On 15 Feb 2016, at 23:39, Ramsay Jones <ramsay@xxxxxxxxxxxxxxxxxxxx> wrote: > > > On 15/02/16 21:40, Jeff King wrote: >> On Mon, Feb 15, 2016 at 09:36:23PM +0000, Ramsay Jones wrote: >> >>>> +test_expect_success '--show-origin stdin' ' >>>> + cat >expect <<-\EOF && >>>> + stdin: user.custom=true >>> >>> So, as with the previous patch, I think this should be: >>> file:<stdin> user.custom=true >> >> That's ambiguous with a file named "<stdin>", which was the point of >> having the two separate prefixes in the first place. >> >> I think in practice we _could_ get by with an ambiguous output (it's not >> like "<stdin>" is a common filename), but that was discussed earlier in >> the thread, and Lars decided to go for something unambiguous. > > sure, I just don't think it would cause a problem in practice. > How about using '-' for <stdin>? Hmm, you can actually create > such a file in the filesystem! Oh well, I guess its not a big deal. > >> >> That doesn't necessarily have to bleed over into the error messages, >> though (which could continue to use "<stdin>" if we want to put in a >> little extra code to covering the cases separately. > > Yep. OK, I am happy to add the extra code. However, out of curiosity, can you explain in what cases you actually use configs from stdin? I wasn't aware of this feature before working on this patch and I still wonder when I would use it. If it is only a seldom used feature then I am not sure if adding the extra code to restore the existing error message is worth the effort? Thanks, Lars -- 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