proposal: use ~ instead of =~ for regular expression constraints

[myitcv]

Currently (per the spec), the =~ operator is defined as follows:

s =~ r is true if s matches the regular expression r.

This issue tracks changing that operator to instead be simply ~, with the same definition.

Motivation etc.

(to follow from @mpvl)

[seh]

What is the motivation for the change? I find that =~ goes along with other inequality operators such as <= and >=, though perhaps that analogy would argue in favor of ~=.

[myitcv]

What is the motivation for the change?

Thanks. I'm just in the process of creating the umbrella issues we referenced from Community Call #1, hence why some issues are a bit light on detail. I'll defer to @mpvl to flesh this out (and will add a section above to highlight that it's pending).

[mpvl]

@seh, one big motivation is how ugly it is to use regexps in a pattern constraint with an alias:

[X= =~"regexp"]: T

[seh]

I agree: That sure is ugly. I assume that you don’t like the transposed proposal any better:

[X= ~="regexp"]: T

[eonpatapon]

I agree it would be better in the context of pattern constraint.

I also regularly forgot if it is =~ or ~= when writing configs.

Logically I think =~ is the correct way as we also have !~. Would be awkward to have ~= and !~.

Having ~ and !~ feels ok to me.