On Wed, 31 Oct 2001, Chris Hostetter wrote: > Forgive me if this is an idea that's allready been considered/rejected in > the past, but shouldn't it be possible to add a "tester" configuration > option to CPAN.pm such that anytime tests are run (either directly, or > indirectly via "install"), testers.cpan.org is > contacted to determine if the results are "usefull" (ie: there are no test > results for the current package/version/platform, or the listed results > contradict the current results) it automatically submits them > (presumably via email, similar to "cpantest" script).
Although I don't think the idea of automatically sending test results has been considered, interfacing the cpantest script with CPAN.pm has been discussed before. Determining what's "useful" or not could be tricky, as there may be variations among similar platforms that might affect the test results. Automatically sending test results for anything but a "pass" is probably not desireable, as usually some comments in the report are in order explaining the results, which means human interaction anyway. > My reason for asking is that even though I may not have the time/energy > to become an active tester, I would gladly wait an extra few seconds every > time I install a new module to let the CPAN shell contribute the > test results on my behalf. > > While it may not make sense to have the behavior be the default, I'm sure > that a significant percentage of the people who use CPAN would turn an > option like that on if it existed. An (alpha) version of a module that integrates the cpantest script within the CPAN.pm shell is available at ftp://theoryx5.uwinnipeg.ca/pub/other/CPAN-Report-0.91.tar.gz This allows one, within the shell, to submit reports via a "cpantest" method, using much the same syntax as the cpantest script. The ability to automatically send test results after installing a module, at least for successful installations, could be added to this, if there's a desire. best regards, randy kobes
