Re: [Pnet-developers] [patch #3599] ildasm offsets

[James Michael DuPont]

--- Russell Stuart <address@hidden> wrote:

> On Thu, 2004-12-30 at 07:48, Rhys Weatherley wrote:
> > Debug code should generally be removed after the bug has been
> located.  It 
> > keeps the system cleaner and easier to understand.  A few modes,
> like CVM 
> > debugging, are present because we have to use them so often.   But
> they are 
> > the exception, not the rule.
> 
> There is a way around this: most projects (pnet included,
> from what I have seen), welcome unit test code.  If you
> find a bug, add a test the demonstrates it, and later
> proves it doesn't exist.

And most project also welcome test cases and bug reports.
even if they dont have fixes.

All my bug reports are reproducable, all have a simple test case
attached. 


My patches are not meant to be checked in, but can be used by other
people if needed. I would like to keep them open until I get to making
them a full blown option or new tool.

I have no problem with maintainig debugging versions of ildasm, ilrun ,
ilverify and ilalink. They could contain more debugging information for
people who USE them. If you like we can make some type of ifdef for
them. Also we can put the code int a separate file so it wont disturb.


> There are not too many people who will object to you
> modifying the ilasm code to make it more testable,
> providing it is not too intrusive.  Code that is easy
> to test is generally easier to debug as well.

Adding in the offset to the IL makes it reable. I am just thinking
about   objdump output, that is basically what i want.

Maybe i should make an objdump clone for pnet?

mike

=====
James Michael DuPont
http://introspector.sourceforge.net/


                
__________________________________ 
Do you Yahoo!? 
Yahoo! Mail - Find what you need with new enhanced search.
http://info.mail.yahoo.com/mail_250