aboutsummaryrefslogtreecommitdiffstats
path: root/lib/tools/emacs/erlang.el
diff options
context:
space:
mode:
Diffstat (limited to 'lib/tools/emacs/erlang.el')
-rw-r--r--lib/tools/emacs/erlang.el62
1 files changed, 49 insertions, 13 deletions
diff --git a/lib/tools/emacs/erlang.el b/lib/tools/emacs/erlang.el
index bc7a190fb4..8f98d6c85c 100644
--- a/lib/tools/emacs/erlang.el
+++ b/lib/tools/emacs/erlang.el
@@ -1516,7 +1516,7 @@ Other commands:
;; A dollar sign right before the double quote that ends a
;; string is not a character escape.
;;
- ;; And a "string" has with a double quote not escaped by a
+ ;; And a "string" consists of a double quote not escaped by a
;; dollar sign, any number of non-backslash non-newline
;; characters or escaped backslashes, a dollar sign
;; (otherwise we wouldn't care) and a double quote. This
@@ -1525,6 +1525,8 @@ Other commands:
;; know whether matching started inside a string: limiting
;; search to a single line keeps things sane.
. (("\\(?:^\\|[^$]\\)\"\\(?:[^\"\n]\\|\\\\\"\\)*\\(\\$\\)\"" 1 "w")
+ ;; Likewise for atoms
+ ("\\(?:^\\|[^$]\\)'\\(?:[^'\n]\\|\\\\'\\)*\\(\\$\\)'" 1 "w")
;; And the dollar sign in $\" escapes two characters, not
;; just one.
("\\(\\$\\)\\\\\\\"" 1 "'"))))))
@@ -2986,18 +2988,52 @@ This assumes that the preceding expression is either simple
(forward-sexp (- arg))
(let ((col (current-column)))
(skip-chars-backward " \t")
- ;; Needed to match the colon in "'foo':'bar'".
- (if (not (memq (preceding-char) '(?# ?:)))
- col
- ;; Special hack to handle: (note line break)
- ;; [#myrecord{
- ;; foo = foo}]
- (or
- (ignore-errors
- (backward-char 1)
- (forward-sexp -1)
- (current-column))
- col)))))
+ ;; Special hack to handle: (note line break)
+ ;; [#myrecord{
+ ;; foo = foo}]
+ ;; where the call (forward-sexp -1) will fail when point is at the `#'.
+ (or
+ (ignore-errors
+ ;; Needed to match the colon in "'foo':'bar'".
+ (cond ((eq (preceding-char) ?:)
+ (backward-char 1)
+ (forward-sexp -1)
+ (current-column))
+ ((eq (preceding-char) ?#)
+ ;; We may now be at:
+ ;; - either a construction of a new record
+ ;; - or update of a record, in which case we want
+ ;; the column of the expression to be updated.
+ ;;
+ ;; To see which of the two cases we are at, we first
+ ;; move an expression backwards, check for keywords,
+ ;; then immediately an expression forwards. Moving
+ ;; backwards skips past tokens like `,' or `->', but
+ ;; when moving forwards again, we won't skip past such
+ ;; tokens. We use this: if, after having moved
+ ;; forwards, we're back where we started, then it was
+ ;; a record update.
+ ;; The check for keywords is to detect cases like:
+ ;; case Something of #record_construction{...}
+ (backward-char 1)
+ (let ((record-start (point))
+ (record-start-col (current-column)))
+ (forward-sexp -1)
+ (let ((preceding-expr-col (current-column))
+ ;; white space definition according to erl_scan
+ (white-space "\000-\040\200-\240"))
+ (if (erlang-at-keyword)
+ ;; The (forward-sexp -1) call moved past a keyword
+ (1+ record-start-col)
+ (forward-sexp 1)
+ (skip-chars-forward white-space record-start)
+ ;; Are we back where we started? If so, it was an update.
+ (if (= (point) record-start)
+ preceding-expr-col
+ (goto-char record-start)
+ (1+ (current-column)))))))
+ (t col)))
+ col))))
(defun erlang-indent-parenthesis (stack-position)
(let ((previous (erlang-indent-find-preceding-expr)))