diff options
| author | Tor Andersson <tor.andersson@artifex.com> | 2016-08-25 12:44:13 +0200 |
|---|---|---|
| committer | Tor Andersson <tor.andersson@artifex.com> | 2016-08-30 16:55:25 +0200 |
| commit | 85ee87997e3ee4eb579084f92d109b9b78dcf9c7 (patch) | |
| tree | 25bf20b8d2d5acf7a4ff61c52fdd41528f48da6d | |
| parent | bf32163059811c822c46e2e17142f517cf9a0bac (diff) | |
| download | mupdf-85ee87997e3ee4eb579084f92d109b9b78dcf9c7.tar.xz | |
Use U+FFFD instead of '?' for bad encodings in text extraction.
| -rw-r--r-- | source/fitz/stext-search.c | 2 | ||||
| -rw-r--r-- | source/fitz/util.c | 2 | ||||
| -rw-r--r-- | source/pdf/pdf-op-run.c | 2 | ||||
| -rw-r--r-- | source/pdf/pdf-unicode.c | 2 | ||||
| -rw-r--r-- | source/xps/xps-glyphs.c | 2 |
5 files changed, 5 insertions, 5 deletions
diff --git a/source/fitz/stext-search.c b/source/fitz/stext-search.c index 4afbe6b7..2246f627 100644 --- a/source/fitz/stext-search.c +++ b/source/fitz/stext-search.c @@ -259,7 +259,7 @@ fz_copy_selection(fz_context *ctx, fz_stext_page *page, fz_rect rect) fz_stext_char_bbox(ctx, &hitbox, span, i); c = span->text[i].c; if (c < 32) - c = '?'; + c = 0xFFFD; if (hitbox.x1 >= x0 && hitbox.x0 <= x1 && hitbox.y1 >= y0 && hitbox.y0 <= y1) { fz_write_buffer_rune(ctx, buffer, c); diff --git a/source/fitz/util.c b/source/fitz/util.c index fc49d84f..4c7b3232 100644 --- a/source/fitz/util.c +++ b/source/fitz/util.c @@ -447,7 +447,7 @@ fz_new_buffer_from_stext_page(fz_context *ctx, fz_stext_page *text, const fz_rec fz_stext_char_bbox(ctx, &hitbox, span, i); c = span->text[i].c; if (c < 32) - c = '?'; + c = 0xFFFD; if (hitbox.x1 >= x0 && hitbox.x0 <= x1 && hitbox.y1 >= y0 && hitbox.y0 <= y1) { saw_text = 1; diff --git a/source/pdf/pdf-op-run.c b/source/pdf/pdf-op-run.c index 96084cdf..aad0aebf 100644 --- a/source/pdf/pdf-op-run.c +++ b/source/pdf/pdf-op-run.c @@ -902,7 +902,7 @@ pdf_show_char(fz_context *ctx, pdf_run_processor *pr, int cid) } if (ucslen == 0 || (ucslen == 1 && ucsbuf[0] == 0)) { - ucsbuf[0] = '?'; + ucsbuf[0] = 0xFFFD; ucslen = 1; } diff --git a/source/pdf/pdf-unicode.c b/source/pdf/pdf-unicode.c index ca84341d..65bda460 100644 --- a/source/pdf/pdf-unicode.c +++ b/source/pdf/pdf-unicode.c @@ -96,7 +96,7 @@ pdf_load_to_unicode(fz_context *ctx, pdf_document *doc, pdf_font_desc *font, if (strings[cpt]) font->cid_to_ucs[cpt] = pdf_lookup_agl(strings[cpt]); else - font->cid_to_ucs[cpt] = '?'; + font->cid_to_ucs[cpt] = 0xFFFD; /* replacement character */ } } diff --git a/source/xps/xps-glyphs.c b/source/xps/xps-glyphs.c index a5d3f196..407953e8 100644 --- a/source/xps/xps-glyphs.c +++ b/source/xps/xps-glyphs.c @@ -389,7 +389,7 @@ xps_parse_glyphs_imp(fz_context *ctx, xps_document *doc, const fz_matrix *ctm, while ((us && un > 0) || (is && *is)) { - int char_code = '?'; + int char_code = 0xFFFD; int code_count = 1; int glyph_count = 1; |