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, 13 insertions, 13 deletions

diff --git a/source/pdf/pdf-function.c b/source/pdf/pdf-function.c
index 7ead09ca..1969cbba 100644
--- a/source/pdf/pdf-function.c
+++ b/source/pdf/pdf-function.c
@@ -919,7 +919,7 @@ load_sample_func(fz_context *ctx, pdf_document *doc, pdf_function *func, pdf_obj
 
 	func->u.sa.samples = NULL;
 
-	obj = pdf_dict_gets(ctx, dict, "Size");
+	obj = pdf_dict_get(ctx, dict, PDF_NAME_Size);
 	if (pdf_array_len(ctx, obj) < func->base.m)
 		fz_throw(ctx, FZ_ERROR_GENERIC, "too few sample function dimension sizes");
 	if (pdf_array_len(ctx, obj) > func->base.m)
@@ -934,7 +934,7 @@ load_sample_func(fz_context *ctx, pdf_document *doc, pdf_function *func, pdf_obj
 		}
 	}
 
-	obj = pdf_dict_gets(ctx, dict, "BitsPerSample");
+	obj = pdf_dict_get(ctx, dict, PDF_NAME_BitsPerSample);
 	func->u.sa.bps = bps = pdf_to_int(ctx, obj);
 
 	for (i = 0; i < func->base.m; i++)
@@ -942,7 +942,7 @@ load_sample_func(fz_context *ctx, pdf_document *doc, pdf_function *func, pdf_obj
 		func->u.sa.encode[i][0] = 0;
 		func->u.sa.encode[i][1] = func->u.sa.size[i] - 1;
 	}
-	obj = pdf_dict_gets(ctx, dict, "Encode");
+	obj = pdf_dict_get(ctx, dict, PDF_NAME_Encode);
 	if (pdf_is_array(ctx, obj))
 	{
 		int ranges = fz_mini(func->base.m, pdf_array_len(ctx, obj) / 2);
@@ -962,7 +962,7 @@ load_sample_func(fz_context *ctx, pdf_document *doc, pdf_function *func, pdf_obj
 		func->u.sa.decode[i][1] = func->range[i][1];
 	}
 
-	obj = pdf_dict_gets(ctx, dict, "Decode");
+	obj = pdf_dict_get(ctx, dict, PDF_NAME_Decode);
 	if (pdf_is_array(ctx, obj))
 	{
 		int ranges = fz_mini(func->base.n, pdf_array_len(ctx, obj) / 2);
@@ -1141,7 +1141,7 @@ load_exponential_func(fz_context *ctx, pdf_document *doc, pdf_function *func, pd
 		fz_warn(ctx, "exponential functions have at most one input");
 	func->base.m = 1;
 
-	obj = pdf_dict_gets(ctx, dict, "N");
+	obj = pdf_dict_get(ctx, dict, PDF_NAME_N);
 	func->u.e.n = pdf_to_real(ctx, obj);
 
 	/* See exponential functions (PDF 1.7 section 3.9.2) */
@@ -1167,7 +1167,7 @@ load_exponential_func(fz_context *ctx, pdf_document *doc, pdf_function *func, pd
 		func->u.e.c1[i] = 1;
 	}
 
-	obj = pdf_dict_gets(ctx, dict, "C0");
+	obj = pdf_dict_get(ctx, dict, PDF_NAME_C0);
 	if (pdf_is_array(ctx, obj))
 	{
 		int ranges = fz_mini(func->base.n, pdf_array_len(ctx, obj));
@@ -1178,7 +1178,7 @@ load_exponential_func(fz_context *ctx, pdf_document *doc, pdf_function *func, pd
 			func->u.e.c0[i] = pdf_to_real(ctx, pdf_array_get(ctx, obj, i));
 	}
 
-	obj = pdf_dict_gets(ctx, dict, "C1");
+	obj = pdf_dict_get(ctx, dict, PDF_NAME_C1);
 	if (pdf_is_array(ctx, obj))
 	{
 		int ranges = fz_mini(func->base.n, pdf_array_len(ctx, obj));
@@ -1232,7 +1232,7 @@ load_stitching_func(fz_context *ctx, pdf_document *doc, pdf_function *func, pdf_
 		fz_warn(ctx, "stitching functions have at most one input");
 	func->base.m = 1;
 
-	obj = pdf_dict_gets(ctx, dict, "Functions");
+	obj = pdf_dict_get(ctx, dict, PDF_NAME_Functions);
 	if (!pdf_is_array(ctx, obj))
 		fz_throw(ctx, FZ_ERROR_GENERIC, "stitching function has no input functions");
 
@@ -1269,7 +1269,7 @@ load_stitching_func(fz_context *ctx, pdf_document *doc, pdf_function *func, pdf_
 		fz_rethrow(ctx);
 	}
 
-	obj = pdf_dict_gets(ctx, dict, "Bounds");
+	obj = pdf_dict_get(ctx, dict, PDF_NAME_Bounds);
 	if (!pdf_is_array(ctx, obj))
 		fz_throw(ctx, FZ_ERROR_GENERIC, "stitching function has no bounds");
 	{
@@ -1297,7 +1297,7 @@ load_stitching_func(fz_context *ctx, pdf_document *doc, pdf_function *func, pdf_
 		func->u.st.encode[i * 2 + 1] = 0;
 	}
 
-	obj = pdf_dict_gets(ctx, dict, "Encode");
+	obj = pdf_dict_get(ctx, dict, PDF_NAME_Encode);
 	if (pdf_is_array(ctx, obj))
 	{
 		int ranges = fz_mini(k, pdf_array_len(ctx, obj) / 2);
@@ -1645,11 +1645,11 @@ pdf_load_function(fz_context *ctx, pdf_document *doc, pdf_obj *dict, int in, int
 	func->base.debug = pdf_debug_function;
 #endif
 
-	obj = pdf_dict_gets(ctx, dict, "FunctionType");
+	obj = pdf_dict_get(ctx, dict, PDF_NAME_FunctionType);
 	func->type = pdf_to_int(ctx, obj);
 
 	/* required for all */
-	obj = pdf_dict_gets(ctx, dict, "Domain");
+	obj = pdf_dict_get(ctx, dict, PDF_NAME_Domain);
 	func->base.m = fz_clampi(pdf_array_len(ctx, obj) / 2, 1, FZ_FN_MAXM);
 	for (i = 0; i < func->base.m; i++)
 	{
@@ -1658,7 +1658,7 @@ pdf_load_function(fz_context *ctx, pdf_document *doc, pdf_obj *dict, int in, int
 	}
 
 	/* required for type0 and type4, optional otherwise */
-	obj = pdf_dict_gets(ctx, dict, "Range");
+	obj = pdf_dict_get(ctx, dict, PDF_NAME_Range);
 	if (pdf_is_array(ctx, obj))
 	{
 		func->has_range = 1;