"Parsing error, this should not happen, send us a report!" about app-cpanminus-reporter HOT 4 CLOSED

This is a tough question, I admit. I tend to believe that the reporter
and the tool writer share some responsibility. If people test perls that
deviate strongly from the standard, their reports will not be useful. We
can presume, they are trying to be helpful and will only do small
tweaking here and there. I do not consider it the responsibility of the
toolchain writers to prevent reports from silghtly altered environments.
It's probably good to not send a report when the toolchain definitely
knows that a patch has been applied. CPAN.pm shell refuses to send a
report on a distro that got patched via distroprefs, but we do not carry
this to the extreme to stop reporting on such a perl forever.

It remains to be suggested that people can add comments to their reports
when they have something personally to say, so we should let them do
what they find useful. Bottom line: do not try to prevent reports that
are provided in good faith.

I must reserve the right to change my opinion, things may move in
certain so far not encountered directions and may make a new evaluation
of thought necessary.

from app-cpanminus-reporter.

I locally tweaked the log to have no Entering /whatever in it, hit cpanm-reporter again, and it worked.

from app-cpanminus-reporter.

Hey, @kentfredric! Thanks for the report and for all the patience - I know it's been a really long time.

Actually, if you fiddle with a distribution, cpanm-reporter should not send a report, right? I mean, it'd be cheating, as the tested version is not the same one as the one on CPAN.

We should probably make the parser explicitly test for that and skip those distributions, I'll give it a go eventually, but patches (and general opinions on the subject) are very much welcome!

Cheers!

from app-cpanminus-reporter.

Possibly. But fixing a target so it works should have little bearing on the reporting of the other things.

Personally I think "non-vanilla dists" ( ie: patched by user, however it was patched ) reporting "pass" can be a good thing, but in order for it to be a good thing, it needs to be more obvious in the report that this was done.

Then ANDK's analysis can see "X is OK" -> "Why is X working where everyone else failed" -> "X is specially patched" to work.

Then its a matter of tracking down an existing bug with said patch, or tracking down the test reporter and asking them what they did.

from app-cpanminus-reporter.