[Mono-devel-list] mono debugger in infinite loop

[Martin Baulig]

Philip Nelson <panmanphil at yahoo.com> writes:

> I spent a good deal of time upgrading my suse 8.0 system to the gnome 2.2.1
> libaries so I could try out the new mono debugger.  Got it compiled and
> executing with the tarball debugger, the cvs version of gtk-sharp and mono .23.
>  My real purpose in all this is to solve more of the problems with NAnt.exe
> running under linux, which I have partially working now.

Hi,

first of all, thanks a lot for trying out the debugger, I really appreciate it a lot if
people try it out and give me some feedback and bug reports :-)

> So, with great anticipation ...
> 
> mono /usr/bin/Debugger.exe NAnt.exe -help

Did you compile NAnt with debugging support ?

I compiled mine with

        $ cd ~/monocvs/mcs/nant/
        $ mcs -g /out:NAnt.exe /recurse:*.cs

> blew up somewhere deep in gtk-sharp (?) with a null reference exception

Can you please post the full backtrace ?

> Ok. mono /usr/bin/Debugger.exe
> load program and viola, the debugger shows a yellow highlight on the first line
> of source.

Ok, I suppose that's also NAnt.exe and it stopped in NAnt.cs, line 98 (on the
`if (!project.Run ())` statement), that's what it does for me.

> Step, Step, and then step into the first reference to a system
> type, AppDomain in this case, never to return.  I could break, but have to
> close the session after that.  

Which method in AppDomain do you mean and did you compile your corlib with debugging
support ?

> Ok, try the same thing from the command line.  I can step and again it goes
> into a loop of some sort when hitting the AppDomain reference.
> 
> OK, try something simpler, a 10 line program that loads up a DataSet, builds a
> DataTable in memory and calls WriteXml() to save.  This was a simple test that
> already worked (except for writexml not being implemented).  In this case I
> could step a few steps until I called a method on the dataset and then an
> infinite loop.

Please send me this small test program so I can have a look at it myself.  Which method
did you attempt to call where it crashed ?

> Any suggestions?  The only other possible oddity with my system is that I have
> applied the low latency patches to my kernel, other than that it should be
> pretty stock suse 8.0

I don't think it's your system which is broken, you just found some bugs in the debugger
which I need to fix.

-- 
Martin Baulig
martin at gnome.org
martin at ximian.com