On Thu, Aug 29, 2024 at 1:10 PM Jeff King <peff@xxxxxxxx> wrote: > On Thu, Aug 29, 2024 at 12:03:37PM +0200, Patrick Steinhardt wrote: > > Would become this: > > > > t/chainlint/pipe.actual:8: error: expected ampersands (&&) > > 7 fish | > > 8 cow > > ^ > > 9 > > > > While this would be neat, I guess it would also be way more work than > > the current series you have posted. And whether that work is ultimately > > really worth it may be another question. Probably not. > > I think that output is quite readable. One bonus is that it follows the > usual "quickfix" format, so there's editor support for jumping to the > problematic spot. The "quickfix" notation has more appeal (to me) than the "^" annotation since it is immediately useful in an editor and doesn't require extra cogitation. A couple concerns which come to mind are that it makes the output even more noisy (which could be distracting), and that the path, if relative, might not correspond to the editor's "cwd" or to a path in the editor's search list, so the promised "jump to next problem" automation may not materialize. > It probably is more verbose if you have multiple errors right next to > each other (since now we just show the annotated source text). But that > is going to be relatively rare compared to single mistakes, I'd think. Maybe. Maybe not. A first-draft test by a newcomer might very well break the &&-chain in many spots. Nevertheless, I'm not likely to implement this any time soon (or at all).
