Last day to vote/reject on proposed EOL names

[Alexander Belchenko]

I think this should satisfy most of the user and their needs.
I'm OK with very-verbose-and-long-options-to-make-it-clear.
It's fine.

Just one suggestion about help: I think you have to say right after first table with options
about "exact" as default when there is no rules specified. Because you list "native"
as first item, so some people may think it's default. Just IMO.

Ian Clatworthy пишет:
> Brian de Alwis wrote:
> 
>> The configuration expressed as a single option is semantically identical
>> to the configuration split as two separately labelled options.  In the
>> split-out case, join the two options with a colon and then proceed as in
>> the single-option case.  You're pointing out that not all combinations
>> are safe -- so the code should error out on unsafe combinations.
>>
>> Splitting the options with meaningful labels helps in making the
>> configuration self-documenting.  I know I'll have to look up the
>> documentation each time to figure out the position meanings of
>> "crlf:lf", especially since I'll only (hopefully) rarely need to look at
>> this option.
> 
> So I've been considering everyone's feedback and here's my decision.
> I'm going to go with a single option because:
> 
> 1. I know it's much easier to implement
> 2. I strongly suspect it will be much easier to support.
> 
> More likely than not, it will be me trying to work out why some poor
> user is having problems and that will be harder with multiple settings,
> possibly configured for different patterns and in different rule
> files (once we support that).
> 
> I'm very sensitive to the self-documenting issue though, so I'm happy
> to make the uncommonly used values verbose. See the attached help for
> the pending updated patch.
> 
> Ian C.
>