projects / apple / icu.git / blame
| Commit | Line | Data |
|---|---|---|
| 729e4ab9 A |
1 | /* |
| 2 | *************************************************************************** | |
| 3 | * Copyright (C) 2008-2009, International Business Machines Corporation | |
| 4 | * and others. All Rights Reserved. | |
| 5 | *************************************************************************** | |
| 6 | * file name: uspoof_build.cpp | |
| 7 | * encoding: US-ASCII | |
| 8 | * tab size: 8 (not used) | |
| 9 | * indentation:4 | |
| 10 | * | |
| 11 | * created on: 2008 Dec 8 | |
| 12 | * created by: Andy Heninger | |
| 13 | * | |
| 14 | * Unicode Spoof Detection Data Builder | |
| 15 | * Builder-related functions are kept in separate files so that applications not needing | |
| 16 | * the builder can more easily exclude them, typically by means of static linking. | |
| 17 | * | |
| 18 | * There are three relatively independent sets of Spoof data, | |
| 19 | * Confusables, | |
| 20 | * Whole Script Confusables | |
| 21 | * ID character extensions. | |
| 22 | * | |
| 23 | * The data tables for each are built separately, each from its own definitions | |
| 24 | */ | |
| 25 | ||
| 26 | #include "unicode/utypes.h" | |
| 27 | #include "unicode/uspoof.h" | |
| 28 | #include "unicode/unorm.h" | |
| 29 | #include "unicode/uregex.h" | |
| 30 | #include "unicode/ustring.h" | |
| 31 | #include "cmemory.h" | |
| 32 | #include "uspoof_impl.h" | |
| 33 | #include "uhash.h" | |
| 34 | #include "uvector.h" | |
| 35 | #include "uassert.h" | |
| 36 | #include "uarrsort.h" | |
| 37 | #include "uspoof_conf.h" | |
| 38 | #include "uspoof_wsconf.h" | |
| 39 | ||
| 40 | #if !UCONFIG_NO_NORMALIZATION | |
| 41 | ||
| 42 | U_NAMESPACE_USE | |
| 43 | ||
| 44 | ||
| 45 | // The main data building function | |
| 46 | ||
| 47 | U_CAPI USpoofChecker * U_EXPORT2 | |
| 48 | uspoof_openFromSource(const char *confusables, int32_t confusablesLen, | |
| 49 | const char *confusablesWholeScript, int32_t confusablesWholeScriptLen, | |
| 50 | int32_t *errorType, UParseError *pe, UErrorCode *status) { | |
| 51 | ||
| 52 | if (U_FAILURE(*status)) { | |
| 53 | return NULL; | |
| 54 | } | |
| 55 | #if UCONFIG_NO_REGULAR_EXPRESSIONS | |
| 56 | *status = U_UNSUPPORTED_ERROR; | |
| 57 | return NULL; | |
| 58 | #else | |
| 59 | if (errorType!=NULL) { | |
| 60 | *errorType = 0; | |
| 61 | } | |
| 62 | if (pe != NULL) { | |
| 63 | pe->line = 0; | |
| 64 | pe->offset = 0; | |
| 65 | pe->preContext[0] = 0; | |
| 66 | pe->postContext[0] = 0; | |
| 67 | } | |
| 68 | ||
| 69 | // Set up a shell of a spoof detector, with empty data. | |
| 70 | SpoofData *newSpoofData = new SpoofData(*status); | |
| 71 | SpoofImpl *This = new SpoofImpl(newSpoofData, *status); | |
| 72 | ||
| 73 | // Compile the binary data from the source (text) format. | |
| 74 | ConfusabledataBuilder::buildConfusableData(This, confusables, confusablesLen, errorType, pe, *status); | |
| 75 | buildWSConfusableData(This, confusablesWholeScript, confusablesWholeScriptLen, pe, *status); | |
| 76 | ||
| 77 | if (U_FAILURE(*status)) { | |
| 78 | delete This; | |
| 79 | This = NULL; | |
| 80 | } | |
| 81 | return (USpoofChecker *)This; | |
| 82 | #endif // UCONFIG_NO_REGULAR_EXPRESSIONS | |
| 83 | } | |
| 84 | ||
| 85 | #endif |