Debuger error when closing the application.
lazarus / 12-Debugger

----------------------------------------------------------------------
taazz · 2012-09-06 15:20

I have encounter an error when I close the debugging application with
Lazarus 1.0 r38415 FPC 2.6.0 i386-win32-win32/win64 on win7 32bit.

Pressing the stop button I'm able to restart it as needed and I
usually use ctrl+f2 to stop the application so its not a big problem
for me but I would like to make sure that the team knows about it so I
can add a bug report if that is required.

see attached image for the info.

----------------------------------------------------------------------
Martin_fr · 2012-09-06 15:26

Looks like http://sourceware.org/bugzilla/show_bug.cgi?id=14018

This can happen as son as threads are involved. (many dialogs start
threads ((Topen/TSaveDialog)

You can downgrade to Gdb 7.2

use SVN
http://svn.freepascal.org/svn/lazarus/binaries/i386-win32/gdb/bin/
or from the mingw site

In most aspects 7.2 should have all features.  I am not sure (int may
have, but I do not know) if you debug dll with starter app.

----------------------------------------------------------------------
taazz · 2012-09-06 15:39

Well based on the link you have provided adding a break point should
start a new thread the funny thing is I do not get any problems with
my break points so far and the test app that produced the above error
had only a form with a constructor, destructor and a dummy method that
does nothing than raising an exception. So, no threads in the
application or break points are required to see the problem.

Just add in some extra info. I'll try to downgrade to 7.2 in the
weekend probably.

thank you for the info.

----------------------------------------------------------------------
Martin_fr · 2012-09-06 15:55

Yes adding a breakpoint may add a thread.

Actually interrupting/pause the app daes (and that may be needed to
add a breakpoint in a running app)

However this is a timing issue. Most times all is well when using
threads. But if the timing is just not right...

Then again, gdb could have several issues leading to similar output