> Having waited for months for this patch to be merged in, it is very > disappointing > to say the least. Wish it had been merged and if changes are needed they can > always be > made subsequently. That has been my understanding of the development model.
If you really want to get into it... I'll certainly accept some of the blame for taking too long to review this patch. However, you didn't do yourself any favors by: a) making one huge ugly patch and b) being rather disagreeable when someone actually tried to review it. As far as the development model goes, it is certainly true that for new things, we can merge first and fix later. But when we're touching something like IPoIB, which is pretty critical to just about everyone using the IB stack at all, the standard is a little different: we need to be much more conservative. And even for new stuff, starting from a good base is pretty important; it's easy to pick on coding style problems, and indeed they do make review harder, but it's even more important to have the underlying logic and structure be simple and maintainable. Anyway, I'll post my current patch series shortly. I think I was able to make the patch quite a bit neater and more reviewable: your patch added > 400 lines, while the main part of my series adds < 200 lines. - R. _______________________________________________ general mailing list [email protected] http://lists.openfabrics.org/cgi-bin/mailman/listinfo/general To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
