On 14/02/2014 08:33, Mattias Gaertner wrote:
On Thu, 13 Feb 2014 19:05:57 -0800 (PST)
leledumbo <leledumbo_c...@yahoo.co.id> wrote:
Reading symbols from /media/LinuxShared/lazarus/lazarus...done.
(gdb) b FPC_RAISEEXCEPTION
Breakpoint 1 at 0x8072f25
(gdb) run
Starting program: /media/LinuxShared/lazarus/lazarus
Fontconfig warning: "/etc/fonts/conf.d/50-user.conf", line 14: reading
configurations from ~/.fonts.conf is deprecated.
TMainIDE.ParseCmdLineOptions:
PrimaryConfigPath="/home/leledumbo/.lazarus"
SecondaryConfigPath="/etc/lazarus"
TSourceEditorSharedValues.GetMainLinkScanner not pascal highlighted:
unit1.pas
[TEnvironmentOptions.Save] error writing
"/home/leledumbo/.lazarus/environmentoptions.xml": EDOMError in
DOMDocument.CreateElement PackageEditor_
NiceChartD7 0
Strange. It is raising a FPC exception. gdb should break here, but it
does not.
Martin, do you have an idea how this is possible?
Not sure, maybe (Marc pointed that out ot me once) it is the same, why
the IDE does take the address of RAISE_Exception, and set the break at
the address.
GDB looks at the assembler of the function and makes a decision where
the prologue ends. With FPC omitting the stackframe it may get that wrong.
Though I have used the
b FPC_RAISEEXCEPTION
myself, and it had worked, but it may depend on versions, etc.
Try
info address
b *0x8072f25
--------------
here is output from gdb I just tried
(gdb) b FPC_RAISEEXCEPTION
Breakpoint 1 at 0x40da33: file ../inc/except.inc, line 195.
(gdb) info address FPC_RAISEEXCEPTION
Symbol "FPC_RAISEEXCEPTION" is at 0x40da20 in a file compiled without debugging.
(gdb)
info address is 19 bytes earlier
In my case that is correct (looking at disaasemble), but my FPC was build with
-O1
--
_______________________________________________
Lazarus mailing list
Lazarus@lists.lazarus.freepascal.org
http://lists.lazarus.freepascal.org/mailman/listinfo/lazarus
