Mail Archives: djgpp/2001/01/18/20:52:06
Eli Zaretskii wrote:
>
> On Thu, 18 Jan 2001, Cesar Rabak wrote:
>
> > If we make the modifications, in order to submit the code and binaries
> > modified, is there any kind of regression tests?
>
> If you tried the testing script I put into the source distribution,
> you could run it before and after the change. It doesn't run the
> entire test suite yet, but it's better than nothing.
Sure!
>
> But most important, simply use the debugger for some time in serious
> work, and if it doesn't show any signs of problems, the changes are
> probably right. I have write access to the GDB CVS tree, so I can
> commit changes specific to DJGPP if I think they are okay.
Will do this, of course! We still have to find the bug in our simulation
software ;-(
>
> But I don't understand the ``and binaries'' part above. Surely, you
> meant to submit source-level diffs, no?
I thought it would be a good idea to have available an already compiled
gdb and its library available as soon the diffs get validated.
I will submit the diffs, and I'm surmising that automatically leads to
an update of the packages after a while is it correct?
>
> > I'm asking this because the changes seem to be small. May I stumble in a
> > wall. . .
>
> They are small, but they touch a very sensitive place in debug
> support: exceptions, memory protection, etc. Caveat emptor...
I noticed that... will try and share the results here (even if not
successful, may help others)!
- Raw text -