diff options
| -rw-r--r-- | include/mupdf/pdf/object.h | 1 | ||||
| -rw-r--r-- | source/pdf/pdf-parse.c | 40 |
2 files changed, 41 insertions, 0 deletions
diff --git a/include/mupdf/pdf/object.h b/include/mupdf/pdf/object.h index a1825eb2..21ed8595 100644 --- a/include/mupdf/pdf/object.h +++ b/include/mupdf/pdf/object.h @@ -17,6 +17,7 @@ pdf_obj *pdf_new_int(fz_context *ctx, pdf_document *doc, int64_t i); pdf_obj *pdf_new_real(fz_context *ctx, pdf_document *doc, float f); pdf_obj *pdf_new_name(fz_context *ctx, pdf_document *doc, const char *str); pdf_obj *pdf_new_string(fz_context *ctx, pdf_document *doc, const char *str, size_t len); +pdf_obj *pdf_new_text_string(fz_context *ctx, pdf_document *doc, const char *s); pdf_obj *pdf_new_indirect(fz_context *ctx, pdf_document *doc, int num, int gen); pdf_obj *pdf_new_array(fz_context *ctx, pdf_document *doc, int initialcap); pdf_obj *pdf_new_dict(fz_context *ctx, pdf_document *doc, int initialcap); diff --git a/source/pdf/pdf-parse.c b/source/pdf/pdf-parse.c index 797d4c56..451d9e1d 100644 --- a/source/pdf/pdf-parse.c +++ b/source/pdf/pdf-parse.c @@ -317,6 +317,46 @@ pdf_to_utf8_name(fz_context *ctx, pdf_document *doc, pdf_obj *src) return dst; } +static pdf_obj * +pdf_new_text_string_utf16be(fz_context *ctx, pdf_document *doc, const char *s) +{ + int c, i = 0, n = fz_utflen(s); + unsigned char *p = fz_malloc(ctx, n * 2 + 2); + pdf_obj *obj; + p[i++] = 254; + p[i++] = 255; + while (*s) + { + s += fz_chartorune(&c, s); + p[i++] = (c>>8) & 0xff; + p[i++] = (c) & 0xff; + } + fz_try(ctx) + obj = pdf_new_string(ctx, doc, (char*)p, i); + fz_always(ctx) + fz_free(ctx, p); + fz_catch(ctx) + fz_rethrow(ctx); + return obj; +} + +/* + * Create a PDF 'text string' by encoding input string as either ASCII or UTF-16BE. + * In theory, we could also use PDFDocEncoding. + */ +pdf_obj * +pdf_new_text_string(fz_context *ctx, pdf_document *doc, const char *s) +{ + int i = 0; + while (s[i] != 0) + { + if (((unsigned char)s[i]) >= 128) + return pdf_new_text_string_utf16be(ctx, doc, s); + ++i; + } + return pdf_new_string(ctx, doc, s, i); +} + pdf_obj * pdf_parse_array(fz_context *ctx, pdf_document *doc, fz_stream *file, pdf_lexbuf *buf) { |