On Tue, Sep 02, 2008 at 01:23:31PM -0700, chromatic wrote: > If the CPAN Testers client already requires developers to modify their > distributions to tease useful information out of reports, then CPAN Testers > has failed at being an accurate representation of how distributions fare for > normal users.
As a developer, I find that the best way of getting more information about a failure, if the contents of the report isn't good enough, is to politely ask the person responsible for the report for more info. This applies equally to CPAN-testers reports and to those that people submit manually using RT. > I already know that my distributions don't work if you don't install the > dependencies I'm pretty damned sure that this a straw man. Can you point at any regular tester who *right now* is regularly failing to follow the dependency chain? > or if you use an unsupported version of Perl The software can't magically divine what versions of perl you care about. I'm sure you are aware of that. Therefore you need to specify which versions you care about. Obviously this information needs to be in some structured form, not just you telling people on a mailing list, or noting it in the documentation or whatever. Again, is there any tester who *right now* is regularly getting this bit wrong? > It certainly does the users no good, especially if they just want to know if > they should bother trying to install one particular version on their machine. > (Can you answer that from the CPAN Testers results on any distribution's > search.cpan.org page without knowing intimate details of how CPAN Testers > works?) Yes. From this page: http://www.cpantesters.org/show/Data-Compare.html I can see that the most recent version of Data-Compare has no failure reports, and I can see which versions of perl it has been tested with on which platforms. I can also see that version 0.13 didn't work on two completely unrelated platforms (Linux and Windows) with perl 5.6.something. There are also other tools which present the information in other ways, eg: http://deps.cpantesters.org/?module=Data::Compare;perl=latest http://bbbike.radzeit.de/~slaven/cpantestersmatrix.cgi?dist=Data-Compare+1.21 both of which are linked from search.cpan.org results pages. > Like Andy Lester suggests, I'm obviously not the target audience of CPAN > Testers. I don't believe normal users are the target audience of CPAN > Testers. Who is? I find it useful as a module developer (I get bug reports for my code) and as a module user (when I have a choice, the data are useful for figuring out which module I should depend on - eg, if two modules both provide the function I need, I'll prefer the one that doesn't suffer from "all the world's a Free Unix" disease because I want my code to run on Solaris as well). It is *more* useful as a module developer though. Follow-ups set to cpan-testers-discuss. -- David Cantrell | top google result for "internet beard fetish club" PLEASE NOTE: This message was meant to offend everyone equally, regardless of race, creed, sexual orientation, politics, choice of beer, operating system, mode of transport, or their editor.