hi ,
I just installed the v3 rc1 and find a bug (i think)
From package manager i try to install GLSCENE and it go into compilation error , this was ok on the older versions.
SDPO and LazSerial install was ok from package manager.
If someone know how to fix it for glscene it will be great , i use it often on my projects

[magu]

Well done guys!
Much appreciated.

[Martin_fr]

Here's the lengthy backtrace from gdb:

Ok, thanks. That is useful. (Well as useful as it will get).

That it crashes when you had a breakpoint is likely pure coincidence.
It is most likely a double free or dangling pointer => And those crash or don't crash depending on memory layout, which can be changed by any unrelated action.

So far, I can tell it is something somewhere in a TListView, and how it is freed by the gtk2 ws.

[Martin_fr]

Here's the lengthy backtrace from gdb:

Ok, thanks. That is useful. (Well as useful as it will get).

Ok, if you have some time.... one more think you could do. It's a bit time consuming, and requires some tools to be installed.
   So if you can't do it, ok.

==> valgrind

Ever used it?

It should be installable with whatever installer tools come with your distro (yum or whatever)

Code: Text  [Select][+][-]

1. valgrind --tool=memcheck --num-callers=30 --log-file=laz.trc   ./lazarus

Mind => your IDE will run crawling slow. So to to all the steps you did to get the error => could be 10 to 15 minutes.

If the IDE crashes, the logfile will contain some useful info.

***** EDIT: I forgot, for this the IDE must be recompiled with -gv

---------------------

Otherwise, I believe you may have set up a debug desktop? Changing the open windows when start/stop debugging? Which windows are opened and closed?

Come to think of it, it may be the breakpoint window => it contain a TListView (IIRC) => and then having a breakpoint in it, may just be what makes the difference (so not a coincidence).
=> Is your breakpoint window docked? docked on its own, or into a tab-group?

Then I can use that to try and see if I can valgrind the issue.