* doc/lispref/searching.texi (Char Classes):
* lisp/emacs-lisp/rx.el (rx): Document [:graph:] to be [:print:]
sans whitespace (not sans space).
* src/character.c (graphicp): Exclude all Unicode whitespace chars,
not just space.
* src/regex.c (ISGRAPH): Exclude U+00A0 (NO-BREAK SPACE).
In POSIX [[:print:]] is equivalent to [ [:graph:]], so change
[:graph:] so that it matches everything that [:print:] does,
except for space.
* doc/lispref/searching.texi (Char Classes):
* etc/NEWS:
* lisp/emacs-lisp/rx.el (rx):
Document [:graph:] to be [:print:] sans ' '.
* src/character.c, src/character.h (graphicp): New function.
* src/regex.c (ISGRAPH) [emacs]: Use it.
(BIT_GRAPH): New macro.
(BIT_PRINT): Increase to 0x200, to make room for BIT_GRAPH.
(re_wctype_to_bit) [! WIDE_CHAR_SUPPORT]:
Return BIT_GRAPH for RECC_GRAPH.
(re_match_2_internal) [emacs]: Use ISGRAPH if BIT_GRAPH,
and ISPRINT if BIT_PRINT.
* src/regex.c (ISPRINT): Call 'printablep' for multibyte characters.
(BIT_PRINT): New bit mask.
(re_wctype_to_bit): Return BIT_PRINT for RECC_PRINT.
* src/character.c (printablep): New function.
* src/character.h (printablep): Add prototype.
* lisp/emacs-lisp/rx.el (rx): Doc fix: document the new behavior
of 'print', 'alnum', and 'alphabetic'.
* doc/lispref/searching.texi (Char Classes): Document the new
behavior of [:print:].
* etc/NEWS: Mention the new behavior of [:print:].
(rx-constituents): Add `any => "."', mark `repeat' as taking any number
of args, add `regex' alias.
(rx-info): Add arg to distinguish head and standalone forms.
(rx-check, rx-form): Pass the corresponding arg.
(rx-**): Simplify.
(rx-repeat): Make it work for any number of args.
(rx-syntax): Make it accept syntax chars as is.
* lisp/obsolete/sregex.el: Move from emacs-lisp/.
* lisp/emacs-lisp/re-builder.el: Remove sregex support.
* lisp/emacs-lisp/edebug.el (sregexq, rx): Remove redundant defs.
rx-anything. Change `not-wordchar' assignment to "\\W" from
"[^[:word:]]".
(rx-group-if): New function.
(rx-parent): New variable.
(rx-and, rx-or): Put shy groups only when necessary.
(rx-bracket): Remove.
(rx-anything): New function.
(rx-any-delete-from-range, rx-any-condense-range)
(rx-check-any-string): New functions.
(rx-check-any): Return result as a list. Don't convert chars to
strings. Don't prepend "\\" to "^". Don't search for close
bracket. Check char category string. Call rx-form instead of
rx-to-string.
(rx-any): Rebuid to complete the function.
(rx-check-not): Fix char category regexp pattern string. Call
rx-form instead of rx-to-string.
(rx-not): Call rx-form instead of rx-to-string. Convert "[^]" to
"[^^]". Call regexp-quote for one char string when not called from
rx-not. Add "\\w", and toggle to upcase. Add the case of
"\\[SCBW]" to toggle.
(rx-=, rx->=, rx -**, rx-repeat, rx-submatch): Call rx-form
instead of rx-to-string.
(rx-kleene): Call rx-form instead of rx-to-string. Call
rx-group-if to adjust putting of shy groups.
(rx-atomic-p): Make check more precisely.
(rx-eval, rx-greedy): Call rx-form instead of rx-to-string.
(rx-regexp): Call rx-group-if.
(rx-form): New function.
(rx-to-string): Call rx-form, rx-group-if. Refine definition of
NO-GROUP.
