This is the mail archive of the
mailing list for the frysk project.
Re: debugging an existing process
- From: Adam Jocksch <ajocksch at redhat dot com>
- To: Wu Zhou <woodzltc at cn dot ibm dot com>
- Cc: frysk at sources dot redhat dot com
- Date: Mon, 31 Jul 2006 10:46:02 -0400
- Subject: Re: debugging an existing process
- References: <email@example.com>
Wu Zhou wrote:
I see that debugging existing process can be used now. So I give it
some try. I am very happy to see that register/memory window work ok.
The source window can also be opened in some situation. But I also
have some comments/questions for this:I agree, making a distinction between these two cases is an excellent
idea. However we have no way of distinguishing between these cases at
the moment. As soon as we can obtain a full stack backtrace we will be
able to tell if there is a stack frame with debug information or not.
1. If the process has no debug information, you will see nothing in
the source window; if the process is executing into some library
function, you will also see nothing in the source window. So is it
better for us to have some mechanism to differentiate these two
My suggestion is like this: if the current stack frame has no
debuginfo, but the current process does have, we can show some message
like this: "no debug information for this stack, but we do have debug
information for this process, would you like to see the source for the
This can give the user more choice to see the source code.
2. When the current stack frame has no debug information available,
the current frysk view will display "unknown file: unknown function".
In my opinion, it might be better to display the instruction address
of the current stack.Something to consider. I'll look into this
3. In the register view, if we can copy the value of some or all the
registers' value, that will be helpful.
Again, something to consider.
4. The menu items in program tab are all disabled. Did this mean we
can't control the process at this time?