El 21/08/2009, a las 3:14, textmate-request@lists.macromates.com escribió:

On Aug 20, 2009, at 6:37 PM, Juan Falgueras wrote:

...

desperate looking for a way of a more relaxed way of searching a string inside another, we need not to take into the account not only the case

s = "Abc" if s =~ /abc/i then ?

but also if you have forgotten an accent, etc:

s = "?Bc" if s =~ /abc/i then ?also should match!

Changing the encoding to the simplest one: ASCII, does not work since iconv, nor ruby force_encoding() work and gives you errors in case you try to convert "?" to "a"

This might be one option for you:

$ irb -KU -r iconv

...

...

s = "?Bc"

=> "?Bc"

...

...

Iconv.conv("ASCII//TRANSLIT//IGNORE", "UTF-8",

s).downcase.delete("^a-z") =~ /abc/ => 0

I hope that helps.

James Edward Gray II

Thanks James!

I have narrowed the chars you should delete after transliteration. If you try with the ~full range of accented latin chars

require "iconv"

ss= "ÀÁÂÃÄÅĀĄĂÆÇĆČĊĎĐÈÉÊËĒĘĚĔĖĜĞĠĢĤĦÌÍÎÏĪĨĬĮİIJĴĶŁĽĹĻĿÑŃŇŅŊÒÓÔÕÖØŌŐŎŒŔŘŖŚŠŞŜȘŤŢŦȚÙÚÛÜŪŮŰŬŨŲŴÝŶŸŹŽŻàáâãäåāąăæçćĉċďđèéêëēęěĕėƒĝğġģĥħìíîïīĩĭįıijĵķĸłľĺļŀñńňņʼnòóôõöøōőŏœŕřŗśšşŝșťţŧțùúûüūůűŭũųŵÿŷžżźþßſÐð "

# you can see a small set of chars that are used to transliterate: "'~^`

puts Iconv.conv("ASCII//TRANSLIT//IGNORE", "UTF-8", ss).delete(%q{"'~^`})

This is a less aggressive way of patching this annoying behaviour of transliterate. I would like to see new //CLEANDROP or so, to simply drop out those accents to work with the chars. Or simply a regexp search with a modifier that could make it just to ignore accents…

- juan