On Wed, Sep 13, 2006 at 10:24:04PM +1000, Hoan Ton-That wrote: > I wrote this feature because I was working on a repository > with configuration files in it. I was always being pestered > about the changes I made to them, but I never wanted to > record them, because they were local changes (like a STMP > server address). > > Boring files didn't do the job, as they only stop certain files > from being added. Once they have been added, darcs still > asks for their changes. > > You might be working on darcs (as I was) and have modified > the makefile so it still compiles despite warnings. Here is an > example of the new commands in action:
I usually record "local" changes in patches with special and visual (in caps) names. I have never found it tiresome to answer 'N' to these patches while pulling or pushing. Note that I don't have to read the patch names and bother with if I want to pull them or not, since the names stand out. This is much easier on the brain then seeing just another change in the record dialogue, and have to think about if you want to record it or not, and it's only one patch instead of possibly several changes. There is a wishlist item for a way to mark patches as ignored / local, so darcs never asks about them when pulling or pushing. This will make this approach even easier, and more important, things like pull --all will be possible. (There are however possible problems with dependent patches with this solution.) What I'm a bit worried about with your new proposed ignore command is that it (to me) doesn't look like much value for the money (or rather, feature for the extra complexity). And it doesn't solve the "real" problem either, does it? What if I want to put a local #ifdef / #endif in a file while making changes to it, and don't want to be pestered with questions about them on every new record of real changes to that file? And even if ignored files are useful to an extent where it's worth it, do we really need two special commands to manage the ignore file? We don't have the commands boring and unboring. I'm sorry to sound so negative. I do want to encourage any sort of development on darcs, and one of the best motivating drives for people to do development, is to fix things that they want to have working. And as David usually says: the one who writes it (almost always) gets to decide. But I think the solution to this problem can be improved a great deal, especially if it could be more generally usable, and I hope you agree with me. There might be other good ideas how the ignore thing can be done, and I think we should have a thorough discussion before settling on this (just to make darcs the absolutely greatest tool there is :-). -- Tommy Pettersson <[EMAIL PROTECTED]> _______________________________________________ darcs-users mailing list [email protected] http://www.abridgegame.org/mailman/listinfo/darcs-users
