[Lazarus] Loading old project into Lazarus_fixes causes lock-up...

[Martin]

On 29/11/2010 11:18, zeljko wrote:
> On Monday 29 November 2010 10:48, Graeme Geldenhuys wrote:
>> * Debugging is seriously broken in 0.9.29. Watch windows don't update,
>> Local Vars window shows old values, breakpoints don't trigger, break point
>> hit count doesn't work. This all causes major headaches while debugging -
>> hence I don't use the IDE for debugging at all any more. I waisted way to
>> many hours hunting down a bug, just to find out that the IDE is actually
>> the cause of the "false information".
> Martin is working on this afaik.
Have you tested, since I added the comments to your issue?

Yes the string (0 vs 1 based) issue still exist. (but it's not a 0.9.29 
issue, it's been there forever. Yes it desperately needs to be fixed, 
but it isn't that easy to be done. It will be fixed as either gdb learns 
to do it, or any alternative (e.g a pascal driven debugger (which afaik 
a project exists)) will be ready.

I understand that to the end user it makes no different if the problem 
is Lazarus gdb, or fpc => it's all the same a problem.

If you find other issue with the debugger, then please let us know.

And yes, debugging was worse in 0.9.29 than in 0.9.28 => it is a 
snapshot, it is not tested, it is all explained and all warned about. 
See my other mail.


>> * Closing a tab via X button, closes all other tabs except the one I wanted
>> closed. Major frustration!
> Cannot reproduce (Fedora 3 + gtk2-2.12, Fedora 13 + gtk2-2.20)

Yes the bug is well known.

Currently only a partial workaround exists (which is by no way a fix of 
the issue, not even close):
- you can disable "ctrl" for the tab-closing only. Navigation still 
works, but tabs-closing will ignore ctrl.

And yes, I'd too rather would have a fix for the issue, than just a 
cheap and partial workaround. But a full fix will take more time.

Martin