Hi, 2011/6/7 Thomas Adam <thomas@xxxxxxxxxx>: > Hmm. What guarantees can you make about the scalar value of @cmd here > once the split has happened? Do you not care, or do you have a > hard-limit for split to use? The verifications on the different cmd[x] are made in the few lines "if ... elsif... elsif ... else" right after the split. > These explicit calls to "STDOUT" here -- why? It's redundant typing > and unsightly, and certainly not very perl-y. Is there some reason > why you're using this style? We wanted to highlight the fact that we are sending those lines to STDOUT which is piped into the caller of the remote helper. > >> + print STDOUT "? refs/heads/master\n"; >> + print STDOUT '@'."refs/heads/master HEAD\n"; >> + print STDOUT "\n"; > > Hmm. Why not just: > > print STDOUT '@'."refs/heads/master HEAD\n\n"; Here, we wanted to stick to the convention given by the remote-helper documentation that indicates the list command should end with a blank line. > > There's certainly a large number of perl-specific clean-ups I'd be > inclined to do -- and if I get time later, I might show you how. But > don't let this necessarily put off this patch for inclusion or > anything like that. As said in previous mails, we are relatively new using perl (it's been roughly two weeks) so any advice on how to get things more perl-ish are welcome. Thanks for your feedback -- Arnaud Lacurie -- 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