On Tuesday, 28 December 2021 01:16:40 CET Junio C Hamano wrote: > Ævar Arnfjörð Bjarmason <avarab@xxxxxxxxx> writes: > > > Does it really matter? I.e. we've got plenty of options parsing code in > > various places that might complain about issues A and B with your > > command-line, but will only emit one of those at a time. > > ... > > But until such an improvement something as trivial as minimizing the > > diff size (i.e. not needing "if" -> "else if") seems preferrable to > > slavishly maintaining compatibility with the exact sequence of errors > > we'd emit before. > > The patch is already touching every other line. Making the change > into a replacement of solid block into another solid block would > probably make the resulting patch easier to read. > > And by doing so, the submitter demonstrates to reviewers that they > _care_, as such an attention to detail is a sign that the submitter > thought the ramifications of changing an early die() into an > assignment that keeps going. > > So, yes, it does matter. > Sorry. The developer hat was taken off ; will reroll shortly.
