Anchor and restrict citekey regular expression

Limit the characters that are allowed in the citekey to letters, digits,
underscores, and hyphens.  This allows the citekey to be treated as a
word (temporarily expanding word constituents to include underscore and
hyphen), as opposed to explicitly specifying the characters that can
come before a citekey.

3 files changed, 35 insertions, 17 deletions

diff --git a/NEWS b/NEWS
index 7260c01..7c453f1 100644
--- a/NEWS
+++ b/NEWS
@@ -36,6 +36,10 @@
 - =bog-goto-citekey-heading-in-buffer= and
   =bog-goto-citekey-heading-in-notes= will now widen the buffer if the
   heading is outside of the narrowed region.
+
+- The format for citekeys is now restricted to letters, digits,
+  underscores, and hyphens.
+
 - =bog-notes-directory= has been renamed to =bog-note-directory=.
 
 * v0.6.0
diff --git a/bog-tests.el b/bog-tests.el
index 598dc31..0173c5f 100644
--- a/bog-tests.el
+++ b/bog-tests.el
@@ -121,7 +121,19 @@ point at the beginning of the inserted text."
   (let ((citekey "hyphen-name2010word"))
     (with-temp-buffer
       (insert citekey)
+      ;; At beginning
       (goto-char (point-min))
+      (should (equal (bog-citekey-at-point) citekey))
+      ;; On hyphen
+      (skip-chars-forward "-")
+      (should (equal (bog-citekey-at-point) citekey))
+      ;; After hyphen
+      (forward-char)
+      (should (equal (bog-citekey-at-point) citekey))
+      ;; On word
+      (skip-chars-forward "0-9")
+      ;; At year
+      (skip-chars-forward "-a-z")
       (should (equal (bog-citekey-at-point) citekey)))))
 
 ;; `bog-citekey-from-tree'
diff --git a/bog.el b/bog.el
index 6addb31..11bca3d 100644
--- a/bog.el
+++ b/bog.el
@@ -41,13 +41,16 @@
   :group 'org)
 
 (defcustom bog-citekey-format
-  "\\([a-z]+[-a-z]*\\)\\([0-9]\\{4\\}\\)\\([a-z][a-z0-9]*\\)"
+  "\\b\\([a-z]+[-a-z]*\\)\\([0-9]\\{4\\}\\)\\([a-z][a-z0-9]*\\)\\b"
   "Regular expression used to match study citekey.
 
 By default, this matches any sequence of lower case
 letters (allowing hyphenation) that is followed by 4 digits and
 then lower case letters.
 
+The format must be anchored by '\b' and should be restricted to
+letters, digits, '-', and '_'.
+
 This is case-sensitive (i.e., case-fold-search will be set to
 nil).
 
@@ -62,15 +65,6 @@ settings:
   :group 'bog
   :type 'string)
 
-(defcustom bog-allowed-before-citekey
-  "\\(\n\\|\\s-\\|(\\|\\[\\|{\\|<\\|,\\)"
-  "Regex that specifies characters allowed before a citekey.
-This may need to be modified if you have a custom
-`bog-citekey-format' or if you tend to used a certain character
-before citekeys that isn't included above."
-  :group 'bog
-  :type 'string)
-
 (defcustom bog-citekey-property "CUSTOM_ID"
   "Property name used to store citekey.
 The default corresponds to the default value of
@@ -244,17 +238,25 @@ year, and the first meaningful word in the title)."
       (mapconcat (lambda (g) (match-string-no-properties g citekey))
                  groups delim))))
 
+(defmacro bog--with-citekey-syntax (&rest body)
+  "Execute BODY with hyphen and underscore as word constituents."
+  (declare (indent 0))
+  `(with-syntax-table (copy-syntax-table (syntax-table))
+     (modify-syntax-entry ?- "w")
+     (modify-syntax-entry ?_ "w")
+     ,@body))
+
 (defun bog-citekey-at-point ()
   "Return citekey at point.
 The citekey must have the format specified by
-`bog-citekey-format' and, if not at the beginning of the buffer,
-be preceded by a characters in `bog-allowed-before-citekey'."
+`bog-citekey-format'.  Hyphens and underscores are considered as
+word constituents."
   (save-excursion
-    (unless (bobp)
-      (re-search-backward bog-allowed-before-citekey)
-      (forward-char 1))
-    (and (looking-at bog-citekey-format)
-         (match-string-no-properties 0))))
+    (bog--with-citekey-syntax
+      (skip-syntax-backward "w")
+      (let (case-fold-search)
+        (and (looking-at bog-citekey-format)
+             (match-string-no-properties 0))))))
 
 (defun bog-citekey-from-surroundings ()
   "Get the citekey from the context of the Org file."