GDB

[marfig]

Can someone please inform me if there is any work being done in the debugger pluggin? After the third thread it starts to become frustrating not getting a simple yes or no. Not even a reply telling me to go away. Maybe I've been posting on the wrong forums, I can accept that. But I don't think my question is particularly offensive or rude, is it?

I have a few suggestions for it, namely on what it comes to conditional brakpoints in GDB since the current pluggin is not dealing with them correctly. Unfortunately I can't do it. I'm not versed in C++ well enough to implement the changes myself.

[killerbot]

The debugger plug-in is as far as I know created and maintained by our leader himself ;-) Mandrav.
I think there's a reason why you had no replies on that topic. Yiannis (Mandrav) was away on holidays and returned this weekend. It's obvious he has a whole lot of catching up to do, mails, pm's , forum ...
You can always send him a PM with your ideas. No idea when he will have time for it (that catching up you know). I am also interested in this topic, but I don't know the debugger code well, and I think it's best I stay out of it and leave it to Yiannis.

[mandrav]

I have a few suggestions for it, namely on what it comes to conditional brakpoints in GDB since the current pluggin is not dealing with them correctly.

Can you please explain what the problem is?

[marfig]

Thank you for the information killerbot

Can you please explain what the problem is?

When you have conditional breakpoints set that can't be evaluated (when condition is based on an inlined function), gdb outputs an error message informing the condition couldn't be evaluated when it hits the breakpoint. It then breaks there as if the condition had been evaluated to true.

But codeblocks behaves differently. It doesn't mirror this error message and doesn't break. From the user standpoint, the breakpoint was ignored. It would be very useful if the gdb output message could be mirrored by the pluggin, even if the current behavior of not breaking remains. Currently there is no way of telling (other than doing a manual debug) why the condition failed to work as expected from within codeblocks.

A simple class with an inlined function gives a good test platform to retrieve the error message. Regardless, here it is as output by gdb:

Error in testing breakpoint condition:                                                                 
Cannot evaluate function -- may be inlined

[mandrav]

OK, I found the time to look into this.
I have setup a very simple test case for this and I can confirm that the error message is not relayed back to the user. But the breakpoint breaks just fine.
So, unless you can provide me with a test case that doesn't break, all I can do is add the code to relay the error message to the user.

[marfig]

I tried, but I'm not being able to replicate the non-breaking situation I experienced before. I apologize.

Thank you for taking the time to solve this. I reckon seeing the message from gdb will already be a good help.