Hello, On Fri, Jun 19, 2009 at 01:18:19PM +0200, Thomas Schwinge wrote: > On Thu, Jun 18, 2009 at 11:10:19PM +0300, Sergiu Ivanov wrote: > > All .c files in unionfs define _GNU_SOURCE explicitly (though, true, > > they do only #define _GNU_SOURCE /*(no 1)*/, and I've changed my new > > patch accordingly). I chose to do similarly in unionmount.c, because > > moving the definition of _GNU_SOURCE into Makefile (if I understand > > you correctly) will mean changing a lot stuff loosely connected to > > unionmount matters. I guess this transition has its place in a > > separate clean-up patch. > > Absolutely right, and sorry that I didn't check the other unionfs files > before complaining -- that's one major problem with reviewing patches out > of their context. > > Yes, that's totally a separate change then, and you are correct to first > do it like it's done in the other unionfs files. So, this change is > first to be done in the unionfs repository, and then that commit is to be > merged into your unionmount branch, at the same time extending the merge > to cover the unionmount.c file (needing the exactly same change) that is > present in your branch, but not in the master one. I think this is > exactly how we're doing to deal with this issue -- sometime later, as > this is not really that important right now. And then you'll see how > non-trivial (but not too difficult either) merging is done.
Yay! :-) I've said something correct :-) Turning back to seriousness: I'm leaving the issue about #define _GNU_SOURCE for the future, just according to your words. Also, I think I would feel excited to see how non-trivial merging is done :-) Regards, scolobb
