Rework handling of PDF names for speed and memory.

Currently, every PDF name is allocated in a pdf_obj structure, and
comparisons are done using strcmp. Given that we can predict most
of the PDF names we'll use in a given file, this seems wasteful.

The pdf_obj type is opaque outside the pdf-object.c file, so we can
abuse it slightly without anyone outside knowing.

We collect a sorted list of names used in PDF (resources/pdf/names.txt),
and we add a utility (namedump) that preprocesses this into 2 header
files.

The first (include/mupdf/pdf/pdf-names-table.h, included as part of
include/mupdf/pdf/object.h), defines a set of "PDF_NAME_xxxx"
entries. These are pdf_obj *'s that callers can use to mean "A PDF
object that means literal name 'xxxx'"

The second (source/pdf/pdf-name-impl.h) is a C array of names.

We therefore update the code so that rather than passing "xxxx" to
functions (such as pdf_dict_gets(...)) we now pass PDF_NAME_xxxx (to
pdf_dict_get(...)). This is a fairly natural (if widespread) change.

The pdf_dict_getp (and sibling) functions that take a path (e.g.
"foo/bar/baz") are therefore supplemented with equivalents that
take a list (pdf_dict_getl(... , PDF_NAME_foo, PDF_NAME_bar,
PDF_NAME_baz, NULL)).

The actual implementation of this relies on the fact that small
pointer values are never valid values. For a given pdf_obj *p,
if NULL < (intptr_t)p < PDF_NAME__LIMIT then p is a literal
entry in the name table.

This enables us to do fast pointer compares and to skip expensive
strcmps.

Also, bring "null", "true" and "false" into the same style as PDF names.

Rather than using full pdf_obj structures for null/true/false, use
special pointer values just above the PDF_NAME_ table. This saves
memory and makes comparisons easier.

1 files changed, 37 insertions, 39 deletions

diff --git a/source/pdf/pdf-stream.c b/source/pdf/pdf-stream.c
index 15fdd29a..ad1e5e2c 100644
--- a/source/pdf/pdf-stream.c
+++ b/source/pdf/pdf-stream.c
@@ -26,10 +26,10 @@ pdf_stream_has_crypt(fz_context *ctx, pdf_obj *stm)
 	pdf_obj *obj;
 	int i;
 
-	filters = pdf_dict_getsa(ctx, stm, "Filter", "F");
+	filters = pdf_dict_geta(ctx, stm, PDF_NAME_Filter, PDF_NAME_F);
 	if (filters)
 	{
-		if (!strcmp(pdf_to_name(ctx, filters), "Crypt"))
+		if (pdf_name_eq(ctx, filters, PDF_NAME_Crypt))
 			return 1;
 		if (pdf_is_array(ctx, filters))
 		{
@@ -37,7 +37,7 @@ pdf_stream_has_crypt(fz_context *ctx, pdf_obj *stm)
 			for (i = 0; i < n; i++)
 			{
 				obj = pdf_array_get(ctx, filters, i);
-				if (!strcmp(pdf_to_name(ctx, obj), "Crypt"))
+				if (pdf_name_eq(ctx, obj, PDF_NAME_Crypt))
 					return 1;
 			}
 		}
@@ -82,31 +82,29 @@ pdf_load_jbig2_globals(fz_context *ctx, pdf_document *doc, pdf_obj *dict)
 static fz_stream *
 build_filter(fz_context *ctx, fz_stream *chain, pdf_document *doc, pdf_obj *f, pdf_obj *p, int num, int gen, fz_compression_params *params)
 {
-	char *s = pdf_to_name(ctx, f);
-
-	int predictor = pdf_to_int(ctx, pdf_dict_gets(ctx, p, "Predictor"));
-	pdf_obj *columns_obj = pdf_dict_gets(ctx, p, "Columns");
+	int predictor = pdf_to_int(ctx, pdf_dict_get(ctx, p, PDF_NAME_Predictor));
+	pdf_obj *columns_obj = pdf_dict_get(ctx, p, PDF_NAME_Columns);
 	int columns = pdf_to_int(ctx, columns_obj);
-	int colors = pdf_to_int(ctx, pdf_dict_gets(ctx, p, "Colors"));
-	int bpc = pdf_to_int(ctx, pdf_dict_gets(ctx, p, "BitsPerComponent"));
+	int colors = pdf_to_int(ctx, pdf_dict_get(ctx, p, PDF_NAME_Colors));
+	int bpc = pdf_to_int(ctx, pdf_dict_get(ctx, p, PDF_NAME_BitsPerComponent));
 
 	if (params)
 		params->type = FZ_IMAGE_RAW;
 
-	if (!strcmp(s, "ASCIIHexDecode") || !strcmp(s, "AHx"))
+	if (pdf_name_eq(ctx, f, PDF_NAME_ASCIIHexDecode) || pdf_name_eq(ctx, f, PDF_NAME_AHx))
 		return fz_open_ahxd(ctx, chain);
 
-	else if (!strcmp(s, "ASCII85Decode") || !strcmp(s, "A85"))
+	else if (pdf_name_eq(ctx, f, PDF_NAME_ASCII85Decode) || pdf_name_eq(ctx, f, PDF_NAME_A85))
 		return fz_open_a85d(ctx, chain);
 
-	else if (!strcmp(s, "CCITTFaxDecode") || !strcmp(s, "CCF"))
+	else if (pdf_name_eq(ctx, f, PDF_NAME_CCITTFaxDecode) || pdf_name_eq(ctx, f, PDF_NAME_CCF))
 	{
-		pdf_obj *k = pdf_dict_gets(ctx, p, "K");
-		pdf_obj *eol = pdf_dict_gets(ctx, p, "EndOfLine");
-		pdf_obj *eba = pdf_dict_gets(ctx, p, "EncodedByteAlign");
-		pdf_obj *rows = pdf_dict_gets(ctx, p, "Rows");
-		pdf_obj *eob = pdf_dict_gets(ctx, p, "EndOfBlock");
-		pdf_obj *bi1 = pdf_dict_gets(ctx, p, "BlackIs1");
+		pdf_obj *k = pdf_dict_get(ctx, p, PDF_NAME_K);
+		pdf_obj *eol = pdf_dict_get(ctx, p, PDF_NAME_EndOfLine);
+		pdf_obj *eba = pdf_dict_get(ctx, p, PDF_NAME_EncodedByteAlign);
+		pdf_obj *rows = pdf_dict_get(ctx, p, PDF_NAME_Rows);
+		pdf_obj *eob = pdf_dict_get(ctx, p, PDF_NAME_EndOfBlock);
+		pdf_obj *bi1 = pdf_dict_get(ctx, p, PDF_NAME_BlackIs1);
 		if (params)
 		{
 			/* We will shortstop here */
@@ -130,9 +128,9 @@ build_filter(fz_context *ctx, fz_stream *chain, pdf_document *doc, pdf_obj *f, p
 				bi1 ? pdf_to_bool(ctx, bi1) : 0);
 	}
 
-	else if (!strcmp(s, "DCTDecode") || !strcmp(s, "DCT"))
+	else if (pdf_name_eq(ctx, f, PDF_NAME_DCTDecode) || pdf_name_eq(ctx, f, PDF_NAME_DCT))
 	{
-		pdf_obj *ct = pdf_dict_gets(ctx, p, "ColorTransform");
+		pdf_obj *ct = pdf_dict_get(ctx, p, PDF_NAME_ColorTransform);
 		if (params)
 		{
 			/* We will shortstop here */
@@ -143,7 +141,7 @@ build_filter(fz_context *ctx, fz_stream *chain, pdf_document *doc, pdf_obj *f, p
 		return fz_open_dctd(ctx, chain, ct ? pdf_to_int(ctx, ct) : -1, 0, NULL);
 	}
 
-	else if (!strcmp(s, "RunLengthDecode") || !strcmp(s, "RL"))
+	else if (pdf_name_eq(ctx, f, PDF_NAME_RunLengthDecode) || pdf_name_eq(ctx, f, PDF_NAME_RL))
 	{
 		if (params)
 		{
@@ -153,7 +151,7 @@ build_filter(fz_context *ctx, fz_stream *chain, pdf_document *doc, pdf_obj *f, p
 		}
 		return fz_open_rld(ctx, chain);
 	}
-	else if (!strcmp(s, "FlateDecode") || !strcmp(s, "Fl"))
+	else if (pdf_name_eq(ctx, f, PDF_NAME_FlateDecode) || pdf_name_eq(ctx, f, PDF_NAME_Fl))
 	{
 		if (params)
 		{
@@ -171,9 +169,9 @@ build_filter(fz_context *ctx, fz_stream *chain, pdf_document *doc, pdf_obj *f, p
 		return chain;
 	}
 
-	else if (!strcmp(s, "LZWDecode") || !strcmp(s, "LZW"))
+	else if (pdf_name_eq(ctx, f, PDF_NAME_LZWDecode) || pdf_name_eq(ctx, f, PDF_NAME_LZW))
 	{
-		pdf_obj *ec = pdf_dict_gets(ctx, p, "EarlyChange");
+		pdf_obj *ec = pdf_dict_get(ctx, p, PDF_NAME_EarlyChange);
 		if (params)
 		{
 			/* We will shortstop here */
@@ -191,20 +189,20 @@ build_filter(fz_context *ctx, fz_stream *chain, pdf_document *doc, pdf_obj *f, p
 		return chain;
 	}
 
-	else if (!strcmp(s, "JBIG2Decode"))
+	else if (pdf_name_eq(ctx, f, PDF_NAME_JBIG2Decode))
 	{
 		fz_jbig2_globals *globals = NULL;
-		pdf_obj *obj = pdf_dict_gets(ctx, p, "JBIG2Globals");
+		pdf_obj *obj = pdf_dict_get(ctx, p, PDF_NAME_JBIG2Globals);
 		if (pdf_is_indirect(ctx, obj))
 			globals = pdf_load_jbig2_globals(ctx, doc, obj);
 		/* fz_open_jbig2d takes possession of globals */
 		return fz_open_jbig2d(ctx, chain, globals);
 	}
 
-	else if (!strcmp(s, "JPXDecode"))
+	else if (pdf_name_eq(ctx, f, PDF_NAME_JPXDecode))
 		return chain; /* JPX decoding is special cased in the image loading code */
 
-	else if (!strcmp(s, "Crypt"))
+	else if (pdf_name_eq(ctx, f, PDF_NAME_Crypt))
 	{
 		pdf_obj *name;
 
@@ -214,14 +212,14 @@ build_filter(fz_context *ctx, fz_stream *chain, pdf_document *doc, pdf_obj *f, p
 			return chain;
 		}
 
-		name = pdf_dict_gets(ctx, p, "Name");
+		name = pdf_dict_get(ctx, p, PDF_NAME_Name);
 		if (pdf_is_name(ctx, name))
-			return pdf_open_crypt_with_filter(ctx, chain, doc->crypt, pdf_to_name(ctx, name), num, gen);
+			return pdf_open_crypt_with_filter(ctx, chain, doc->crypt, name, num, gen);
 
 		return chain;
 	}
 
-	fz_warn(ctx, "unknown filter name (%s)", s);
+	fz_warn(ctx, "unknown filter name (%s)", pdf_to_name(ctx, f));
 	return chain;
 }
 
@@ -284,7 +282,7 @@ pdf_open_raw_filter(fz_context *ctx, fz_stream *chain, pdf_document *doc, pdf_ob
 	/* don't close chain when we close this filter */
 	fz_keep_stream(ctx, chain);
 
-	len = pdf_to_int(ctx, pdf_dict_gets(ctx, stmobj, "Length"));
+	len = pdf_to_int(ctx, pdf_dict_get(ctx, stmobj, PDF_NAME_Length));
 	chain = fz_open_null(ctx, chain, len, offset);
 
 	hascrypt = pdf_stream_has_crypt(ctx, stmobj);
@@ -304,8 +302,8 @@ pdf_open_filter(fz_context *ctx, pdf_document *doc, fz_stream *chain, pdf_obj *s
 	pdf_obj *filters;
 	pdf_obj *params;
 
-	filters = pdf_dict_getsa(ctx, stmobj, "Filter", "F");
-	params = pdf_dict_getsa(ctx, stmobj, "DecodeParms", "DP");
+	filters = pdf_dict_geta(ctx, stmobj, PDF_NAME_Filter, PDF_NAME_F);
+	params = pdf_dict_geta(ctx, stmobj, PDF_NAME_DecodeParms, PDF_NAME_DP);
 
 	chain = pdf_open_raw_filter(ctx, chain, doc, stmobj, num, num, gen, offset);
 
@@ -345,8 +343,8 @@ pdf_open_inline_stream(fz_context *ctx, pdf_document *doc, pdf_obj *stmobj, int
 	pdf_obj *filters;
 	pdf_obj *params;
 
-	filters = pdf_dict_getsa(ctx, stmobj, "Filter", "F");
-	params = pdf_dict_getsa(ctx, stmobj, "DecodeParms", "DP");
+	filters = pdf_dict_geta(ctx, stmobj, PDF_NAME_Filter, PDF_NAME_F);
+	params = pdf_dict_geta(ctx, stmobj, PDF_NAME_DecodeParms, PDF_NAME_DP);
 
 	/* don't close chain when we close this filter */
 	fz_keep_stream(ctx, chain);
@@ -470,7 +468,7 @@ pdf_load_raw_renumbered_stream(fz_context *ctx, pdf_document *doc, int num, int
 
 	dict = pdf_load_object(ctx, doc, num, gen);
 
-	len = pdf_to_int(ctx, pdf_dict_gets(ctx, dict, "Length"));
+	len = pdf_to_int(ctx, pdf_dict_get(ctx, dict, PDF_NAME_Length));
 
 	pdf_drop_obj(ctx, dict);
 
@@ -517,8 +515,8 @@ pdf_load_image_stream(fz_context *ctx, pdf_document *doc, int num, int gen, int
 
 	dict = pdf_load_object(ctx, doc, num, gen);
 
-	len = pdf_to_int(ctx, pdf_dict_gets(ctx, dict, "Length"));
-	obj = pdf_dict_gets(ctx, dict, "Filter");
+	len = pdf_to_int(ctx, pdf_dict_get(ctx, dict, PDF_NAME_Length));
+	obj = pdf_dict_get(ctx, dict, PDF_NAME_Filter);
 	len = pdf_guess_filter_length(len, pdf_to_name(ctx, obj));
 	n = pdf_array_len(ctx, obj);
 	for (i = 0; i < n; i++)