branch: externals/consult
commit 110b560ae1e0ad398d5ce2b4aaf0e91f87f38297
Author: Daniel Mendler <[email protected]>
Commit: Daniel Mendler <[email protected]>
Regexp translation: Add crude support for some syntax classes
---
consult.el | 8 ++++++--
1 file changed, 6 insertions(+), 2 deletions(-)
diff --git a/consult.el b/consult.el
index 78c6a6638c..0760537e93 100644
--- a/consult.el
+++ b/consult.el
@@ -1147,7 +1147,10 @@ if IGNORE-CASE is non-nil."
;; since PCRE does not make this distinction. Usually the
;; context determines if \b is the beginning or the end.
'(("\\<" . "\\b") ("\\>" . "\\b")
- ("\\_<" . "\\b") ("\\_>" . "\\b"))
+ ("\\_<" . "\\b") ("\\_>" . "\\b")
+ ("\\s-" . "[ \\n\\t\\r]") ("\\S-" . "[^ \\n\\t\\r]")
+ ("\\sw" . "[a-zA-Z0-9]") ("\\Sw" . "[^a-zA-Z0-0]")
+ ("\\s_" . "[a-zA-Z0-9_-]") ("\\S_" . "[^a-zA-Z0-0_-]"))
;; Treat \` and \' as beginning and end of line. This is more
;; widely supported and makes sense for line-based commands.
'(("\\`" . "^") ("\\'" . "$"))
@@ -1172,7 +1175,7 @@ if IGNORE-CASE is non-nil."
;; Support for Emacs regular expressions is fairly complete for basic
;; usage. There are a few unsupported Emacs regexp features:
;; - \= point matching
- ;; - Syntax classes \sx \Sx
+ ;; - Most syntax classes \sx \Sx
;; - Character classes \cx \Cx
;; - Explicitly numbered groups (?3:group)
(replace-regexp-in-string
@@ -1182,6 +1185,7 @@ if IGNORE-CASE is non-nil."
(seq (or bos "^") (any "*+?")) ;; Historical: + or * at the
beginning
(seq (opt "\\") (any "(){|}")) ;; Escape parens/braces/pipe
(seq "\\" (any "'<>`")) ;; Special escapes
+ (seq "\\" (any "Ss") (any "-w_")) ;; Whitespace, word, symbol
syntax class
(seq "\\_" (any "<>")))) ;; Beginning or end of symbol
(lambda (x) (or (cdr (assoc x consult--convert-regexp-table)) x))
regexp 'fixedcase 'literal)))
