Author Topic: The 15 August 2011 build (7386) DEBUGGER BRANCH version is out.  (Read 22168 times)

Offline Oleg_Sam

  • Advanced newcomer
  • *
  • Posts: 25
Re: The 15 August 2011 build (7386) DEBUGGER BRANCH version is out.
« Reply #30 on: September 13, 2011, 06:37:21 am »
I create 2 patches for CB:

1. Add support Environment variable on the Debugger path settings on debugger branch.

2. Disable Pre/Post buld steps, when using external Makefile.
The  Pre/Post buld steps must be disable when using external Makefile because Makefile must be independent from CB environment, so all pre/post build steps must be inside in Makefile. Projects may be build on command line shell without CB environment...

All patches I testing on debugger branch svn 7444 on winxp sp3 and works fine.

Thanks. 

Offline Oleg_Sam

  • Advanced newcomer
  • *
  • Posts: 25
Re: The 15 August 2011 build (7386) DEBUGGER BRANCH version is out.
« Reply #31 on: September 13, 2011, 06:39:14 am »
Debugger path patch

Offline oBFusCATed

  • Developer
  • Lives here!
  • *****
  • Posts: 7787
Re: The 15 August 2011 build (7386) DEBUGGER BRANCH version is out.
« Reply #32 on: September 13, 2011, 09:32:37 am »
Non-debugger patches, should be made against trunk and discussed on the trunks topics.
Here we are interested only in the debugger.

BTW: Why would you want to disable the pre/post build steps? They could be useful if you're using a makefile you can't control. And if you need to build with the makefile only you could do so, no one stops you.
<debugger plugin maintainer>
(most of the time I ignore long posts)

Offline scarphin

  • Regular
  • ***
  • Posts: 425
Re: The 15 August 2011 build (7386) DEBUGGER BRANCH version is out.
« Reply #33 on: September 13, 2011, 09:20:06 pm »
I'm experiencing a crash on win7 sp1 with rev7387. That's how it happens:
1-Launch cb without any project loaded.
2-Double-click 'CodeBlocks.cbp' file on my local trunk folder.

There comes the crash. Is it specific to me or anyone else able to reproduce it?

Note: CB doesn't crash if I load cb by double-clicking on the project file.
« Last Edit: September 13, 2011, 09:21:52 pm by scarphin »

Offline scarphin

  • Regular
  • ***
  • Posts: 425
Re: The 15 August 2011 build (7386) DEBUGGER BRANCH version is out.
« Reply #34 on: September 13, 2011, 09:35:57 pm »
Report from 'codeblocks.rpt' file:
Code: [Select]
Error occured on Tuesday, September 13, 2011 at 22:20:15.

C:\GNU\CodeBlocks\codeblocks.exe caused an Access Violation at location 65eae3bd in module C:\GNU\CodeBlocks\share\codeblocks\plugins\codecompletion.dll Reading from location 00000000.

Registers:
eax=00000000 ebx=00000000 ecx=00000001 edx=00000000 esi=00000483 edi=0028e67c
eip=65eae3bd esp=0028e544 ebp=0028e56c iopl=0         nv up ei pl zr na po nc
cs=0023  ss=002b  ds=002b  es=002b  fs=0053  gs=002b             efl=00010246

Call stack:
65EAE3BD  C:\GNU\CodeBlocks\share\codeblocks\plugins\codecompletion.dll:65EAE3BD
6CCC7670  C:\GNU\CodeBlocks\wxmsw28u_gcc_cb.dll:6CCC7670  _ZN12wxEvtHandler21ProcessEventIfMatchesERK21wxEventTableEntryBasePS_R7wxEvent
6CCC77A9  C:\GNU\CodeBlocks\wxmsw28u_gcc_cb.dll:6CCC77A9  _ZN16wxEventHashTable11HandleEventER7wxEventP12wxEvtHandler
6CCC7B74  C:\GNU\CodeBlocks\wxmsw28u_gcc_cb.dll:6CCC7B74  _ZN12wxEvtHandler12ProcessEventER7wxEvent
6CDD70ED  C:\GNU\CodeBlocks\wxmsw28u_gcc_cb.dll:6CDD70ED  _ZN11wxTimerBase6NotifyEv
6CD07BDC  C:\GNU\CodeBlocks\wxmsw28u_gcc_cb.dll:6CD07BDC  _ZN7wxTimer4InitEv
74F362FA  C:\Windows\syswow64\USER32.dll:74F362FA  gapfnScSendMessage
74F36D3A  C:\Windows\syswow64\USER32.dll:74F36D3A  GetThreadDesktop
74F377C4  C:\Windows\syswow64\USER32.dll:74F377C4  CharPrevW
74F3788A  C:\Windows\syswow64\USER32.dll:74F3788A  DispatchMessageW
74F5C81F  C:\Windows\syswow64\USER32.dll:74F5C81F  IsDialogMessageW
6CCF59BC  C:\GNU\CodeBlocks\wxmsw28u_gcc_cb.dll:6CCF59BC  _ZN11wxEventLoop17PreProcessMessageEP6tagMSG
6CCF556D  C:\GNU\CodeBlocks\wxmsw28u_gcc_cb.dll:6CCF556D  _ZN11wxEventLoop14ProcessMessageEP6tagMSG
6CCF576B  C:\GNU\CodeBlocks\wxmsw28u_gcc_cb.dll:6CCF576B  _ZN11wxEventLoop8DispatchEv
6CCDDC1C  C:\GNU\CodeBlocks\wxmsw28u_gcc_cb.dll:6CCDDC1C  _ZN5wxApp5YieldEb
6CCCBC72  C:\GNU\CodeBlocks\wxmsw28u_gcc_cb.dll:6CCCBC72  _Z7wxYieldv
6CCC4B53  C:\GNU\CodeBlocks\wxmsw28u_gcc_cb.dll:6CCC4B53  _Z9wxExecuteRK8wxStringiP9wxProcess
6CCCA31B  C:\GNU\CodeBlocks\wxmsw28u_gcc_cb.dll:6CCCA31B  _ZN15wxMessageOutput3SetEPS_
65EC1408  C:\GNU\CodeBlocks\share\codeblocks\plugins\codecompletion.dll:65EC1408
65EBF38A  C:\GNU\CodeBlocks\share\codeblocks\plugins\codecompletion.dll:65EBF38A
65EC276B  C:\GNU\CodeBlocks\share\codeblocks\plugins\codecompletion.dll:65EC276B
65EC1DA8  C:\GNU\CodeBlocks\share\codeblocks\plugins\codecompletion.dll:65EC1DA8
65EAB622  C:\GNU\CodeBlocks\share\codeblocks\plugins\codecompletion.dll:65EAB622
65F09C84  C:\GNU\CodeBlocks\share\codeblocks\plugins\codecompletion.dll:65F09C84  _ZN12cbToolPlugin9BuildMenuEP9wxMenuBar
61880F24  C:\GNU\CodeBlocks\codeblocks.dll:61880F24  _ZN7Manager12ProcessEventER15CodeBlocksEvent
61897A18  C:\GNU\CodeBlocks\codeblocks.dll:61897A18  _ZN13PluginManager13NotifyPluginsER15CodeBlocksEvent
618B534A  C:\GNU\CodeBlocks\codeblocks.dll:618B534A  _ZN14ProjectManager10SetProjectEP9cbProjectb
618B8A93  C:\GNU\CodeBlocks\codeblocks.dll:618B8A93  _ZN14ProjectManager11LoadProjectERK8wxStringb
004461E6  C:\GNU\CodeBlocks\codeblocks.exe:004461E6
00445CB9  C:\GNU\CodeBlocks\codeblocks.exe:00445CB9
00445567  C:\GNU\CodeBlocks\codeblocks.exe:00445567
0040830F  C:\GNU\CodeBlocks\codeblocks.exe:0040830F
0040198D  C:\GNU\CodeBlocks\codeblocks.exe:0040198D
6CCB2025  C:\GNU\CodeBlocks\wxmsw28u_gcc_cb.dll:6CCB2025  _ZN11wxDDEClient16DeleteConnectionEPv
74F4E97A  C:\Windows\syswow64\USER32.dll:74F4E97A  DdeCreateStringHandleW
74F4E928  C:\Windows\syswow64\USER32.dll:74F4E928  DdeCreateStringHandleW
74F4FE0C  C:\Windows\syswow64\USER32.dll:74F4FE0C  DdeGetData
74F4FD92  C:\Windows\syswow64\USER32.dll:74F4FD92  DdeGetData
74F4FD62  C:\Windows\syswow64\USER32.dll:74F4FD62  DdeGetData
74F362FA  C:\Windows\syswow64\USER32.dll:74F362FA  gapfnScSendMessage
74F36D3A  C:\Windows\syswow64\USER32.dll:74F36D3A  GetThreadDesktop
74F377C4  C:\Windows\syswow64\USER32.dll:74F377C4  CharPrevW
74F3788A  C:\Windows\syswow64\USER32.dll:74F3788A  DispatchMessageW
74F5C81F  C:\Windows\syswow64\USER32.dll:74F5C81F  IsDialogMessageW
6CCF59BC  C:\GNU\CodeBlocks\wxmsw28u_gcc_cb.dll:6CCF59BC  _ZN11wxEventLoop17PreProcessMessageEP6tagMSG
6CCF556D  C:\GNU\CodeBlocks\wxmsw28u_gcc_cb.dll:6CCF556D  _ZN11wxEventLoop14ProcessMessageEP6tagMSG
6CCF576B  C:\GNU\CodeBlocks\wxmsw28u_gcc_cb.dll:6CCF576B  _ZN11wxEventLoop8DispatchEv
6CD8D518  C:\GNU\CodeBlocks\wxmsw28u_gcc_cb.dll:6CD8D518  _ZN17wxEventLoopManual3RunEv
6CD6BB19  C:\GNU\CodeBlocks\wxmsw28u_gcc_cb.dll:6CD6BB19  _ZN9wxAppBase8MainLoopEv
004058C4  C:\GNU\CodeBlocks\codeblocks.exe:004058C4
6CC73248  C:\GNU\CodeBlocks\wxmsw28u_gcc_cb.dll:6CC73248  _Z14wxUninitializev
6CCCD392  C:\GNU\CodeBlocks\wxmsw28u_gcc_cb.dll:6CCCD392  _Z7wxEntryP11HINSTANCE__S0_Pci
00401D71  C:\GNU\CodeBlocks\codeblocks.exe:00401D71
004823C6  C:\GNU\CodeBlocks\codeblocks.exe:004823C6
004010DB  C:\GNU\CodeBlocks\codeblocks.exe:004010DB
00401158  C:\GNU\CodeBlocks\codeblocks.exe:00401158
7503339A  C:\Windows\syswow64\kernel32.dll:7503339A  BaseThreadInitThunk
77439ED2  C:\Windows\SysWOW64\ntdll.dll:77439ED2  RtlInitializeExceptionChain
77439EA5  C:\Windows\SysWOW64\ntdll.dll:77439EA5  RtlInitializeExceptionChain

Offline oBFusCATed

  • Developer
  • Lives here!
  • *****
  • Posts: 7787
Re: The 15 August 2011 build (7386) DEBUGGER BRANCH version is out.
« Reply #35 on: September 13, 2011, 10:32:42 pm »
Does it happen with the trunk version of the nightly?
What happens if you disabled the CC?
<debugger plugin maintainer>
(most of the time I ignore long posts)

Offline scarphin

  • Regular
  • ***
  • Posts: 425
Re: The 15 August 2011 build (7386) DEBUGGER BRANCH version is out.
« Reply #36 on: September 14, 2011, 01:02:41 am »
Trunk also crashes. When I disable code-completion, both (trunk and debuggers branch) work. I can then enable cc plugin. Can u reproduce the crash or is it just me?

Offline oBFusCATed

  • Developer
  • Lives here!
  • *****
  • Posts: 7787
Re: The 15 August 2011 build (7386) DEBUGGER BRANCH version is out.
« Reply #37 on: September 14, 2011, 01:11:14 am »
CC is known to be crashy in this revision, disable it and wait for the next nightly...
<debugger plugin maintainer>
(most of the time I ignore long posts)

Offline Manolo

  • Advanced newcomer
  • *
  • Posts: 38
Re: The 15 August 2011 build (7386) DEBUGGER BRANCH version is out.
« Reply #38 on: September 14, 2011, 10:43:53 am »
Hello
I use MinGw/GDB. I'm getting this issue while debugging:
When the GDB just execute this line of code:
  wxString *psty = (wxString*) NULL;
I put the mouse over 'psty' and a SIGSEGV is prompt:

Code: [Select]
> whatis psty
type = wxString *
>>>>>>cb_gdb:
> output psty
(wxString *) 0x0>>>>>>cb_gdb:
> output /c psty->wx_str()[0]@psty->length()
The program being debugged was signaled while in a function called from GDB.
GDB has restored the context to what it was before the call.
To change this behavior use "set unwindonsignal off".
Evaluation of the expression containing the function
(wxString::wx_str() const) will be abandoned.
Program received signal SIGSEGV, Segmentation fault.
0x6fc611d6 in libstdc++-6!_ZNKSbIwSt11char_traitsIwESaIwEE5c_strEv () from c:\mingw\bin\libstdc++-6.dll
>>>>>>cb_gdb:

I'm also using the script I posted some days ago:
http://forums.codeblocks.org/index.php/topic,15050.msg101768.html#msg101768

Of course, the problems arise because psty is not pointing to a valid wxString object.
What I ask here and myself is: How can C:B be aware of this before trying to output this invalid var?

Regards

Offline oBFusCATed

  • Developer
  • Lives here!
  • *****
  • Posts: 7787
Re: The 15 August 2011 build (7386) DEBUGGER BRANCH version is out.
« Reply #39 on: September 14, 2011, 10:48:39 am »
I guess we can check for NULL in before evaluating some scripts...
But I suggest to use python enabled gdb + pretty printers for wxwidgets (there are available for 2.9+). This is the better approach.
<debugger plugin maintainer>
(most of the time I ignore long posts)

Offline ollydbg

  • Developer
  • Lives here!
  • *****
  • Posts: 4184
  • Interests on OpenCV and Robotics
    • Chinese OpenCV forum moderator
Re: The 15 August 2011 build (7386) DEBUGGER BRANCH version is out.
« Reply #40 on: September 14, 2011, 11:16:22 am »
Hello
I use MinGw/GDB. I'm getting this issue while debugging:
When the GDB just execute this line of code:
  wxString *psty = (wxString*) NULL;
I put the mouse over 'psty' and a SIGSEGV is prompt:

Code: [Select]
> whatis psty
type = wxString *
>>>>>>cb_gdb:
> output psty
(wxString *) 0x0>>>>>>cb_gdb:
> output /c psty->wx_str()[0]@psty->length()
The program being debugged was signaled while in a function called from GDB.
GDB has restored the context to what it was before the call.
To change this behavior use "set unwindonsignal off".
Evaluation of the expression containing the function
(wxString::wx_str() const) will be abandoned.
Program received signal SIGSEGV, Segmentation fault.
0x6fc611d6 in libstdc++-6!_ZNKSbIwSt11char_traitsIwESaIwEE5c_strEv () from c:\mingw\bin\libstdc++-6.dll
>>>>>>cb_gdb:

I'm also using the script I posted some days ago:
http://forums.codeblocks.org/index.php/topic,15050.msg101768.html#msg101768

Of course, the problems arise because psty is not pointing to a valid wxString object.
What I ask here and myself is: How can C:B be aware of this before trying to output this invalid var?

Regards


I just test on gdb (the gdb build myself with python support, you can find download links in this forum), the log is below, it seems there is no crash.
Code: [Select]
> p psty
$1 = (wxString *) 0x0
>>>>>>cb_gdb:
> whatis psty
type = wxString *
>>>>>>cb_gdb:
> output /c psty->m_pchData[0]@((wxStringData*)psty->m_pchData - 1)->nDataLength
Cannot access memory at address 0x0
>>>>>>cb_gdb:
If some piece of memory should be reused, turn them to variables (or const variables).
If some piece of operations should be reused, turn them to functions.
If they happened together, then turn them to classes.

Offline oBFusCATed

  • Developer
  • Lives here!
  • *****
  • Posts: 7787
Re: The 15 August 2011 build (7386) DEBUGGER BRANCH version is out.
« Reply #41 on: September 14, 2011, 12:05:20 pm »
The difference probably is that you're accessing a member, but Manalo is calling a function...
<debugger plugin maintainer>
(most of the time I ignore long posts)