OLD | NEW |
1 #ifndef EXPERIMENTAL_PDFVIEWER_PDFPARSER_NATIVE_SKPDFNATIVETOKENIZER_H_ | 1 #ifndef EXPERIMENTAL_PDFVIEWER_PDFPARSER_NATIVE_SKPDFNATIVETOKENIZER_H_ |
2 #define EXPERIMENTAL_PDFVIEWER_PDFPARSER_NATIVE_SKPDFNATIVETOKENIZER_H_ | 2 #define EXPERIMENTAL_PDFVIEWER_PDFPARSER_NATIVE_SKPDFNATIVETOKENIZER_H_ |
3 | 3 |
4 class SkPdfNativeTokenizer : public SkPdfPodofoTokenizer { | 4 #include "SkTDArray.h" |
| 5 #include "SkTDict.h" |
| 6 #include <math.h> |
| 7 #include <string.h> |
| 8 |
| 9 class SkPdfMapper; |
| 10 class SkPdfDictionary; |
| 11 |
| 12 // White Spaces |
| 13 #define kNUL_PdfWhiteSpace '\x00' |
| 14 #define kHT_PdfWhiteSpace '\x09' |
| 15 #define kLF_PdfWhiteSpace '\x0A' |
| 16 #define kFF_PdfWhiteSpace '\x0C' |
| 17 #define kCR_PdfWhiteSpace '\x0D' |
| 18 #define kSP_PdfWhiteSpace '\x20' |
| 19 |
| 20 // PdfDelimiters |
| 21 #define kOpenedRoundBracket_PdfDelimiter '(' |
| 22 #define kClosedRoundBracket_PdfDelimiter ')' |
| 23 #define kOpenedInequityBracket_PdfDelimiter '<' |
| 24 #define kClosedInequityBracket_PdfDelimiter '>' |
| 25 #define kOpenedSquareBracket_PdfDelimiter '[' |
| 26 #define kClosedSquareBracket_PdfDelimiter ']' |
| 27 #define kOpenedCurlyBracket_PdfDelimiter '{' |
| 28 #define kClosedCurlyBracket_PdfDelimiter '}' |
| 29 #define kNamed_PdfDelimiter '/' |
| 30 #define kComment_PdfDelimiter '%' |
| 31 |
| 32 #define kEscape_PdfSpecial '\\' |
| 33 #define kBackspace_PdfSpecial '\x08' |
| 34 |
| 35 // TODO(edisonn): what is the faster way for compiler/machine type to evaluate t
his expressions? |
| 36 // we should evaluate all options. might be even different from one machine to a
nother |
| 37 // 1) expand expression, let compiler optimize it |
| 38 // 2) binary search |
| 39 // 3) linear search in array |
| 40 // 4) vector (e.f. T type[256] .. return type[ch] ... |
| 41 // 5) manually build the expression with least number of operators, e.g. for con
secutive |
| 42 // chars, we can use an binary equal ignoring last bit |
| 43 #define isPdfWhiteSpace(ch) (((ch)==kNUL_PdfWhiteSpace)||((ch)==kHT_PdfWhiteSpac
e)||((ch)==kLF_PdfWhiteSpace)||((ch)==kFF_PdfWhiteSpace)||((ch)==kCR_PdfWhiteSpa
ce)||((ch)==kSP_PdfWhiteSpace)) |
| 44 |
| 45 #define isPdfEOL(ch) (((ch)==kLF_PdfWhiteSpace)||((ch)==kCR_PdfWhiteSpace)) |
| 46 |
| 47 |
| 48 #define isPdfDelimiter(ch) (((ch)==kOpenedRoundBracket_PdfDelimiter)||\ |
| 49 ((ch)==kClosedRoundBracket_PdfDelimiter)||\ |
| 50 ((ch)==kOpenedInequityBracket_PdfDelimiter)||\ |
| 51 ((ch)==kClosedInequityBracket_PdfDelimiter)||\ |
| 52 ((ch)==kOpenedSquareBracket_PdfDelimiter)||\ |
| 53 ((ch)==kClosedSquareBracket_PdfDelimiter)||\ |
| 54 ((ch)==kOpenedCurlyBracket_PdfDelimiter)||\ |
| 55 ((ch)==kClosedCurlyBracket_PdfDelimiter)||\ |
| 56 ((ch)==kNamed_PdfDelimiter)||\ |
| 57 ((ch)==kComment_PdfDelimiter)) |
| 58 |
| 59 #define isPdfWhiteSpaceOrPdfDelimiter(ch) (isPdfWhiteSpace(ch)||isPdfDelimiter(c
h)) |
| 60 |
| 61 #define isPdfDigit(ch) ((ch)>='0'&&(ch)<='9') |
| 62 #define isPdfNumeric(ch) (isPdfDigit(ch)||(ch)=='+'||(ch)=='-') |
| 63 |
| 64 unsigned char* skipPdfWhiteSpaces(unsigned char* buffer, size_t len); |
| 65 unsigned char* endOfPdfToken(unsigned char* start, size_t len); |
| 66 unsigned char* skipPdfComment(unsigned char* start, size_t len); |
| 67 |
| 68 // TODO(edisonn): typedef read and integer tyepes? make less readable... |
| 69 //typedef double SkPdfReal; |
| 70 //typedef int64_t SkPdfInteger; |
| 71 |
| 72 // an allocator only allocates memory, and it deletes it all when the allocator
is destroyed |
| 73 // this would allow us not to do any garbage collection while we parse or draw a
pdf, and defere it |
| 74 // while the user is looking at the image |
| 75 |
| 76 class SkPdfObject; |
| 77 |
| 78 class SkPdfAllocator { |
| 79 #define BUFFER_SIZE 1024 |
| 80 SkTDArray<SkPdfObject*> fHistory; |
| 81 SkTDArray<void*> fHandles; |
| 82 SkPdfObject* fCurrent; |
| 83 int fCurrentUsed; |
| 84 |
| 85 SkPdfObject* allocBlock(); |
| 86 |
5 public: | 87 public: |
6 SkPdfNativeTokenizer(); | 88 SkPdfAllocator() { |
| 89 fCurrent = allocBlock(); |
| 90 fCurrentUsed = 0; |
| 91 } |
| 92 |
| 93 ~SkPdfAllocator(); |
| 94 |
| 95 SkPdfObject* allocObject(); |
| 96 |
| 97 // TODO(edisonn): free this memory in destructor, track the usage? |
| 98 void* alloc(size_t bytes) { |
| 99 void* data = malloc(bytes); |
| 100 fHandles.push(data); |
| 101 return data; |
| 102 } |
| 103 }; |
| 104 |
| 105 unsigned char* nextObject(unsigned char* start, unsigned char* end, SkPdfObject*
token, SkPdfAllocator* allocator); |
| 106 |
| 107 enum SkPdfTokenType { |
| 108 kKeyword_TokenType, |
| 109 kObject_TokenType, |
| 110 }; |
| 111 |
| 112 struct PdfToken { |
| 113 const char* fKeyword; |
| 114 size_t fKeywordLength; |
| 115 SkPdfObject* fObject; |
| 116 SkPdfTokenType fType; |
| 117 |
| 118 PdfToken() : fKeyword(NULL), fKeywordLength(0), fObject(NULL) {} |
| 119 }; |
| 120 |
| 121 class SkPdfNativeTokenizer { |
| 122 public: |
| 123 SkPdfNativeTokenizer(SkPdfObject* objWithStream, const SkPdfMapper* mapper,
SkPdfAllocator* allocator); |
| 124 SkPdfNativeTokenizer(unsigned char* buffer, int len, const SkPdfMapper* mapp
er, SkPdfAllocator* allocator); |
| 125 |
7 virtual ~SkPdfNativeTokenizer(); | 126 virtual ~SkPdfNativeTokenizer(); |
| 127 |
| 128 bool readToken(PdfToken* token); |
| 129 bool readTokenCore(PdfToken* token); |
| 130 void PutBack(PdfToken token); |
| 131 |
| 132 private: |
| 133 const SkPdfMapper* fMapper; |
| 134 SkPdfAllocator* fAllocator; |
| 135 |
| 136 unsigned char* fUncompressedStreamStart; |
| 137 unsigned char* fUncompressedStream; |
| 138 unsigned char* fUncompressedStreamEnd; |
| 139 |
| 140 bool fEmpty; |
| 141 bool fHasPutBack; |
| 142 PdfToken fPutBack; |
8 }; | 143 }; |
9 | 144 |
10 #endif // EXPERIMENTAL_PDFVIEWER_PDFPARSER_NATIVE_SKPDFNATIVETOKENIZER_H_ | 145 #endif // EXPERIMENTAL_PDFVIEWER_PDFPARSER_NATIVE_SKPDFNATIVETOKENIZER_H_ |
OLD | NEW |