> What do you think? Should I bring this up on the Git mailing list? Or have I missed something/misunderstood something?
It does sound worth including in a mailing list discussion, even it it's only adding a short clarifying sentence in the man page or its syntax summary (synopsis) Draft patches also help discussions. The mailing list does not require a subscription, but does require 'plain' text (absolutely no HTML) On Thursday, March 23, 2023 at 9:08:09 PM UTC kristoffer...@gmail.com wrote: > I was using `git range-diff`: > > $ git range-diff main feature-v2 HEAD > > And got an error message: > > fatal: need two commit ranges > > I intended to give three arguments to the command. I started to wonder > if I maybe had just typed `main feature-v2` (forgot the third > argument). No, it was three arguments. > > The problem was that I had mistyped one argument; it was `feature-v3`, > not `feature-v2`. > > I thought that this error message was a little misleading or too > generic. I looked at the history of the file (`builtin/range-diff.c`) > and saw this commit: b75747829f (range-diff: optionally accept > pathspecs, 2022-08-26). > > Maybe it works better with a `pathspec` marker (`--`)? > > $ git range-diff HEAD main feature-v2 -- > fatal: not a revision: 'feature-v2' > > usage: git range-diff [<options>] <old-base>..<old-tip> > <new-base>..<new-tip> > or: git range-diff [<options>] <old-tip>...<new-tip> > or: git range-diff [<options>] <base> <old-tip> <new-tip> > > Yes it does: `not a revision: 'feature-v2'`. > > So I tried to contrast the behavior on the current release with the > behavior on the release before the aforementioned commit. > > $ git checkout v2.40.0 > $ make clean > $ NO_CURL=true make -j 4 > $ # Misspelled ref `seen` as `seent` > $ ./bin-wrappers/git range-diff master next seent > fatal: need two commit ranges > > usage: git range-diff [<options>] <old-base>..<old-tip> > <new-base>..<new-tip> > or: git range-diff [<options>] <old-tip>...<new-tip> > or: git range-diff [<options>] <base> <old-tip> <new-tip> > […] > > Expected behavior: tell me that `seent` is not a revision. > > Actual behavior: generic error message. > > But I get a nice error message if I append `--`: > > $ ./bin-wrappers/git range-diff master next seent -- > fatal: not a revision: 'seent' > > usage: git range-diff [<options>] <old-base>..<old-tip> > <new-base>..<new-tip> > or: git range-diff [<options>] <old-tip>...<new-tip> > or: git range-diff [<options>] <base> <old-tip> <new-tip> > > Contrast the behavior on `v2.37.2`, which is one of the tags before > rev. b75747829f (range-diff: optionally accept pathspecs, > 2022-08-26).[1] > > $ git checkout v2.37.2 > $ make clean > $ NO_CURL=true make -j 4 > $ ./bin-wrappers/git range-diff master next seent > fatal: not a revision: 'seent' > > usage: git range-diff [<options>] <old-base>..<old-tip> > <new-base>..<new-tip> > or: git range-diff [<options>] <old-tip>...<new-tip> > or: git range-diff [<options>] <base> <old-tip> <new-tip> > […] > > $ # Tag before rev. b75747829f (range-diff: optionally > $ # accept pathspecs, 2022-08-26) > $ git checkout v2.37.2 > $ ./bin-wrappers/git range-diff HEAD master seent > fatal: ambiguous argument 'HEAD..seent': unknown revision or path not > in the working tree. > Use '--' to separate paths from revisions, like this: > 'git <command> [<revision>...] -- [<file>...]' > error: could not parse log for 'HEAD..seent' > > [1]: I figured out later that `v2.37.6` looks like the last release > which does not have this commit > > This error message is exactly what I expect to see when I mistype a ref > (my git(1) conditioning). > > But the `v2.40.0` error message with `--` is better: `fatal: not a > revision: 'seent'` > > So what if I give a `--`: > > $ ./bin-wrappers/git range-diff master next seent -- > fatal: ambiguous argument 'master..seent': unknown revision or path > not in the working tree. > Use '--' to separate paths from revisions, like this: > 'git <command> [<revision>...] -- [<file>...]' > error: could not parse log for 'master..seent' > > It’s the same. Oh, right—I guess rev. b75747829f (range-diff: optionally > accept pathspecs, 2022-08-26) that `range-diff` about `--`. > > So in conclusion: IMO `range-diff` has a slight usability regression for > when you mistype the ref. It would be nice if the error message without > a pathspec (`--`/`dash_dash`) was as nice as one without it. > > --- > > What do you think? Should I bring this up on the Git mailing list? Or > have I missed something/misunderstood something? > -- You received this message because you are subscribed to the Google Groups "Git for human beings" group. To unsubscribe from this group and stop receiving emails from it, send an email to git-users+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/git-users/2b97e5cc-cf20-4eae-a691-dabbbe2869e1n%40googlegroups.com.
