Avoid some UNSAFE_BUFFER noise in fpdf_parser_utility.h.
This is a widely-included header doing C-Style array indexing.
Change-Id: I0e46d311b3b6765804ebc5db834f34c9c3c87bf2
Reviewed-on: https://pdfium-review.googlesource.com/c/pdfium/+/116930
Commit-Queue: Tom Sepez <tsepez@chromium.org>
Reviewed-by: Lei Zhang <thestig@chromium.org>
Reviewed-by: Thomas Sepez <tsepez@google.com>
diff --git a/core/fpdfapi/parser/fpdf_parser_utility.h b/core/fpdfapi/parser/fpdf_parser_utility.h
index 509661a..1ec02d9 100644
--- a/core/fpdfapi/parser/fpdf_parser_utility.h
+++ b/core/fpdfapi/parser/fpdf_parser_utility.h
@@ -8,10 +8,12 @@
#define CORE_FPDFAPI_PARSER_FPDF_PARSER_UTILITY_H_
#include <iosfwd>
+#include <limits>
#include <optional>
#include <vector>
#include "core/fxcrt/bytestring.h"
+#include "core/fxcrt/compiler_specific.h"
#include "core/fxcrt/retain_ptr.h"
class CPDF_Array;
@@ -22,17 +24,30 @@
// Use the accessors below instead of directly accessing kPDFCharTypes.
extern const char kPDFCharTypes[256];
+inline uint8_t GetPDFCharTypeFromArray(uint8_t c) {
+ static_assert(std::numeric_limits<decltype(c)>::min() == 0);
+ static_assert(std::numeric_limits<decltype(c)>::max() <
+ std::size(kPDFCharTypes));
+
+ // SAFETY: previous static_asserts show table covers entire range
+ // of the index's type.
+ return UNSAFE_BUFFERS(kPDFCharTypes[c]);
+}
+
inline bool PDFCharIsWhitespace(uint8_t c) {
- return kPDFCharTypes[c] == 'W';
+ return GetPDFCharTypeFromArray(c) == 'W';
}
+
inline bool PDFCharIsNumeric(uint8_t c) {
- return kPDFCharTypes[c] == 'N';
+ return GetPDFCharTypeFromArray(c) == 'N';
}
+
inline bool PDFCharIsDelimiter(uint8_t c) {
- return kPDFCharTypes[c] == 'D';
+ return GetPDFCharTypeFromArray(c) == 'D';
}
+
inline bool PDFCharIsOther(uint8_t c) {
- return kPDFCharTypes[c] == 'R';
+ return GetPDFCharTypeFromArray(c) == 'R';
}
inline bool PDFCharIsLineEnding(uint8_t c) {
