develooper Front page | perl.perl5.porters | Postings from February 2013

Re: Missing diagnostics - non C task

Thread Previous | Thread Next
Karl Williamson
February 8, 2013 04:18
Re: Missing diagnostics - non C task
Message ID:
On 02/07/2013 08:11 PM, James E Keenan wrote:
> On 2/7/13 11:06 AM, demerphq wrote:
>> We apparently auto scan for diagnostics, and then auto TODO ones that
>> are older than a certain age.
> Would it be possible to post this as an RT?  We could then reference it
> in upcoming hackathons?
> Also, I'm not sure who is the "we" you are referring to and where that
> "auto scan" or "auto TODO"-ing takes place.  Could you elaborate a bit
> so that a non-core/P5P person could pick this up more quickly?
> Thank you very much for raising the issue.
> Jim Keenan

The test program t/porting/diag.t, run by doing a 'make test', greps 
through the Perl source looking for things that appear to be diagnostic 
messages, and then looks to see if each one is listed in 
pod/perldiag.pod.  If not, it fails that test.  In the .t are 
hard-coded a bunch of diagnostics which are grandfathered in, because any of

1) there was no explanation for them at the time the test program was 

2) someone was too lazy to write one, and just added it to the 
grandfather list;

3) the .t was buggy and didn't correctly match the diagnostic with the 
existing explanation

4) the person was too busy at the moment, and is planning to add 
explanations before the next release is shipped.

The goal is to go through that list and write the prose explanations to 
add to the pod.

The problem is that someone who isn't very familiar with the code in 
question can write a misleading explanation.  Even someone who is very 
familiar would do well to re-examine the code to verify that s/he isn't 
misleading people by not catching all the appropriate nuances.

Thread Previous | Thread Next Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at | Group listing | About