[rcr] supply option hash to Regexp.new

[Simon Strandgaard]

Regexp.new is old, and relies on C'ish magic values.

Lets write something ugly with the current initialize:
Regexp.new('a.*a', 5, 'U')
Thats the same as /a.*a/uxim

The problem is these legacy values
Regexp::IGNORECASE = 1
Regexp::EXTENDED = 2
Regexp::MULTILINE = 4

I propose to let initialize take a hash.

Regexp.new('a.*a', :ignorecase=>true, :extended=>true, :multiline=>true, :encoding=>:UTF8)


Ideas for other options: :interpret_warnings_as_errors, :verbose_mismatch

 

[Robert Klemme]

+1

 

[Simon Strandgaard]

+1

Robert, +1 ?? if we calculate the flags for 'uxim'

then 'u' => 64, 'x' => 2, 'i' => 1, 'm' => 4 .... sum=71

 

[David A. Black]

Hi --

Robert, +1 ??

I think +1 means "I vote yes" (or in this case perhaps, "I will vote
yes when the RCR is submitted on RCRchive"

I also like the idea of discarding these cryptic Regexp things.


David

 

[Simon Strandgaard]

I think +1 means "I vote yes" (or in this case perhaps, "I will vote
yes when the RCR is submitted on RCRchive"

I also like the idea of discarding these cryptic Regexp things.

ok.. apparently +1 translates into 'i vote yes'.. I guessed so.
Just wanted to point out my that the value 5 was sligthly wrong.
It should have been 7. 5+1 would only be 6 (halfway).

I better fill in an rcr for this.

 

[Simon Strandgaard]

I better fill in an rcr for this.

I have added an RCR entry.. vote on it here

http://rcrchive.net/rcr/RCR/RCR258

 

[Robert Klemme]

ok.. apparently +1 translates into 'i vote yes'..
Right.

I guessed so.
Just wanted to point out my that the value 5 was sligthly wrong.
It should have been 7. 5+1 would only be 6 (halfway).

Halfway is often better than no progress at all. )

Regards

robert