lucy-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Marvin Humphrey <mar...@rectangular.com>
Subject Re: [lucy-dev] AUTHOR_MODE and Podchecker
Date Mon, 12 Mar 2012 15:38:33 GMT
On Sun, Mar 11, 2012 at 2:23 PM, Nick Wellnhofer <wellnhofer@aevum.de> wrote:
> On 09/03/2012 20:16, Logan Bell wrote:
>>
>> I did go and fix the pod checker issues, however, it might be good to just
>> disable pod checker unless the tests are being run by an author. Below is
>> an example of what I mean in 801-pod_checker.t.
>>
>> BEGIN {
>>    *plan skip_all =>  "Author mode not enabled SKIPPING"*
>>        unless $ENV{AUTHOR_MODE};
>>    if ( $]<  5.010 ) {
>>        plan( 'skip_all', "Old Pod::Checker is buggy" );
>>    }
>>    else {
>>        plan('no_plan');
>>    }
>> }
>
>
> +1 for this.

-0, provided that we run this test as part of the community release prep
extended tests, as we do with running test_valgrind.  -1 if this test will
never be run.

    http://wiki.apache.org/lucy/ReleasePrep

Does anybody here actually run with AUTHOR_MODE set?  I'm assuming that nobody
does and that therefore the consequence of this action will be that bad POD
gets committed and that somebody other than the original contributor will get
the privilege of cleaning up the mess just prior to release.

I'm not real excited about letting the original contributor off the hook for
contributing malformed POD -- it's like turning off 'use strict', and it
doesn't help anybody.  I also dislike the idea that if we try to nip this in
the bud we will have to inform someone directly about their bad POD after code
review -- IMO having the machine tell you that you screwed up is less
personal.

Marvin Humphrey

Mime
View raw message