Add more function declarations

* lisp/subr.el (buffer-narrowed-p, sha1, match-substitute-replacement)
(version-to-list, version<, version<=, version=)
(function-get, subregexp-context-p, split-string)
(combine-and-quote-strings, split-string-and-unquote)
(replace-regexp-in-string, syntax-after)
(string-trim-left, string-trim):
* lisp/emacs-lisp/subr-x.el (hash-table-empty-p, hash-table-keys)
(hash-table-values, string-glyph-split)
(string-clean-whitespace, string-fill, string-limit)
(string-pixel-width):
* lisp/env.el (substitute-env-vars, substitute-env-in-file-name)
(setenv-internal):
* lisp/emacs-lisp/rx.el (rx-to-string):
* lisp/emacs-lisp/regexp-opt.el (regexp-opt-depth)
(regexp-opt-charset):
Add appropriate declarations: pure, side-effect-free, and/or
important-return-value.

lisp/emacs-lisp/regexp-opt.el

@@ -154,6 +154,7 @@ usually more efficient than that of a simplified version:
   "Return the depth of REGEXP.
 This means the number of non-shy regexp grouping constructs
 \(parenthesized expressions) in REGEXP."
+  (declare (pure t) (side-effect-free t))
   (save-match-data
     ;; Hack to signal an error if REGEXP does not have balanced parentheses.
     (string-match regexp "")
@@ -270,6 +271,7 @@ Merges keywords to avoid backtracking in Emacs's regexp matcher."
 CHARS should be a list of characters.
 If CHARS is the empty list, the return value is a regexp that
 never matches anything."
+  (declare (pure t) (side-effect-free t))
   ;; The basic idea is to find character ranges.  Also we take care in the
   ;; position of character set meta characters in the character set regexp.
   ;;