fitz/: Fix pymupdf#2238 - use 'overlap' rather than 'contains' when e…

…xtracting text.

We now include chars that overlap with the clipbox, instead of only those that
are entirely contained within the clipbox.

fitz/fitz.i

@@ -11528,7 +11528,7 @@ struct TextPage {
                             fz_rect linerect = fz_empty_rect;
                             for (ch = line->first_char; ch; ch = ch->next) {
                                 fz_rect cbbox = JM_char_bbox(gctx, line, ch);
-                                if (!fz_contains_rect(tp_rect, cbbox) &&
+                                if (!JM_rects_overlap(tp_rect, cbbox) &&
                                     !fz_is_infinite_rect(tp_rect)) {
                                     continue;
                                 }
@@ -11542,7 +11542,7 @@ struct TextPage {
                             blockrect = fz_union_rect(blockrect, linerect);
                         }
                         text = JM_EscapeStrFromBuffer(gctx, res);
-                    } else if (fz_contains_rect(tp_rect, block->bbox) || fz_is_infinite_rect(tp_rect)) {
+                    } else if (JM_rects_overlap(tp_rect, block->bbox) || fz_is_infinite_rect(tp_rect)) {
                         fz_image *img = block->u.i.image;
                         fz_colorspace *cs = img->colorspace;
                         text = PyUnicode_FromFormat("<image: %s, width: %d, height: %d, bpc: %d>", fz_colorspace_name(gctx, cs), img->w, img->h, img->bpc);
@@ -11610,7 +11610,7 @@ struct TextPage {
                         buflen = 0;                       // reset char counter
                         for (ch = line->first_char; ch; ch = ch->next) {
                             fz_rect cbbox = JM_char_bbox(gctx, line, ch);
-                            if (!fz_contains_rect(tp_rect, cbbox) &&
+                            if (!JM_rects_overlap(tp_rect, cbbox) &&
                                 !fz_is_infinite_rect(tp_rect)) {
                                 continue;
                             }

fitz/helper-other.i

@@ -1295,7 +1295,11 @@ fz_archive *JM_archive_from_py(fz_context *ctx, fz_archive *arch, PyObject *path
 }
 
 
-
+int JM_rects_overlap(const fz_rect a, const fz_rect b)
+{
+    fz_rect c = fz_intersect_rect(a, b);
+    return !fz_is_empty_rect(c);
+}
 
 //-----------------------------------------------------------------------------
 // dummy structure for various tools and utilities

fitz/helper-stext.i

@@ -228,7 +228,7 @@ JM_new_buffer_from_stext_page(fz_context *ctx, fz_stext_page *page)
             if (block->type == FZ_STEXT_BLOCK_TEXT) {
                 for (line = block->u.t.first_line; line; line = line->next) {
                     for (ch = line->first_char; ch; ch = ch->next) {
-                        if (!fz_contains_rect(rect, JM_char_bbox(ctx, line, ch)) &&
+                        if (!JM_rects_overlap(rect, JM_char_bbox(ctx, line, ch)) &&
                             !fz_is_infinite_rect(rect)) {
                             continue;
                         }
@@ -420,7 +420,7 @@ JM_search_stext_page(fz_context *ctx, fz_stext_page *page, const char *needle)
             for (line = block->u.t.first_line; line; line = line->next) {
                 for (ch = line->first_char; ch; ch = ch->next) {
                     if (!fz_is_infinite_rect(rect) &&
-                        !fz_contains_rect(rect, JM_char_bbox(ctx, line, ch))) {
+                        !JM_rects_overlap(rect, JM_char_bbox(ctx, line, ch))) {
                             goto next_char;
                         }
 try_new_match:
@@ -482,7 +482,7 @@ JM_print_stext_page_as_text(fz_context *ctx, fz_output *out, fz_stext_page *page
                 for (ch = line->first_char; ch; ch = ch->next) {
                     chbbox = JM_char_bbox(ctx, line, ch);
                     if (fz_is_infinite_rect(rect) ||
-                        fz_contains_rect(rect, chbbox)) {
+                        JM_rects_overlap(rect, chbbox)) {
                         last_char = ch->c;
                         n = fz_runetochar(utf, ch->c);
                         for (i = 0; i < n; i++) {
@@ -571,7 +571,7 @@ JM_make_spanlist(fz_context *ctx, PyObject *line_dict,
 
     for (ch = line->first_char; ch; ch = ch->next) {
         fz_rect r = JM_char_bbox(ctx, line, ch);
-        if (!fz_contains_rect(tp_rect, r) &&
+        if (!JM_rects_overlap(tp_rect, r) &&
             !fz_is_infinite_rect(tp_rect)) {
             continue;
         }
@@ -769,7 +769,7 @@ void JM_make_textpage_dict(fz_context *ctx, fz_stext_page *tp, PyObject *page_di
     int block_n = -1;
     for (block = tp->first_block; block; block = block->next) {
         block_n++;
-        if (!fz_contains_rect(tp_rect, block->bbox) &&
+        if (!JM_rects_overlap(tp_rect, block->bbox) &&
             !fz_is_infinite_rect(tp_rect) &&
             block->type == FZ_STEXT_BLOCK_IMAGE) {
             continue;
@@ -817,7 +817,7 @@ JM_copy_rectangle(fz_context *ctx, fz_stext_page *page, fz_rect area)
 				int line_had_text = 0;
 				for (ch = line->first_char; ch; ch = ch->next) {
 					fz_rect r = JM_char_bbox(ctx, line, ch);
-					if (fz_contains_rect(area, r)) {
+					if (JM_rects_overlap(area, r)) {
 						line_had_text = 1;
 						if (need_new_line) {
 							fz_append_string(ctx, buffer, "\n");