Fix typo in text device where lines would group into blocks too eagerly.

The default page userspace transform changed to a top-down coordinate
space, and I forgot this detail when updating the text device branch.

Also remove the final block sorting pass to give preference to the original
PDF text order.

1 files changed, 7 insertions, 15 deletions

diff --git a/fitz/dev_text.c b/fitz/dev_text.c
index fb6ebdc6..5e7f8164 100644
--- a/fitz/dev_text.c
+++ b/fitz/dev_text.c
@@ -190,10 +190,12 @@ lookup_block_for_line(fz_context *ctx, fz_text_page *page, fz_text_line *line)
 	for (i = 0; i < page->len; i++)
 	{
 		fz_text_block *block = page->blocks + i;
-		int w = block->bbox.x1 - block->bbox.x0;
-		if (block->bbox.y0 - line->bbox.y1 < size * PARAGRAPH_DIST)
-			if (line->bbox.x0 < block->bbox.x1 && line->bbox.x1 > block->bbox.x0)
-				if (ABS(line->bbox.x0 - block->bbox.x0) < w / 4)
+		float w = block->bbox.x1 - block->bbox.x0;
+		float dx = line->bbox.x0 - block->bbox.x0;
+		float dy = line->bbox.y0 - block->bbox.y1;
+		if (dy > -size * 1.5f && dy < size * PARAGRAPH_DIST)
+			if (line->bbox.x0 <= block->bbox.x1 && line->bbox.x1 >= block->bbox.x0)
+				if (ABS(dx) < w / 2)
 					return block;
 	}
 
@@ -498,15 +500,6 @@ fz_text_ignore_text(fz_device *dev, fz_text *text, fz_matrix ctm)
 	fz_text_extract(dev->ctx, tdev, text, ctm, style);
 }
 
-static int cmp_block(const void *av, const void *bv)
-{
-	const fz_text_block *a = av;
-	const fz_text_block *b = bv;
-	int x = a->bbox.x0 - b->bbox.x0;
-	if (x) return x;
-	return -(a->bbox.y0 - b->bbox.y0);
-}
-
 static void
 fz_text_free_user(fz_device *dev)
 {
@@ -516,8 +509,7 @@ fz_text_free_user(fz_device *dev)
 	append_span(ctx, &tdev->cur_line, &tdev->cur_span);
 	insert_line(ctx, tdev->page, &tdev->cur_line);
 
-	qsort(tdev->page->blocks, tdev->page->len, sizeof *tdev->page->blocks, cmp_block);
-
+	/* TODO: smart sorting of blocks in reading order */
 	/* TODO: unicode NFC normalization */
 	/* TODO: bidi logical reordering */