On Tuesday 09 March 2010, Antonio Borneo wrote:
> I'm not sure I fully catch your points.
> Anyway, I agree that code got even more confused and that comments are
> not consistent with code anymore.
>
> Why not changing the whole procedure in a cleaner way?
> I've put in attachment my proposal
Could you reduce the gratuitous changes which just hide
the content of your patch? Like shuffling comments, Such
things just make patches hard to review.
> Execution is more linear and without goto statement.
YOu talk as if a well structured "goto" is a problem in itself.
That's not true ... but if that bothers you, all you had to
do was turn that into a "do { ... } while(updated)" loop.
(And note that the reason for that additional loop was to handle
changes affecting the core loop criteria ... a quick read
didn't show you still addressing those issues.)
> Moreover, it splits the first-last interval and should improve performance.
I don't follow what you mean by this. Do you mean it addresses
the REVISIT comment by issuing multiple driver requests if it works
out that there are discontiguous regions? That's not necessarily
going to be faster... and such a change would have been more clearly
done as a separate patch.
- Dave
_______________________________________________
Openocd-development mailing list
[email protected]
https://lists.berlios.de/mailman/listinfo/openocd-development
