Yeah, sorry, but those are not particularly reasonable arguments. A code cleanup is by definition different from regular commits that concern functionality. It is very simple to distinguish between the two and it is particularly simple to distinguish this in a git-blame (*cough* git-blame -w *cough*). And the answer to the question "why is it like this" in case of code style commits is rather simple. Again - We are programmers, we know how to filter. Don't pretend you cannot filter, it's simply not a useful argument.
I think that the only real concern here is that history bloat is cosmetically unappealing and as I've said - I'm willing to get that down to a handful of commits per PR going forward (save, obviously, for commits related to comments by you guys, but we could find a way to combine those into one). I'm really more than happy to make this work, but I'm asking you to pay a small price for this in respecting my own commit and work history the way it actually went. Throwing up technical side-issues as arguments is, I suppose, the way we developers usually try to handle everything, but let's be clear about this: I'm telling you that it is my own, personal, requirement to do it in this way. I have already put in a lot of work in this and I'm offering you to do a tremendous amount more work. I hope you aren't so foolish as to throw that opportunity away. My work would be, in time, only a blip on the commit history. Bikeshedding about whether it would be a blip of one or two microseconds seems almost comical. Anyways, let me know if that is a final decision. The only thing I can do here is appeal to reason and ask to respect my one requirement. If that doesn't work for you, I'd be happy to pull my PRs and safe myself further work and you guys further trouble of dealing with this. On Wed, Sep 18, 2013 at 1:18 PM, Thomas Bruederli <[email protected]>wrote: > David Deutsch wrote: > > So you're saying a clean commit history is more important than giving me > > proper credit where it is due? Again, that is not acceptable. Besides - > why > > is the commit history that important to begin with? > > I have to agree with Alec here. Commit history is important for us and so > are proper commit messages. I want to be able to find out who wrote a > particular line of code and why it is like it is. That includes tracking > changes (preferably with ticket numbers in the commit message) to find the > reason why I better do not change that or what circumstances are to > consider when changing something. That's the reason why I initially wasn't > really keen on code refactoring because it doesn't change the functionality > but bloats the history. > > Of course the arguments for cleaning up the code are strong enough to > sacrifice that history let's still try to minimize it. Preferably we want > one single commit per PR that says "Code cleanup by David Deutsch" or > something. I know that there'll be more but what we currently have in the > pending PRs of yours is rather messy and not helpful. I admit that's > primarily because of all the discussions we had to reach the agreement > which we now have and I also understand that other PRs will be way shorter. > > It's definitely not about not wanting to give you the credits you deserve > for all your hard work, please don't get us wrong on this. If you have > other propositions how we can credit your work, please let us know. > > But still I kindly request you to re-create the pull requests with one > commit per processing step (as proposed in [1]) and with descriptive commit > messages. > > Kind regards, > Thomas > > [1] https://gist.github.com/daviddeutsch/6376013 > _______________________________________________ > Roundcube Development discussion mailing list > [email protected] > http://lists.roundcube.net/mailman/listinfo/dev >
_______________________________________________ Roundcube Development discussion mailing list [email protected] http://lists.roundcube.net/mailman/listinfo/dev
