[gtksourceview/wip/regex-search] Regex search: document known issue with word boundaries



commit a0392dd6b6c5649bc28b52ea8e075591481b9bef
Author: Sébastien Wilmet <swilmet gnome org>
Date:   Sun Jul 21 18:06:36 2013 +0200

    Regex search: document known issue with word boundaries

 gtksourceview/gtksourcesearch.c |   14 ++++++++++++++
 1 files changed, 14 insertions(+), 0 deletions(-)
---
diff --git a/gtksourceview/gtksourcesearch.c b/gtksourceview/gtksourcesearch.c
index eb53f8d..15f057b 100644
--- a/gtksourceview/gtksourcesearch.c
+++ b/gtksourceview/gtksourcesearch.c
@@ -151,6 +151,20 @@
  *   several idle loop iterations. (2) Once the subject string is retrieved
  *   completely, we launch the regex matching in a thread. (3) Once the thread
  *   is finished, we highlight the matches in the buffer. And voilà.
+ *
+ * Known issue:
+ * To search at word boundaries, \b is added at the beginning and at the
+ * end of the pattern. But \b is not the same as
+ * gtk_text_iter_starts_word() and gtk_text_iter_ends_word(). \b for
+ * example doesn't take the underscore as a word boundary.
+ * Using gtk_text_iter_starts_word() and ends_word() for regex searches
+ * is not easily possible: if the GRegex return a match, but doesn't
+ * start and end a word, maybe a shorter match (for a greedy pattern)
+ * start and end a word, or a longer match (for an ungreedy pattern). To
+ * be able to use the gtk_text_iter_starts_word() and ends_word()
+ * functions for regex search, g_regex_match_all_full() must be used, to
+ * retrieve _all_ matches, and test the word boundaries until a match is
+ * found at word boundaries.
  */
 
 /*


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]