1.1 --- /dev/null Thu Jan 01 00:00:00 1970 +0000
1.2 +++ b/intl/icu/source/i18n/uspoof_conf.h Wed Dec 31 06:09:35 2014 +0100
1.3 @@ -0,0 +1,132 @@
1.4 +/*
1.5 +******************************************************************************
1.6 +*
1.7 +* Copyright (C) 2008-2011, International Business Machines
1.8 +* Corporation and others. All Rights Reserved.
1.9 +*
1.10 +******************************************************************************
1.11 +* file name: uspoof_conf.h
1.12 +* encoding: US-ASCII
1.13 +* tab size: 8 (not used)
1.14 +* indentation:4
1.15 +*
1.16 +* created on: 2009Jan05
1.17 +* created by: Andy Heninger
1.18 +*
1.19 +* Internal classes for compiling confusable data into its binary (runtime) form.
1.20 +*/
1.21 +
1.22 +#ifndef __USPOOF_BUILDCONF_H__
1.23 +#define __USPOOF_BUILDCONF_H__
1.24 +
1.25 +#if !UCONFIG_NO_NORMALIZATION
1.26 +
1.27 +#if !UCONFIG_NO_REGULAR_EXPRESSIONS
1.28 +
1.29 +#include "uspoof_impl.h"
1.30 +
1.31 +U_NAMESPACE_BEGIN
1.32 +
1.33 +// SPUString
1.34 +// Holds a string that is the result of one of the mappings defined
1.35 +// by the confusable mapping data (confusables.txt from Unicode.org)
1.36 +// Instances of SPUString exist during the compilation process only.
1.37 +
1.38 +struct SPUString : public UMemory {
1.39 + UnicodeString *fStr; // The actual string.
1.40 + int32_t fStrTableIndex; // Index into the final runtime data for this string.
1.41 + // (or, for length 1, the single string char itself,
1.42 + // there being no string table entry for it.)
1.43 + SPUString(UnicodeString *s);
1.44 + ~SPUString();
1.45 +};
1.46 +
1.47 +
1.48 +// String Pool A utility class for holding the strings that are the result of
1.49 +// the spoof mappings. These strings will utimately end up in the
1.50 +// run-time String Table.
1.51 +// This is sort of like a sorted set of strings, except that ICU's anemic
1.52 +// built-in collections don't support those, so it is implemented with a
1.53 +// combination of a uhash and a UVector.
1.54 +
1.55 +
1.56 +class SPUStringPool : public UMemory {
1.57 + public:
1.58 + SPUStringPool(UErrorCode &status);
1.59 + ~SPUStringPool();
1.60 +
1.61 + // Add a string. Return the string from the table.
1.62 + // If the input parameter string is already in the table, delete the
1.63 + // input parameter and return the existing string.
1.64 + SPUString *addString(UnicodeString *src, UErrorCode &status);
1.65 +
1.66 +
1.67 + // Get the n-th string in the collection.
1.68 + SPUString *getByIndex(int32_t i);
1.69 +
1.70 + // Sort the contents; affects the ordering of getByIndex().
1.71 + void sort(UErrorCode &status);
1.72 +
1.73 + int32_t size();
1.74 +
1.75 + private:
1.76 + UVector *fVec; // Elements are SPUString *
1.77 + UHashtable *fHash; // Key: UnicodeString Value: SPUString
1.78 +};
1.79 +
1.80 +
1.81 +// class ConfusabledataBuilder
1.82 +// An instance of this class exists while the confusable data is being built from source.
1.83 +// It encapsulates the intermediate data structures that are used for building.
1.84 +// It exports one static function, to do a confusable data build.
1.85 +
1.86 +class ConfusabledataBuilder : public UMemory {
1.87 + private:
1.88 + SpoofImpl *fSpoofImpl;
1.89 + UChar *fInput;
1.90 + UHashtable *fSLTable;
1.91 + UHashtable *fSATable;
1.92 + UHashtable *fMLTable;
1.93 + UHashtable *fMATable;
1.94 + UnicodeSet *fKeySet; // A set of all keys (UChar32s) that go into the four mapping tables.
1.95 +
1.96 + // The binary data is first assembled into the following four collections, then
1.97 + // copied to its final raw-memory destination.
1.98 + UVector *fKeyVec;
1.99 + UVector *fValueVec;
1.100 + UnicodeString *fStringTable;
1.101 + UVector *fStringLengthsTable;
1.102 +
1.103 + SPUStringPool *stringPool;
1.104 + URegularExpression *fParseLine;
1.105 + URegularExpression *fParseHexNum;
1.106 + int32_t fLineNum;
1.107 +
1.108 + ConfusabledataBuilder(SpoofImpl *spImpl, UErrorCode &status);
1.109 + ~ConfusabledataBuilder();
1.110 + void build(const char * confusables, int32_t confusablesLen, UErrorCode &status);
1.111 +
1.112 + // Add an entry to the key and value tables being built
1.113 + // input: data from SLTable, MATable, etc.
1.114 + // outut: entry added to fKeyVec and fValueVec
1.115 + void addKeyEntry(UChar32 keyChar, // The key character
1.116 + UHashtable *table, // The table, one of SATable, MATable, etc.
1.117 + int32_t tableFlag, // One of USPOOF_SA_TABLE_FLAG, etc.
1.118 + UErrorCode &status);
1.119 +
1.120 + // From an index into fKeyVec & fValueVec
1.121 + // get a UnicodeString with the corresponding mapping.
1.122 + UnicodeString getMapping(int32_t index);
1.123 +
1.124 + // Populate the final binary output data array with the compiled data.
1.125 + void outputData(UErrorCode &status);
1.126 +
1.127 + public:
1.128 + static void buildConfusableData(SpoofImpl *spImpl, const char * confusables,
1.129 + int32_t confusablesLen, int32_t *errorType, UParseError *pe, UErrorCode &status);
1.130 +};
1.131 +U_NAMESPACE_END
1.132 +
1.133 +#endif
1.134 +#endif // !UCONFIG_NO_REGULAR_EXPRESSIONS
1.135 +#endif // __USPOOF_BUILDCONF_H__
