Re: [Tools-discuss] Idnits-2.16.02 feedback: -latest pattern

[Carsten Bormann <[e-mail]>]

On Apr 6, 2019, at 18:58, Henrik Levkowetz <[e-mail]> wrote:
> 
> Hi Carsten,
> 
> On 2019-04-06 18:43, Carsten Bormann wrote:
>> Idnits should recognize the -latest pattern that Martin Thomson’s
>> template uses and turn these four messages into one (or maybe even
>> fix the fourth, because that very much can be figured out based on
>> the pattern):
> 
> I don't see that from the information given.
> 
> 1) If idnits was made more permissive to accept these, and they actually
>    occurred in a submitted document by mistake, they would not be flagged

They would be, in one message instead of four (that all need to be actively ignored by a human each time an idnits check is made).

> 2) I don't see why idnits should be run on something which hasn't been
>    processed all the way through to exact submission state, which seems to
>    be what's happening here -- if the processing pipeline is able to produce
>    something ready for submission, why is idnits run on something else?

Because it is expedient.  Documents spend most of their time in the -latest stage; doing a quick kdrfc/idnits is a good thing to do after not having touched a document for a couple of weeks.

Grüße, Carsten