Regex selectors more tolerant to matches not ending on char boundaries

2 files changed, 12 insertions, 5 deletions

diff --git a/src/selectors.cc b/src/selectors.cc
index f5fdb088..781ea0e0 100644
--- a/src/selectors.cc
+++ b/src/selectors.cc
@@ -510,8 +510,8 @@ void select_all_matches(SelectionList& selections, const Regex& regex)
 
         for (; re_it != re_end; ++re_it)
         {
-            auto& begin = (*re_it)[0].first;
-            auto& end   = (*re_it)[0].second;
+            auto begin = ensure_char_start(buffer, (*re_it)[0].first);
+            auto end = ensure_char_start(buffer, (*re_it)[0].second);
 
             if (begin == sel_end)
                 continue;
@@ -549,8 +549,9 @@ void split_selections(SelectionList& selections, const Regex& regex)
             if (end == buf_end)
                 continue;
 
+            end = ensure_char_start(buffer, end);
             result.push_back(keep_direction({ begin.coord(), (begin == end) ? end.coord() : utf8::previous(end, begin).coord() }, sel));
-            begin = (*re_it)[0].second;
+            begin = ensure_char_start(buffer, (*re_it)[0].second);
         }
         if (begin.coord() <= sel.max())
             result.push_back(keep_direction({ begin.coord(), sel.max() }, sel));
diff --git a/src/selectors.hh b/src/selectors.hh
index 3dc4e5ba..09ba45e1 100644
--- a/src/selectors.hh
+++ b/src/selectors.hh
@@ -256,6 +256,12 @@ bool find_match_in_buffer(const Buffer& buffer, const BufferIterator pos,
                 find_last_match(buffer.begin(), buffer.end(), matches, ex));
 }
 
+inline BufferIterator ensure_char_start(const Buffer& buffer, const BufferIterator& it)
+{
+    return it != buffer.end() ?
+        utf8::character_start(it, buffer.iterator_at(it.coord().line)) : it;
+}
+
 template<Direction direction>
 Selection find_next_match(const Buffer& buffer, const Selection& sel, const Regex& regex)
 {
@@ -269,8 +275,8 @@ Selection find_next_match(const Buffer& buffer, const Selection& sel, const Rege
                                     : utf8::previous(begin, buffer.begin());
     if ((found = find_match_in_buffer<direction>(buffer, pos, matches, regex)))
     {
-        begin = matches[0].first;
-        end   = matches[0].second;
+        begin = ensure_char_start(buffer, matches[0].first);
+        end = ensure_char_start(buffer, matches[0].second);
         for (auto& match : matches)
             captures.emplace_back(match.first, match.second);
     }