On Mon, Nov 23, 2009 at 06:05:47PM +0100, Petr Baudis wrote: > > You wouldn't need to do anything that drastic. You would just need to > > pass "--progress --all-progress" instead of only --all-progress. But you > > have provided the data point that such a change would break at least one > > user. > > > > We could also leave --all-progress as-is and add new option to mean "if > > you are already doing progress, do all progress". > > Hmm, maybe I'm confused - I just call > > git remote update > > and don't pass any progress switches - would your change still affect > me? Can I pass --progress to `git remote update`? Oh, I misunderstood; I thought you were calling pack-objects directly. So you are actually relying on the "even though isatty(2) is not true, we always print progress messages" behavior? I think that behavior is buggy. It hurts everybody pushing via cron, and it violates the usual rule for when we show progress messages. I don't think you can get a --progress pushed all the way down to the pack-objects in this case; we would need to add code to override the isatty check. That being said, your example of "remote update" means you are dealing with fetch, and we are not touching the fetch code path at all. -Peff -- To unsubscribe from this list: send the line "unsubscribe git" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html