Syntax table entries for non-alphanumeric chars

[ikirill]

See #410

List of characters generated with the unicode-fonts package

(mapc
 (lambda (block)
   (when (and (string-match-p "operators\\|punctuation\\|symbol\\|arrow"
                              (car block))
              (not (string-match-p "alpha\\|letterlike" (car block))))
     (insert
      (format "\n%-45s ; %s"
              (prin1-to-string
               `(modify-syntax-entry (cons ,(nth 1 block) ,(nth 2 block)) "_" table))
              (car block)))))
 unicode-fonts-blocks)

I hope I didn't miss any important characters.

[purcell]

Excellent, thanks!

• Do we know if this is the same list of ranges Haskell itself is using?
• Can you check if ⧺ can now be removed from the string of characters above? If so, you can amend the commit and force-push it to update the PR.
• mapc over a list would perhaps be a little cleaner than repeatedly calling modify-syntax-entry.

[ikirill]

• I added enclosed alphanumeric characters, they are not alphanumeric (http://www.fileformat.info/info/unicode/char/2460/index.htm)
• ghc uses Unicode character category (see https://github.com/ghc/ghc/blob/master/compiler/basicTypes/Lexeme.hs#L203), which is per-character, not per-block. I'm sure this list of blocks is precise enough for most uses.
• Yes, DOUBLE PLUS works fine.
• I used dolist.

[purcell]

Great! Merged.