On 24/04/14 19:35, Alexander Shiyan wrote: >> The right way to do it is, as I wrote above, by gradually changing the >> old driver with a patch series. And my question is, why not do it that >> way? Then it would be possible to review the patches one by one, seeing >> what has changed. > > "gradually changing"... > I repeat that this is not an old modified driver, but written new. Yes, I understand that. Again, my question is, why didn't you modify the old driver? That's how things should normally be done. Instead, you made a totally new one, making proper review against the old driver impossible. > if you imagine a new file as a diff to the old, this can be clearly seen. > > There is no reason to waste time on a series of changes since I > can not even check these changes on real hardware, but only in the > last stage when the driver will be the current version. Hmm what? So is the old driver totally broken, and cannot be used at the moment? Or why you can't test on real hardware? Note that I don't know anything about the fb hardware in question, nor the driver. Maybe there's a valid reason to write a new driver from scratch. But there very rarely is. And "because I already wrote a new driver, and it's a waste of time for me to throw away my work and patch the old one", is not a very good reason. Tomi
Attachment:
signature.asc
Description: OpenPGP digital signature
