Fix detection of URLs in HTML source code (URLs inside quotes)

Fixes #785

docs/changelog.rst

@@ -42,6 +42,9 @@ Changelog
   (:iss:`754`)
 
 
+- Fix detection of URLs in HTML source code (URLs inside quotes) (:iss:`785`)
+
+
 0.11.3 [2018-07-10]
 ------------------------------
 

kittens/hints/main.py

@@ -170,7 +170,7 @@ def regex_finditer(pat, minimum_match_length, text):
             yield s, e
 
 
-closing_bracket_map = {'(': ')', '[': ']', '{': '}', '<': '>'}
+closing_bracket_map = {'(': ')', '[': ']', '{': '}', '<': '>', '*': '*', '"': '"', "'": "'"}
 opening_brackets = ''.join(closing_bracket_map)
 postprocessor_map = {}
 
@@ -189,15 +189,15 @@ def url(text, s, e):
             e -= len(url) - idx
     while text[e - 1] in '.,?!' and e > 1:  # remove trailing punctuation
         e -= 1
+    # truncate url at closing bracket/quote
+    if s > 0 and e <= len(text) and text[s-1] in opening_brackets:
+        q = closing_bracket_map[text[s-1]]
+        idx = text.find(q, s)
+        if idx > s:
+            e = idx
     # Restructured Text URLs
     if e > 3 and text[e-2:e] == '`_':
         e -= 2
-    # Remove trailing bracket if matched by leading bracket
-    if s > 0 and e < len(text) and text[s-1] in opening_brackets and text[e-1] == closing_bracket_map[text[s-1]]:
-        e -= 1
-    # Remove trailing quote if matched by leading quote
-    if s > 0 and e < len(text) and text[s-1] in '\'"' and text[e-1] == text[s-1]:
-        e -= 1
 
     return s, e
 

kitty/line.c

@@ -132,10 +132,11 @@ line_url_start_at(Line *self, index_type x) {
 }
 
 index_type
-line_url_end_at(Line *self, index_type x, bool check_short) {
+line_url_end_at(Line *self, index_type x, bool check_short, char_type sentinel) {
     index_type ans = x;
     if (x >= self->xnum || (check_short && self->xnum <= MIN_URL_LEN + 3)) return 0;
-    while (ans < self->xnum && is_url_char(self->cpu_cells[ans].ch)) ans++;
+    if (sentinel) { while (ans < self->xnum && self->cpu_cells[ans].ch != sentinel && is_url_char(self->cpu_cells[ans].ch)) ans++; }
+    else { while (ans < self->xnum && is_url_char(self->cpu_cells[ans].ch)) ans++; }
     if (ans) ans--;
     while (ans > x && can_strip_from_end_of_url(self->cpu_cells[ans].ch)) ans--;
     return ans;
@@ -148,9 +149,11 @@ url_start_at(Line *self, PyObject *x) {
 }
 
 static PyObject*
-url_end_at(Line *self, PyObject *x) {
+url_end_at(Line *self, PyObject *args) {
 #define url_end_at_doc "url_end_at(x) -> Return the end cell number for a URL containing x or 0 if not found"
-    return PyLong_FromUnsignedLong((unsigned long)line_url_end_at(self, PyLong_AsUnsignedLong(x), true));
+    unsigned int x, sentinel = 0;
+    if (!PyArg_ParseTuple(args, "I|I", &x, &sentinel)) return NULL;
+    return PyLong_FromUnsignedLong((unsigned long)line_url_end_at(self, x, true, sentinel));
 }
 
 // }}}
@@ -560,7 +563,7 @@ static PyMethodDef methods[] = {
     METHOD(is_continued, METH_NOARGS)
     METHOD(width, METH_O)
     METHOD(url_start_at, METH_O)
-    METHOD(url_end_at, METH_O)
+    METHOD(url_end_at, METH_VARARGS)
     METHOD(sprite_at, METH_O)
 
     {NULL}  /* Sentinel */

kitty/lineops.h

@@ -59,7 +59,7 @@ void line_set_char(Line *, unsigned int , uint32_t , unsigned int , Cursor *, bo
 void line_right_shift(Line *, unsigned int , unsigned int );
 void line_add_combining_char(Line *, uint32_t , unsigned int );
 index_type line_url_start_at(Line *self, index_type x);
-index_type line_url_end_at(Line *self, index_type x, bool);
+index_type line_url_end_at(Line *self, index_type x, bool, char_type);
 index_type line_as_ansi(Line *self, Py_UCS4 *buf, index_type buflen);
 unsigned int line_length(Line *self);
 size_t cell_as_unicode(CPUCell *cell, bool include_cc, Py_UCS4 *buf, char_type);

kitty/mouse.c

@@ -209,32 +209,57 @@ extend_selection(Window *w) {
 }
 
 static inline void
-extend_url(Screen *screen, Line *line, index_type *x, index_type *y) {
+extend_url(Screen *screen, Line *line, index_type *x, index_type *y, char_type sentinel) {
     unsigned int count = 0;
     while(count++ < 10) {
         if (*x != line->xnum - 1) break;
         line = screen_visual_line(screen, *y + 1);
         if (!line) break; // we deliberately allow non-continued lines as some programs, like mutt split URLs with newlines at line boundaries
-        index_type new_x = line_url_end_at(line, 0, false);
+        index_type new_x = line_url_end_at(line, 0, false, sentinel);
         if (!new_x) break;
         *y += 1; *x = new_x;
     }
 }
 
+static inline char_type
+get_url_sentinel(Line *line, index_type url_start) {
+    char_type before = 0, sentinel;
+    if (url_start > 0 && url_start < line->xnum) before = line->cpu_cells[url_start - 1].ch;
+    switch(before) {
+        case '"':
+        case '\'':
+        case '*':
+            sentinel = before; break;
+        case '(':
+            sentinel = ')'; break;
+        case '[':
+            sentinel = ']'; break;
+        case '{':
+            sentinel = '}'; break;
+        case '<':
+            sentinel = '>'; break;
+        default:
+            sentinel = 0; break;
+    }
+    return sentinel;
+}
+
 static inline void
 detect_url(Screen *screen, unsigned int x, unsigned int y) {
     bool has_url = false;
     index_type url_start, url_end = 0;
     Line *line = screen_visual_line(screen, y);
+    char_type sentinel;
     if (line) {
         url_start = line_url_start_at(line, x);
-        if (url_start < line->xnum) url_end = line_url_end_at(line, x, true);
+        sentinel = get_url_sentinel(line, url_start);
+        if (url_start < line->xnum) url_end = line_url_end_at(line, x, true, sentinel);
         has_url = url_end > url_start;
     }
     if (has_url) {
         mouse_cursor_shape = HAND;
         index_type y_extended = y;
-        extend_url(screen, line, &url_end, &y_extended);
+        extend_url(screen, line, &url_end, &y_extended, sentinel);
         screen_mark_url(screen, url_start, y, url_end, y_extended);
     } else {
         mouse_cursor_shape = BEAM;

kitty_tests/hints.py

@@ -29,3 +29,4 @@ class TestHints(BaseTest):
         t(u + '\nxxx', u + 'xxx', len(u))
         t('link:{}[xxx]'.format(u), u)
         t('`xyz <{}>`_.'.format(u), u)
+        t('<a href="{}">moo'.format(u), u)