>>>>> On Fri, 28 Mar 2008 23:17:48 -0400, "David Golden" <[EMAIL PROTECTED]> >>>>> said:
> I wouldn't suggest it be treated as a "PASS" but rather as a "DISCARD" > -- meaning that no report is sent. While I get the point that it's a > "bug" to ship future timestamps, I suspect that in most cases, it's > only a temporary issue and most people won't have an issue with it. > (Possibly because most people use a CPAN mirror that has some > substantial replication delay.) Sure, but then you need to branch between short and long timestamp issues. And between Module::Build and MakeMaker. And locale dependent parsing. And take into account the version of make. I'm tempted to introduce a config variable 'allow_future_timestamps' to CPAN.pm which defaults to false and change the logic to refuse to make any distro containing files in the future unless - allow_future_timestamps variable is true or - there is a Build.PL to avoid the issue altogether. Then the remaining conflicts can be resolved between users and authors directly and we testers stop interfering at all. And those who test later can test normally. Opinions? -- andreas
