Re: [PATCH 4/4] remote: new show output style
- From
Jay Soffian <jaysoffian@gmail.com>
- Date
- Feb 19, 2009, 17:57 UTC
- Message-ID
- <76718490902190957td40d425p6b2ae1242060114d@mail.gmail.com>
- In-Reply-To
- <499D82BC.7020500@xiplink.com>
On Thu, Feb 19, 2009 at 11:03 AM, Marc Branchaud <marcnarc@xiplink.com> wrote:
> First, a nit: I don't know if the "w/remote" notation makes sense to > non-English speakers. I also like the alignment achieved by "merges with > remote " (note the trailing space).
I tried out a few different option and none was very satisfactory to me.
1)
master rebases onto remote master
another-branch merges with remote next
some-other-branch rebases onto remote masterHere, the unaligned "with" and "onto" is ugly.
2)
master rebases onto remote master
another-branch merges with remote next
some-other-branch rebases onto remote masterThis looks better to me. However, if none of your branches rebase, then the extra space looks like it is a mistake. e.g.:
master merges with remote master
another-branch merges with remote next
some-other-branch merges with remote masterI could add code to detect whether all the branches merge and then not output the space, but, sigh. And I couldn't think of any other combination of words that had the same character spacing.
So that's how I ended up with "merges w/remote". This is also slightly less wide. I always try to have output fit into 80 columns (how quaint, I know) and a merging branch might have multiple upstreams. e.g.
another-branch merges w/foo, bar, baz
IOW, the output in the patch wasn't arbitrary. I did think about it quite a bit. Which isn't to say it's right, just it's the best I came up with.
I'm somewhat confused by "w/remote" making sense to non-English speakers as it's English output.
> Second, I think it would be good to also change the format of the 'git push' > list, for consistency:
I left that out on purpose. The only folks with push refspecs put them their manually, and the raw refspec is clearer and more concise than any English words can convey. That was my reasoning anyway.
Thanks for the feedback.
j.