ICU-12646 Error code for long confusable entries. Documenting spoof data binary file...

X-SVN-Rev: 39360

diff --git a/icu4c/source/i18n/uspoof_conf.cpp b/icu4c/source/i18n/uspoof_conf.cpp
index 9584e9484e5124c97749799f4a9d9a8247d69fe6..587dc4e01c3d53c20024e8508b3fd1aab8cec7f9 100644 (file)
--- a/icu4c/source/i18n/uspoof_conf.cpp
+++ b/icu4c/source/i18n/uspoof_conf.cpp
@@ -344,6 +344,13 @@ void ConfusabledataBuilder::build(const char * confusables, int32_t confusablesL
             SPUString *targetMapping = static_cast<SPUString *>(uhash_iget(fTable, keyChar));
             U_ASSERT(targetMapping != NULL);
 
+            // Set an error code if trying to consume a long string.  Otherwise,
+            // codePointAndLengthToKey will abort on a U_ASSERT.
+            if (targetMapping->fStr->length() > 256) {
+                status = U_ILLEGAL_ARGUMENT_ERROR;
+                return;
+            }
+
             int32_t key = ConfusableDataUtils::codePointAndLengthToKey(keyChar,
                 targetMapping->fStr->length());
             int32_t value = targetMapping->fCharOrStrTableIndex;

diff --git a/icu4c/source/i18n/uspoof_impl.h b/icu4c/source/i18n/uspoof_impl.h
index 82b8b5d5704c5af4d5abf1bad6e5a3b4e9a9132c..530767038f3b71c5887be32df0142fc994835b93 100644 (file)
--- a/icu4c/source/i18n/uspoof_impl.h
+++ b/icu4c/source/i18n/uspoof_impl.h
@@ -132,7 +132,7 @@ public:
 
 
 //
-//  Confusable Mappings Data Structures
+//  Confusable Mappings Data Structures, version 2.0
 //
 //    For the confusable data, we are essentially implementing a map,
 //       key:    a code point
@@ -156,6 +156,16 @@ public:
 //
 //       There is no nul character or other mark between adjacent strings.
 //
+//----------------------------------------------------------------------------
+//
+//  Changes from format version 1 to format version 2:
+//      1) Removal of the whole-script confusable data tables.
+//      2) Removal of the SL/SA/ML/MA and multi-table flags in the key bitmask.
+//      3) Expansion of string length value in the key bitmask from 2 bits to 8 bits.
+//      4) Removal of the string lengths table since 8 bits is sufficient for the
+//         lengths of all entries in confusables.txt.
+
+
 
 // Internal functions for manipulating confusable data table keys
 #define USPOOF_CONFUSABLE_DATA_FORMAT_VERSION 2  // version for ICU 58