../deps/icu-small/source/i18n/translit.cpp

Bug Summary

File:	out/../deps/icu-small/source/i18n/translit.cpp
Warning:	line 1351, column 5 Returning null reference
Annotated Source Code

Press '?' to see keyboard shortcuts
Show analyzer invocation
clang -cc1 -cc1 -triple x86_64-unknown-linux-gnu -analyze -disable-free -clear-ast-before-backend -disable-llvm-verifier -discard-value-names -main-file-name translit.cpp -analyzer-checker=core -analyzer-checker=apiModeling -analyzer-checker=unix -analyzer-checker=deadcode -analyzer-checker=cplusplus -analyzer-checker=security.insecureAPI.UncheckedReturn -analyzer-checker=security.insecureAPI.getpw -analyzer-checker=security.insecureAPI.gets -analyzer-checker=security.insecureAPI.mktemp -analyzer-checker=security.insecureAPI.mkstemp -analyzer-checker=security.insecureAPI.vfork -analyzer-checker=nullability.NullPassedToNonnull -analyzer-checker=nullability.NullReturnedFromNonnull -analyzer-output plist -w -setup-static-analyzer -mrelocation-model pic -pic-level 2 -pic-is-pie -mframe-pointer=all -fmath-errno -ffp-contract=on -fno-rounding-math -mconstructor-aliases -funwind-tables=2 -target-cpu x86-64 -tune-cpu generic -debugger-tuning=gdb -fcoverage-compilation-dir=/home/maurizio/node-v18.6.0/out -resource-dir /usr/local/lib/clang/16.0.0 -D V8_DEPRECATION_WARNINGS -D V8_IMMINENT_DEPRECATION_WARNINGS -D _GLIBCXX_USE_CXX11_ABI=1 -D NODE_OPENSSL_CONF_NAME=nodejs_conf -D NODE_OPENSSL_HAS_QUIC -D __STDC_FORMAT_MACROS -D OPENSSL_NO_PINSHARED -D OPENSSL_THREADS -D U_COMMON_IMPLEMENTATION=1 -D U_I18N_IMPLEMENTATION=1 -D U_IO_IMPLEMENTATION=1 -D U_TOOLUTIL_IMPLEMENTATION=1 -D U_ATTRIBUTE_DEPRECATED= -D _CRT_SECURE_NO_DEPRECATE= -D U_STATIC_IMPLEMENTATION=1 -D UCONFIG_NO_SERVICE=1 -D U_ENABLE_DYLOAD=0 -D U_HAVE_STD_STRING=1 -D UCONFIG_NO_BREAK_ITERATION=0 -I ../deps/icu-small/source/common -I ../deps/icu-small/source/i18n -I ../deps/icu-small/source/tools/toolutil -internal-isystem /usr/lib/gcc/x86_64-redhat-linux/8/../../../../include/c++/8 -internal-isystem /usr/lib/gcc/x86_64-redhat-linux/8/../../../../include/c++/8/x86_64-redhat-linux -internal-isystem /usr/lib/gcc/x86_64-redhat-linux/8/../../../../include/c++/8/backward -internal-isystem /usr/local/lib/clang/16.0.0/include -internal-isystem /usr/local/include -internal-isystem /usr/lib/gcc/x86_64-redhat-linux/8/../../../../x86_64-redhat-linux/include -internal-externc-isystem /include -internal-externc-isystem /usr/include -O3 -Wno-unused-parameter -Wno-deprecated-declarations -Wno-strict-aliasing -std=gnu++17 -fdeprecated-macro -fdebug-compilation-dir=/home/maurizio/node-v18.6.0/out -ferror-limit 19 -fgnuc-version=4.2.1 -vectorize-loops -vectorize-slp -analyzer-output=html -faddrsig -D__GCC_HAVE_DWARF2_CFI_ASM=1 -o /tmp/scan-build-2022-08-22-142216-507842-1 -x c++ ../deps/icu-small/source/i18n/translit.cpp
1// © 2016 and later: Unicode, Inc. and others.
2// License & terms of use: http://www.unicode.org/copyright.html
3/*
**********************************************************************
*   Copyright (C) 1999-2016, International Business Machines
*   Corporation and others.  All Rights Reserved.
**********************************************************************
*   Date        Name        Description
*   11/17/99    aliu        Creation.
**********************************************************************
*/

13#include "utypeinfo.h"  // for 'typeid' to work

15#include "unicode/utypes.h"

17#if !UCONFIG_NO_TRANSLITERATION0

19#include "unicode/putil.h"
20#include "unicode/translit.h"
21#include "unicode/locid.h"
22#include "unicode/msgfmt.h"
23#include "unicode/rep.h"
24#include "unicode/resbund.h"
25#include "unicode/unifilt.h"
26#include "unicode/uniset.h"
27#include "unicode/uscript.h"
28#include "unicode/strenum.h"
29#include "unicode/utf16.h"
30#include "cpdtrans.h"
31#include "nultrans.h"
32#include "rbt_data.h"
33#include "rbt_pars.h"
34#include "rbt.h"
35#include "transreg.h"
36#include "name2uni.h"
37#include "nortrans.h"
38#include "remtrans.h"
39#include "titletrn.h"
40#include "tolowtrn.h"
41#include "toupptrn.h"
42#include "uni2name.h"
43#include "brktrans.h"
44#include "esctrn.h"
45#include "unesctrn.h"
46#include "tridpars.h"
47#include "anytrans.h"
48#include "util.h"
49#include "hash.h"
50#include "mutex.h"
51#include "ucln_in.h"
52#include "uassert.h"
53#include "cmemory.h"
54#include "cstring.h"
55#include "uinvchar.h"

57static const UChar TARGET_SEP  = 0x002D; /*-*/
58static const UChar ID_DELIM    = 0x003B; /*;*/
59static const UChar VARIANT_SEP = 0x002F; // '/'

61/**
* Prefix for resource bundle key for the display name for a
* transliterator.  The ID is appended to this to form the key.
* The resource bundle value should be a String.
*/
66static const char RB_DISPLAY_NAME_PREFIX[] = "%Translit%%";

68/**
* Prefix for resource bundle key for the display name for a
* transliterator SCRIPT.  The ID is appended to this to form the key.
* The resource bundle value should be a String.
*/
73static const char RB_SCRIPT_DISPLAY_NAME_PREFIX[] = "%Translit%";

75/**
* Resource bundle key for display name pattern.
* The resource bundle value should be a String forming a
* MessageFormat pattern, e.g.:
* "{0,choice,0#|1#{1} Transliterator|2#{1} to {2} Transliterator}".
*/
81static const char RB_DISPLAY_NAME_PATTERN[] = "TransliteratorNamePattern";

83/**
* Resource bundle key for the list of RuleBasedTransliterator IDs.
* The resource bundle value should be a String[] with each element
* being a valid ID.  The ID will be appended to RB_RULE_BASED_PREFIX
* to obtain the class name in which the RB_RULE key will be sought.
*/
89static const char RB_RULE_BASED_IDS[] = "RuleBasedTransliteratorIDs";

91/**
* The mutex controlling access to registry object.
*/
94static icu::UMutex registryMutex;

96/**
* System transliterator registry; non-null when initialized.
*/
99static icu::TransliteratorRegistry* registry = 0;

101// Macro to check/initialize the registry. ONLY USE WITHIN
102// MUTEX. Avoids function call when registry is initialized.
103#define HAVE_REGISTRY(status)(registry!=0 || initializeRegistry(status)) (registry!=0 || initializeRegistry(status))

105U_NAMESPACE_BEGINnamespace icu_71 {

107UOBJECT_DEFINE_ABSTRACT_RTTI_IMPLEMENTATION(Transliterator)UClassID Transliterator::getStaticClassID() { static char classID
 = 0; return (UClassID)&classID; }

109/**
* Return TRUE if the given UTransPosition is valid for text of
* the given length.
*/
113static inline UBool positionIsValid(UTransPosition& index, int32_t len) {
  return !(index.contextStart < 0 ||
           index.start < index.contextStart ||
           index.limit < index.start ||
           index.contextLimit < index.limit ||
           len < index.contextLimit);
119}

121/**
* Default constructor.
* @param theID the string identifier for this transliterator
* @param theFilter the filter.  Any character for which
* <tt>filter.contains()</tt> returns <tt>FALSE</tt> will not be
* altered by this transliterator.  If <tt>filter</tt> is
* <tt>null</tt> then no filtering is applied.
*/
129Transliterator::Transliterator(const UnicodeString& theID,
                             UnicodeFilter* adoptedFilter) :
  UObject(), ID(theID), filter(adoptedFilter),
  maximumContextLength(0)
133{
  // NUL-terminate the ID string, which is a non-aliased copy.
  ID.append((UChar)0);
  ID.truncate(ID.length()-1);
137}

139/**
* Destructor.
*/
142Transliterator::~Transliterator() {
  if (filter) {
      delete filter;
  }
146}

148/**
* Copy constructor.
*/
151Transliterator::Transliterator(const Transliterator& other) :
  UObject(other), ID(other.ID), filter(0),
  maximumContextLength(other.maximumContextLength)
154{
  // NUL-terminate the ID string, which is a non-aliased copy.
  ID.append((UChar)0);
  ID.truncate(ID.length()-1);

  if (other.filter != 0) {
      // We own the filter, so we must have our own copy
      filter = other.filter->clone();
  }
163}

165Transliterator* Transliterator::clone() const {
  return NULL__null;
167}

169/**
* Assignment operator.
*/
172Transliterator& Transliterator::operator=(const Transliterator& other) {
  if (this == &other) { return *this; }  // self-assignment: no-op
  ID = other.ID;
  // NUL-terminate the ID string
  ID.getTerminatedBuffer();

  maximumContextLength = other.maximumContextLength;
  adoptFilter((other.filter == 0) ? 0 : other.filter->clone());
  return *this;
181}

183/**
* Transliterates a segment of a string.  <code>Transliterator</code> API.
* @param text the string to be transliterated
* @param start the beginning index, inclusive; <code>0 <= start
* <= limit</code>.
* @param limit the ending index, exclusive; <code>start <= limit
* <= text.length()</code>.
* @return the new limit index, or -1
*/
192int32_t Transliterator::transliterate(Replaceable& text,
                                    int32_t start, int32_t limit) const {
  if (start < 0 ||
      limit < start ||
      text.length() < limit) {
      return -1;
  }

  UTransPosition offsets;
  offsets.contextStart= start;
  offsets.contextLimit = limit;
  offsets.start = start;
  offsets.limit = limit;
  filteredTransliterate(text, offsets, FALSE0, TRUE1);
  return offsets.limit;
207}

209/**
* Transliterates an entire string in place. Convenience method.
* @param text the string to be transliterated
*/
213void Transliterator::transliterate(Replaceable& text) const {
  transliterate(text, 0, text.length());
215}

217/**
* Transliterates the portion of the text buffer that can be
* transliterated unambiguosly after new text has been inserted,
* typically as a result of a keyboard event.  The new text in
* <code>insertion</code> will be inserted into <code>text</code>
* at <code>index.contextLimit</code>, advancing
* <code>index.contextLimit</code> by <code>insertion.length()</code>.
* Then the transliterator will try to transliterate characters of
* <code>text</code> between <code>index.start</code> and
* <code>index.contextLimit</code>.  Characters before
* <code>index.start</code> will not be changed.
*
* <p>Upon return, values in <code>index</code> will be updated.
* <code>index.contextStart</code> will be advanced to the first
* character that future calls to this method will read.
* <code>index.start</code> and <code>index.contextLimit</code> will
* be adjusted to delimit the range of text that future calls to
* this method may change.
*
* <p>Typical usage of this method begins with an initial call
* with <code>index.contextStart</code> and <code>index.contextLimit</code>
* set to indicate the portion of <code>text</code> to be
* transliterated, and <code>index.start == index.contextStart</code>.
* Thereafter, <code>index</code> can be used without
* modification in future calls, provided that all changes to
* <code>text</code> are made via this method.
*
* <p>This method assumes that future calls may be made that will
* insert new text into the buffer.  As a result, it only performs
* unambiguous transliterations.  After the last call to this
* method, there may be untransliterated text that is waiting for
* more input to resolve an ambiguity.  In order to perform these
* pending transliterations, clients should call {@link
* #finishKeyboardTransliteration} after the last call to this
* method has been made.
* 
* @param text the buffer holding transliterated and untransliterated text
* @param index an array of three integers.
*
* <ul><li><code>index.contextStart</code>: the beginning index,
* inclusive; <code>0 <= index.contextStart <= index.contextLimit</code>.
*
* <li><code>index.contextLimit</code>: the ending index, exclusive;
* <code>index.contextStart <= index.contextLimit <= text.length()</code>.
* <code>insertion</code> is inserted at
* <code>index.contextLimit</code>.
*
* <li><code>index.start</code>: the next character to be
* considered for transliteration; <code>index.contextStart <=
* index.start <= index.contextLimit</code>.  Characters before
* <code>index.start</code> will not be changed by future calls
* to this method.</ul>
*
* @param insertion text to be inserted and possibly
* transliterated into the translation buffer at
* <code>index.contextLimit</code>.  If <code>null</code> then no text
* is inserted.
* @see #START
* @see #LIMIT
* @see #CURSOR
* @see #handleTransliterate
* @exception IllegalArgumentException if <code>index</code>
* is invalid
*/
281void Transliterator::transliterate(Replaceable& text,
                                 UTransPosition& index,
                                 const UnicodeString& insertion,
                                 UErrorCode &status) const {
  _transliterate(text, index, &insertion, status);
286}

288/**
* Transliterates the portion of the text buffer that can be
* transliterated unambiguosly after a new character has been
* inserted, typically as a result of a keyboard event.  This is a
* convenience method; see {@link
* #transliterate(Replaceable, int[], String)} for details.
* @param text the buffer holding transliterated and
* untransliterated text
* @param index an array of three integers.  See {@link
* #transliterate(Replaceable, int[], String)}.
* @param insertion text to be inserted and possibly
* transliterated into the translation buffer at
* <code>index.contextLimit</code>.
* @see #transliterate(Replaceable, int[], String)
*/
303void Transliterator::transliterate(Replaceable& text,
                                 UTransPosition& index,
                                 UChar32 insertion,
                                 UErrorCode& status) const {
  UnicodeString str(insertion);
  _transliterate(text, index, &str, status);
309}

311/**
* Transliterates the portion of the text buffer that can be
* transliterated unambiguosly.  This is a convenience method; see
* {@link #transliterate(Replaceable, int[], String)} for
* details.
* @param text the buffer holding transliterated and
* untransliterated text
* @param index an array of three integers.  See {@link
* #transliterate(Replaceable, int[], String)}.
* @see #transliterate(Replaceable, int[], String)
*/
322void Transliterator::transliterate(Replaceable& text,
                                 UTransPosition& index,
                                 UErrorCode& status) const {
  _transliterate(text, index, 0, status);
326}

328/**
* Finishes any pending transliterations that were waiting for
* more characters.  Clients should call this method as the last
* call after a sequence of one or more calls to
* <code>transliterate()</code>.
* @param text the buffer holding transliterated and
* untransliterated text.
* @param index the array of indices previously passed to {@link
* #transliterate}
*/
338void Transliterator::finishTransliteration(Replaceable& text,
                                         UTransPosition& index) const {
  if (!positionIsValid(index, text.length())) {
      return;
  }

  filteredTransliterate(text, index, FALSE0, TRUE1);
345}

347/**
* This internal method does keyboard transliteration.  If the
* 'insertion' is non-null then we append it to 'text' before
* proceeding.  This method calls through to the pure virtual
* framework method handleTransliterate() to do the actual
* work.
*/
354void Transliterator::_transliterate(Replaceable& text,
                                  UTransPosition& index,
                                  const UnicodeString* insertion,
                                  UErrorCode &status) const {
  if (U_FAILURE(status)) {
      return;
  }

  if (!positionIsValid(index, text.length())) {
      status = U_ILLEGAL_ARGUMENT_ERROR;
      return;
  }

367//    int32_t originalStart = index.contextStart;
  if (insertion != 0) {
      text.handleReplaceBetween(index.limit, index.limit, *insertion);
      index.limit += insertion->length();
      index.contextLimit += insertion->length();
  }

  if (index.limit > 0 &&
      U16_IS_LEAD(text.charAt(index.limit - 1))(((text.charAt(index.limit - 1))&0xfffffc00)==0xd800)) {
      // Oops, there is a dangling lead surrogate in the buffer.
      // This will break most transliterators, since they will
      // assume it is part of a pair.  Don't transliterate until
      // more text comes in.
      return;
  }

  filteredTransliterate(text, index, TRUE1, TRUE1);

385#if 0
  // TODO
  // I CAN'T DO what I'm attempting below now that the Kleene star
  // operator is supported.  For example, in the rule

  //   ([:Lu:]+) { x } > $1;

  // what is the maximum context length?  getMaximumContextLength()
  // will return 1, but this is just the length of the ante context
  // part of the pattern string -- 1 character, which is a standin
  // for a Quantifier, which contains a StringMatcher, which
  // contains a UnicodeSet.

  // There is a complicated way to make this work again, and that's
  // to add a "maximum left context" protocol into the
  // UnicodeMatcher hierarchy.  At present I'm not convinced this is
  // worth it.

  // ---

  // The purpose of the code below is to keep the context small
  // while doing incremental transliteration.  When part of the left
  // context (between contextStart and start) is no longer needed,
  // we try to advance contextStart past that portion.  We use the
  // maximum context length to do so.
  int32_t newCS = index.start;
  int32_t n = getMaximumContextLength();
  while (newCS > originalStart && n-- > 0) {
      --newCS;
      newCS -= U16_LENGTH(text.char32At(newCS))((uint32_t)(text.char32At(newCS))<=0xffff ? 1 : 2) - 1;
  }
  index.contextStart = uprv_maxuprv_max_71(newCS, originalStart);
417#endif
418}

420/**
* This method breaks up the input text into runs of unfiltered
* characters.  It passes each such run to
* <subclass>.handleTransliterate().  Subclasses that can handle the
* filter logic more efficiently themselves may override this method.
*
* All transliteration calls in this class go through this method.
*/
428void Transliterator::filteredTransliterate(Replaceable& text,
                                         UTransPosition& index,
                                         UBool incremental,
                                         UBool rollback) const {
  // Short circuit path for transliterators with no filter in
  // non-incremental mode.
  if (filter == 0 && !rollback) {
      handleTransliterate(text, index, incremental);
      return;
  }

  //----------------------------------------------------------------------
  // This method processes text in two groupings:
  //
  // RUNS -- A run is a contiguous group of characters which are contained
  // in the filter for this transliterator (filter.contains(ch) == TRUE).
  // Text outside of runs may appear as context but it is not modified.
  // The start and limit Position values are narrowed to each run.
  //
  // PASSES (incremental only) -- To make incremental mode work correctly,
  // each run is broken up into n passes, where n is the length (in code
  // points) of the run.  Each pass contains the first n characters.  If a
  // pass is completely transliterated, it is committed, and further passes
  // include characters after the committed text.  If a pass is blocked,
  // and does not transliterate completely, then this method rolls back
  // the changes made during the pass, extends the pass by one code point,
  // and tries again.
  //----------------------------------------------------------------------
  
  // globalLimit is the limit value for the entire operation.  We
  // set index.limit to the end of each unfiltered run before
  // calling handleTransliterate(), so we need to maintain the real
  // value of index.limit here.  After each transliteration, we
  // update globalLimit for insertions or deletions that have
  // happened.
  int32_t globalLimit = index.limit;
  
  // If there is a non-null filter, then break the input text up.  Say the
  // input text has the form:
  //   xxxabcxxdefxx
  // where 'x' represents a filtered character (filter.contains('x') ==
  // false).  Then we break this up into:
  //   xxxabc xxdef xx
  // Each pass through the loop consumes a run of filtered
  // characters (which are ignored) and a subsequent run of
  // unfiltered characters (which are transliterated).
  
  for (;;) {

      if (filter != NULL__null) {
          // Narrow the range to be transliterated to the first segment
          // of unfiltered characters at or after index.start.

          // Advance past filtered chars
          UChar32 c;
          while (index.start < globalLimit &&
                 !filter->contains(c=text.char32At(index.start))) {
              index.start += U16_LENGTH(c)((uint32_t)(c)<=0xffff ? 1 : 2);
          }

          // Find the end of this run of unfiltered chars
          index.limit = index.start;
          while (index.limit < globalLimit &&
                 filter->contains(c=text.char32At(index.limit))) {
              index.limit += U16_LENGTH(c)((uint32_t)(c)<=0xffff ? 1 : 2);
          }
      }

      // Check to see if the unfiltered run is empty.  This only
      // happens at the end of the string when all the remaining
      // characters are filtered.
      if (index.limit == index.start) {
          // assert(index.start == globalLimit);
          break;
      }

      // Is this run incremental?  If there is additional
      // filtered text (if limit < globalLimit) then we pass in
      // an incremental value of FALSE to force the subclass to
      // complete the transliteration for this run.
      UBool isIncrementalRun =
          (index.limit < globalLimit ? FALSE0 : incremental);
      
      int32_t delta;

      // Implement rollback.  To understand the need for rollback,
      // consider the following transliterator:
      //
      //  "t" is "a > A;"
      //  "u" is "A > b;"
      //  "v" is a compound of "t; NFD; u" with a filter [:Ll:]
      //
      // Now apply "c" to the input text "a".  The result is "b".  But if
      // the transliteration is done incrementally, then the NFD holds
      // things up after "t" has already transformed "a" to "A".  When
      // finishTransliterate() is called, "A" is _not_ processed because
      // it gets excluded by the [:Ll:] filter, and the end result is "A"
      // -- incorrect.  The problem is that the filter is applied to a
      // partially-transliterated result, when we only want it to apply to
      // input text.  Although this example hinges on a compound
      // transliterator containing NFD and a specific filter, it can
      // actually happen with any transliterator which may do a partial
      // transformation in incremental mode into characters outside its
      // filter.
      //
      // To handle this, when in incremental mode we supply characters to
      // handleTransliterate() in several passes.  Each pass adds one more
      // input character to the input text.  That is, for input "ABCD", we
      // first try "A", then "AB", then "ABC", and finally "ABCD".  If at
      // any point we block (upon return, start < limit) then we roll
      // back.  If at any point we complete the run (upon return start ==
      // limit) then we commit that run.

      if (rollback && isIncrementalRun) {

          int32_t runStart = index.start;
          int32_t runLimit = index.limit;
          int32_t runLength =  runLimit - runStart;

          // Make a rollback copy at the end of the string
          int32_t rollbackOrigin = text.length();
          text.copy(runStart, runLimit, rollbackOrigin);

          // Variables reflecting the commitment of completely
          // transliterated text.  passStart is the runStart, advanced
          // past committed text.  rollbackStart is the rollbackOrigin,
          // advanced past rollback text that corresponds to committed
          // text.
          int32_t passStart = runStart;
          int32_t rollbackStart = rollbackOrigin;

          // The limit for each pass; we advance by one code point with
          // each iteration.
          int32_t passLimit = index.start;

          // Total length, in 16-bit code units, of uncommitted text.
          // This is the length to be rolled back.
          int32_t uncommittedLength = 0;

          // Total delta (change in length) for all passes
          int32_t totalDelta = 0;

          // PASS MAIN LOOP -- Start with a single character, and extend
          // the text by one character at a time.  Roll back partial
          // transliterations and commit complete transliterations.
          for (;;) {
              // Length of additional code point, either one or two
              int32_t charLength = U16_LENGTH(text.char32At(passLimit))((uint32_t)(text.char32At(passLimit))<=0xffff ? 1 : 2);
              passLimit += charLength;
              if (passLimit > runLimit) {
                  break;
              }
              uncommittedLength += charLength;

              index.limit = passLimit;

              // Delegate to subclass for actual transliteration.  Upon
              // return, start will be updated to point after the
              // transliterated text, and limit and contextLimit will be
              // adjusted for length changes.
              handleTransliterate(text, index, TRUE1);

              delta = index.limit - passLimit; // change in length

              // We failed to completely transliterate this pass.
              // Roll back the text.  Indices remain unchanged; reset
              // them where necessary.
              if (index.start != index.limit) {
                  // Find the rollbackStart, adjusted for length changes
                  // and the deletion of partially transliterated text.
                  int32_t rs = rollbackStart + delta - (index.limit - passStart);

                  // Delete the partially transliterated text
                  text.handleReplaceBetween(passStart, index.limit, UnicodeString());

                  // Copy the rollback text back
                  text.copy(rs, rs + uncommittedLength, passStart);

                  // Restore indices to their original values
                  index.start = passStart;
                  index.limit = passLimit;
                  index.contextLimit -= delta;
              }

              // We did completely transliterate this pass.  Update the
              // commit indices to record how far we got.  Adjust indices
              // for length change.
              else {
                  // Move the pass indices past the committed text.
                  passStart = passLimit = index.start;

                  // Adjust the rollbackStart for length changes and move
                  // it past the committed text.  All characters we've
                  // processed to this point are committed now, so zero
                  // out the uncommittedLength.
                  rollbackStart += delta + uncommittedLength;
                  uncommittedLength = 0;

                  // Adjust indices for length changes.
                  runLimit += delta;
                  totalDelta += delta;
              }
          }

          // Adjust overall limit and rollbackOrigin for insertions and
          // deletions.  Don't need to worry about contextLimit because
          // handleTransliterate() maintains that.
          rollbackOrigin += totalDelta;
          globalLimit += totalDelta;

          // Delete the rollback copy
          text.handleReplaceBetween(rollbackOrigin, rollbackOrigin + runLength, UnicodeString());

          // Move start past committed text
          index.start = passStart;
      }

      else {
          // Delegate to subclass for actual transliteration.
          int32_t limit = index.limit;
          handleTransliterate(text, index, isIncrementalRun);
          delta = index.limit - limit; // change in length

          // In a properly written transliterator, start == limit after
          // handleTransliterate() returns when incremental is false.
          // Catch cases where the subclass doesn't do this, and throw
          // an exception.  (Just pinning start to limit is a bad idea,
          // because what's probably happening is that the subclass
          // isn't transliterating all the way to the end, and it should
          // in non-incremental mode.)
          if (!incremental && index.start != index.limit) {
              // We can't throw an exception, so just fudge things
              index.start = index.limit;
          }

          // Adjust overall limit for insertions/deletions.  Don't need
          // to worry about contextLimit because handleTransliterate()
          // maintains that.
          globalLimit += delta;
      }

      if (filter == NULL__null || isIncrementalRun) {
          break;
      }

      // If we did completely transliterate this
      // run, then repeat with the next unfiltered run.
  }

  // Start is valid where it is.  Limit needs to be put back where
  // it was, modulo adjustments for deletions/insertions.
  index.limit = globalLimit;
680}

682void Transliterator::filteredTransliterate(Replaceable& text,
                                         UTransPosition& index,
                                         UBool incremental) const {
  filteredTransliterate(text, index, incremental, FALSE0);
686}

688/**
* Method for subclasses to use to set the maximum context length.
* @see #getMaximumContextLength
*/
692void Transliterator::setMaximumContextLength(int32_t maxContextLength) {
  maximumContextLength = maxContextLength;
694}

696/**
* Returns a programmatic identifier for this transliterator.
* If this identifier is passed to <code>getInstance()</code>, it
* will return this object, if it has been registered.
* @see #registerInstance
* @see #getAvailableIDs
*/
703const UnicodeString& Transliterator::getID(void) const {
  return ID;
705}

707/**
* Returns a name for this transliterator that is appropriate for
* display to the user in the default locale.  See {@link
* #getDisplayName(Locale)} for details.
*/
712UnicodeString& U_EXPORT2 Transliterator::getDisplayName(const UnicodeString& ID,
                                            UnicodeString& result) {
  return getDisplayName(ID, Locale::getDefault(), result);
715}

717/**
* Returns a name for this transliterator that is appropriate for
* display to the user in the given locale.  This name is taken
* from the locale resource data in the standard manner of the
* <code>java.text</code> package.
*
* <p>If no localized names exist in the system resource bundles,
* a name is synthesized using a localized
* <code>MessageFormat</code> pattern from the resource data.  The
* arguments to this pattern are an integer followed by one or two
* strings.  The integer is the number of strings, either 1 or 2.
* The strings are formed by splitting the ID for this
* transliterator at the first TARGET_SEP.  If there is no TARGET_SEP, then the
* entire ID forms the only string.
* @param inLocale the Locale in which the display name should be
* localized.
* @see java.text.MessageFormat
*/
735UnicodeString& U_EXPORT2 Transliterator::getDisplayName(const UnicodeString& id,
                                            const Locale& inLocale,
                                            UnicodeString& result) {
  UErrorCode status = U_ZERO_ERROR;

  ResourceBundle bundle(U_ICUDATA_TRANSLIT"icudt" "71" "l" "-" "translit", inLocale, status);

  // Suspend checking status until later...

  result.truncate(0);

  // Normalize the ID
  UnicodeString source, target, variant;
  UBool sawSource;
  TransliteratorIDParser::IDtoSTV(id, source, target, variant, sawSource);
  if (target.length() < 1) {
      // No target; malformed id
      return result;
  }
  if (variant.length() > 0) { // Change "Foo" to "/Foo"
      variant.insert(0, VARIANT_SEP);
  }
  UnicodeString ID(source);
  ID.append(TARGET_SEP).append(target).append(variant);

  // build the char* key
  if (uprv_isInvariantUStringuprv_isInvariantUString_71(ID.getBuffer(), ID.length())) {
      char key[200];
      uprv_strcpy(key, RB_DISPLAY_NAME_PREFIX):: strcpy(key, RB_DISPLAY_NAME_PREFIX);
      int32_t length=(int32_t)uprv_strlen(RB_DISPLAY_NAME_PREFIX):: strlen(RB_DISPLAY_NAME_PREFIX);
      ID.extract(0, (int32_t)(sizeof(key)-length), key+length, (int32_t)(sizeof(key)-length), US_INVicu::UnicodeString::kInvariant);

      // Try to retrieve a UnicodeString from the bundle.
      UnicodeString resString = bundle.getStringEx(key, status);

      if (U_SUCCESS(status) && resString.length() != 0) {
          return result = resString; // [sic] assign & return
      }

774#if !UCONFIG_NO_FORMATTING0
      // We have failed to get a name from the locale data.  This is
      // typical, since most transliterators will not have localized
      // name data.  The next step is to retrieve the MessageFormat
      // pattern from the locale data and to use it to synthesize the
      // name from the ID.

      status = U_ZERO_ERROR;
      resString = bundle.getStringEx(RB_DISPLAY_NAME_PATTERN, status);

      if (U_SUCCESS(status) && resString.length() != 0) {
          MessageFormat msg(resString, inLocale, status);
          // Suspend checking status until later...

          // We pass either 2 or 3 Formattable objects to msg.
          Formattable args[3];
          int32_t nargs;
          args[0].setLong(2); // # of args to follow
          args[1].setString(source);
          args[2].setString(target);
          nargs = 3;

          // Use display names for the scripts, if they exist
          UnicodeString s;
          length=(int32_t)uprv_strlen(RB_SCRIPT_DISPLAY_NAME_PREFIX):: strlen(RB_SCRIPT_DISPLAY_NAME_PREFIX);
          for (int j=1; j<=2; ++j) {
              status = U_ZERO_ERROR;
              uprv_strcpy(key, RB_SCRIPT_DISPLAY_NAME_PREFIX):: strcpy(key, RB_SCRIPT_DISPLAY_NAME_PREFIX);
              args[j].getString(s);
              if (uprv_isInvariantUStringuprv_isInvariantUString_71(s.getBuffer(), s.length())) {
                  s.extract(0, sizeof(key)-length-1, key+length, (int32_t)sizeof(key)-length-1, US_INVicu::UnicodeString::kInvariant);

                  resString = bundle.getStringEx(key, status);

                  if (U_SUCCESS(status)) {
                      args[j] = resString;
                  }
              }
          }

          status = U_ZERO_ERROR;
          FieldPosition pos; // ignored by msg
          msg.format(args, nargs, result, pos, status);
          if (U_SUCCESS(status)) {
              result.append(variant);
              return result;
          }
      }
822#endif
  }

  // We should not reach this point unless there is something
  // wrong with the build or the RB_DISPLAY_NAME_PATTERN has
  // been deleted from the root RB_LOCALE_ELEMENTS resource.
  result = ID;
  return result;
830}

832/**
* Returns the filter used by this transliterator, or <tt>null</tt>
* if this transliterator uses no filter.  Caller musn't delete
* the result!
*/
837const UnicodeFilter* Transliterator::getFilter(void) const {
  return filter;
839}

841/**
* Returns the filter used by this transliterator, or
* <tt>NULL</tt> if this transliterator uses no filter.  The
* caller must eventually delete the result.  After this call,
* this transliterator's filter is set to <tt>NULL</tt>.
*/
847UnicodeFilter* Transliterator::orphanFilter(void) {
  UnicodeFilter *result = filter;
  filter = NULL__null;
  return result;
851}

853/**
* Changes the filter used by this transliterator.  If the filter
* is set to <tt>null</tt> then no filtering will occur.
*
* <p>Callers must take care if a transliterator is in use by
* multiple threads.  The filter should not be changed by one
* thread while another thread may be transliterating.
*/
861void Transliterator::adoptFilter(UnicodeFilter* filterToAdopt) {
  delete filter;
  filter = filterToAdopt;
864}

866/**
* Returns this transliterator's inverse.  See the class
* documentation for details.  This implementation simply inverts
* the two entities in the ID and attempts to retrieve the
* resulting transliterator.  That is, if <code>getID()</code>
* returns "A-B", then this method will return the result of
* <code>getInstance("B-A")</code>, or <code>null</code> if that
* call fails.
*
* <p>This method does not take filtering into account.  The
* returned transliterator will have no filter.
*
* <p>Subclasses with knowledge of their inverse may wish to
* override this method.
*
* @return a transliterator that is an inverse, not necessarily
* exact, of this transliterator, or <code>null</code> if no such
* transliterator is registered.
* @see #registerInstance
*/
886Transliterator* Transliterator::createInverse(UErrorCode& status) const {
  UParseError parseError;
  return Transliterator::createInstance(ID, UTRANS_REVERSE,parseError,status);
889}

891Transliterator* U_EXPORT2
892Transliterator::createInstance(const UnicodeString& ID,
                              UTransDirection dir,
                              UErrorCode& status)
895{
  UParseError parseError;
  return createInstance(ID, dir, parseError, status);
898}

900/**
* Returns a <code>Transliterator</code> object given its ID.
* The ID must be either a system transliterator ID or a ID registered
* using <code>registerInstance()</code>.
*
* @param ID a valid ID, as enumerated by <code>getAvailableIDs()</code>
* @return A <code>Transliterator</code> object with the given ID
* @see #registerInstance
* @see #getAvailableIDs
* @see #getID
*/
911Transliterator* U_EXPORT2
912Transliterator::createInstance(const UnicodeString& ID,
                              UTransDirection dir,
                              UParseError& parseError,
                              UErrorCode& status)
916{
  if (U_FAILURE(status)) {
      return 0;
  }

  UnicodeString canonID;
  UVector list(status);
  if (U_FAILURE(status)) {
      return NULL__null;
  }

  UnicodeSet* globalFilter = nullptr;
  // TODO add code for parseError...currently unused, but
  // later may be used by parsing code...
  if (!TransliteratorIDParser::parseCompoundID(ID, dir, canonID, list, globalFilter)) {
      status = U_INVALID_ID;
      delete globalFilter;
      return NULL__null;
  }
  LocalPointer<UnicodeSet> lpGlobalFilter(globalFilter);
  
  TransliteratorIDParser::instantiateList(list, status);
  if (U_FAILURE(status)) {
      return NULL__null;
  }
  
  U_ASSERT(list.size() > 0)(void)0;
  Transliterator* t = NULL__null;
  
  if (list.size() > 1 || canonID.indexOf(ID_DELIM) >= 0) {
      // [NOTE: If it's a compoundID, we instantiate a CompoundTransliterator even if it only
      // has one child transliterator.  This is so that toRules() will return the right thing
      // (without any inactive ID), but our main ID still comes out correct.  That is, if we
      // instantiate "(Lower);Latin-Greek;", we want the rules to come out as "::Latin-Greek;"
      // even though the ID is "(Lower);Latin-Greek;".
      t = new CompoundTransliterator(list, parseError, status);
  }
  else {
      t = (Transliterator*)list.elementAt(0);
  }
  // Check null pointer
  if (t != NULL__null) {
      t->setID(canonID);
      if (lpGlobalFilter.isValid()) {
          t->adoptFilter(lpGlobalFilter.orphan());
      }
  }
  else if (U_SUCCESS(status)) {
      status = U_MEMORY_ALLOCATION_ERROR;
  }
  return t;
967}

969/**
* Create a transliterator from a basic ID.  This is an ID
* containing only the forward direction source, target, and
* variant.
* @param id a basic ID of the form S-T or S-T/V.
* @return a newly created Transliterator or null if the ID is
* invalid.
*/
977Transliterator* Transliterator::createBasicInstance(const UnicodeString& id,
                                                  const UnicodeString* canon) {
  UParseError pe;
  UErrorCode ec = U_ZERO_ERROR;
  TransliteratorAlias* alias = 0;
  Transliterator* t = 0;

  umtx_lockumtx_lock_71(&registryMutex);
  if (HAVE_REGISTRY(ec)(registry!=0 || initializeRegistry(ec))) {
      t = registry->get(id, alias, ec);
  }
  umtx_unlockumtx_unlock_71(&registryMutex);

  if (U_FAILURE(ec)) {
      delete t;
      delete alias;
      return 0;
  }

  // We may have not gotten a transliterator:  Because we can't
  // instantiate a transliterator from inside TransliteratorRegistry::
  // get() (that would deadlock), we sometimes pass back an alias.  This
  // contains the data we need to finish the instantiation outside the
  // registry mutex.  The alias may, in turn, generate another alias, so
  // we handle aliases in a loop.  The max times through the loop is two.
  // [alan]
  while (alias != 0) {
      U_ASSERT(t==0)(void)0;
      // Rule-based aliases are handled with TransliteratorAlias::
      // parse(), followed by TransliteratorRegistry::reget().
      // Other aliases are handled with TransliteratorAlias::create().
      if (alias->isRuleBased()) {
          // Step 1. parse
          TransliteratorParser parser(ec);
          alias->parse(parser, pe, ec);
          delete alias;
          alias = 0;

          // Step 2. reget
          umtx_lockumtx_lock_71(&registryMutex);
          if (HAVE_REGISTRY(ec)(registry!=0 || initializeRegistry(ec))) {
              t = registry->reget(id, parser, alias, ec);
          }
          umtx_unlockumtx_unlock_71(&registryMutex);

          // Step 3. Loop back around!
      } else {
          t = alias->create(pe, ec);
          delete alias;
          alias = 0;
          break;
      }
      if (U_FAILURE(ec)) {
          delete t;
          delete alias;
          t = NULL__null;
          break;
      }
  }

  if (t != NULL__null && canon != NULL__null) {
      t->setID(*canon);
  }

  return t;
1042}

1044/**
* Returns a <code>Transliterator</code> object constructed from
* the given rule string.  This will be a RuleBasedTransliterator,
* if the rule string contains only rules, or a
* CompoundTransliterator, if it contains ID blocks, or a
* NullTransliterator, if it contains ID blocks which parse as
* empty for the given direction.
*/
1052Transliterator* U_EXPORT2
1053Transliterator::createFromRules(const UnicodeString& ID,
                              const UnicodeString& rules,
                              UTransDirection dir,
                              UParseError& parseError,
                              UErrorCode& status)
1058{
  Transliterator* t = NULL__null;

  TransliteratorParser parser(status);
  parser.parse(rules, dir, parseError, status);

  if (U_FAILURE(status)) {
      return 0;
  }

  // NOTE: The logic here matches that in TransliteratorRegistry.
  if (parser.idBlockVector.size() == 0 && parser.dataVector.size() == 0) {
      t = new NullTransliterator();
  }
  else if (parser.idBlockVector.size() == 0 && parser.dataVector.size() == 1) {
      t = new RuleBasedTransliterator(ID, (TransliterationRuleData*)parser.dataVector.orphanElementAt(0), TRUE1);
  }
  else if (parser.idBlockVector.size() == 1 && parser.dataVector.size() == 0) {
      // idBlock, no data -- this is an alias.  The ID has
      // been munged from reverse into forward mode, if
      // necessary, so instantiate the ID in the forward
      // direction.
      if (parser.compoundFilter != NULL__null) {
          UnicodeString filterPattern;
          parser.compoundFilter->toPattern(filterPattern, FALSE0);
          t = createInstance(filterPattern + UnicodeString(ID_DELIM)
                  + *((UnicodeString*)parser.idBlockVector.elementAt(0)), UTRANS_FORWARD, parseError, status);
      }
      else
          t = createInstance(*((UnicodeString*)parser.idBlockVector.elementAt(0)), UTRANS_FORWARD, parseError, status);


      if (t != NULL__null) {
          t->setID(ID);
      }
  }
  else {
      UVector transliterators(status);
      // TODO ICU-21701 missing U_FAILURE check here.
      //      Error and nullptr checking through this whole block looks suspect.
      int32_t passNumber = 1;

      int32_t limit = parser.idBlockVector.size();
      if (parser.dataVector.size() > limit)
          limit = parser.dataVector.size();

      for (int32_t i = 0; i < limit; i++) {
          if (i < parser.idBlockVector.size()) {
              UnicodeString* idBlock = (UnicodeString*)parser.idBlockVector.elementAt(i);
              if (!idBlock->isEmpty()) {
                  Transliterator* temp = createInstance(*idBlock, UTRANS_FORWARD, parseError, status);
                  if (U_FAILURE(status)) {
                      delete temp;
                      return nullptr;
                  }
                  if (temp != NULL__null && typeid(*temp) != typeid(NullTransliterator)) {
                      transliterators.addElement(temp, status);
                      if (U_FAILURE(status)) {
                          delete temp;
                          return nullptr;
                      }
                  } else {
                      delete temp;
                  }
              }
          }
          if (!parser.dataVector.isEmpty()) {
              TransliterationRuleData* data = (TransliterationRuleData*)parser.dataVector.orphanElementAt(0);
              // TODO: Should passNumber be turned into a decimal-string representation (1 -> "1")?
              RuleBasedTransliterator* temprbt = new RuleBasedTransliterator(UnicodeString(CompoundTransliterator::PASS_STRING) + UnicodeString(passNumber++),
                      data, TRUE1);
              // Check if NULL before adding it to transliterators to avoid future usage of NULL pointer.
              if (temprbt == NULL__null) {
                  if (U_SUCCESS(status)) {
                      status = U_MEMORY_ALLOCATION_ERROR;
                  }
                  return t;
              }
              transliterators.addElement(temprbt, status);
              if (U_FAILURE(status)) {
                  delete temprbt;
                  return t;
              }
              // TODO: ICU-21701 the transliterators vector will leak its contents if anything goes wrong.
              //       Under normal operation, the CompoundTransliterator constructor adopts the
              //       the contents of the vector.
          }
      }

      t = new CompoundTransliterator(transliterators, passNumber - 1, parseError, status);
      // Null pointer check
      if (t != NULL__null) {
          t->setID(ID);
          t->adoptFilter(parser.orphanCompoundFilter());
      }
  }
  if (U_SUCCESS(status) && t == NULL__null) {
      status = U_MEMORY_ALLOCATION_ERROR;
  }
  return t;
1158}

1160UnicodeString& Transliterator::toRules(UnicodeString& rulesSource,
                                     UBool escapeUnprintable) const {
  // The base class implementation of toRules munges the ID into
  // the correct format.  That is: foo => ::foo
  if (escapeUnprintable) {
      rulesSource.truncate(0);
      UnicodeString id = getID();
      for (int32_t i=0; i<id.length();) {
          UChar32 c = id.char32At(i);
          if (!ICU_Utility::escapeUnprintable(rulesSource, c)) {
              rulesSource.append(c);
          }
          i += U16_LENGTH(c)((uint32_t)(c)<=0xffff ? 1 : 2);
      }
  } else {
      rulesSource = getID();
  }
  // KEEP in sync with rbt_pars
  rulesSource.insert(0, UNICODE_STRING_SIMPLE("::")icu::UnicodeString(true, u"::", -1));
  rulesSource.append(ID_DELIM);
  return rulesSource;
1181}

1183int32_t Transliterator::countElements() const {
  const CompoundTransliterator* ct = dynamic_cast<const CompoundTransliterator*>(this);
  return ct != NULL__null ? ct->getCount() : 0;
1186}

1188const Transliterator& Transliterator::getElement(int32_t index, UErrorCode& ec) const {
  if (U_FAILURE(ec)) {
      return *this;
  }
  const CompoundTransliterator* cpd = dynamic_cast<const CompoundTransliterator*>(this);
  int32_t n = (cpd == NULL__null) ? 1 : cpd->getCount();
  if (index < 0 || index >= n) {
      ec = U_INDEX_OUTOFBOUNDS_ERROR;
      return *this;
  } else {
      return (n == 1) ? *this : cpd->getTransliterator(index);
  }
1200}

1202UnicodeSet& Transliterator::getSourceSet(UnicodeSet& result) const {
  handleGetSourceSet(result);
  if (filter != NULL__null) {
      UnicodeSet* filterSet = dynamic_cast<UnicodeSet*>(filter);
      UBool deleteFilterSet = FALSE0;
      // Most, but not all filters will be UnicodeSets.  Optimize for
      // the high-runner case.
      if (filterSet == NULL__null) {
          filterSet = new UnicodeSet();
          // Check null pointer
          if (filterSet == NULL__null) {
              return result;
          }
          deleteFilterSet = TRUE1;
          filter->addMatchSetTo(*filterSet);
      }
      result.retainAll(*filterSet);
      if (deleteFilterSet) {
          delete filterSet;
      }
  }
  return result;
1224}

1226void Transliterator::handleGetSourceSet(UnicodeSet& result) const {
  result.clear();
1228}

1230UnicodeSet& Transliterator::getTargetSet(UnicodeSet& result) const {
  return result.clear();
1232}

1234// For public consumption
1235void U_EXPORT2 Transliterator::registerFactory(const UnicodeString& id,
                                   Transliterator::Factory factory,
                                   Transliterator::Token context) {
  Mutex lock(&registryMutex);
  UErrorCode ec = U_ZERO_ERROR;
  if (HAVE_REGISTRY(ec)(registry!=0 || initializeRegistry(ec))) {
      _registerFactory(id, factory, context);
  }
1243}

1245// To be called only by Transliterator subclasses that are called
1246// to register themselves by initializeRegistry().
1247void Transliterator::_registerFactory(const UnicodeString& id,
                                    Transliterator::Factory factory,
                                    Transliterator::Token context) {
  UErrorCode ec = U_ZERO_ERROR;
  registry->put(id, factory, context, TRUE1, ec);
1252}

1254// To be called only by Transliterator subclasses that are called
1255// to register themselves by initializeRegistry().
1256void Transliterator::_registerSpecialInverse(const UnicodeString& target,
                                           const UnicodeString& inverseTarget,
                                           UBool bidirectional) {
  UErrorCode status = U_ZERO_ERROR;
  TransliteratorIDParser::registerSpecialInverse(target, inverseTarget, bidirectional, status);
1261}

1263/**
* Registers a instance <tt>obj</tt> of a subclass of
* <code>Transliterator</code> with the system.  This object must
* implement the <tt>clone()</tt> method.  When
* <tt>getInstance()</tt> is called with an ID string that is
* equal to <tt>obj.getID()</tt>, then <tt>obj.clone()</tt> is
* returned.
*
* @param obj an instance of subclass of
* <code>Transliterator</code> that defines <tt>clone()</tt>
* @see #getInstance
* @see #unregister
*/
1276void U_EXPORT2 Transliterator::registerInstance(Transliterator* adoptedPrototype) {
  Mutex lock(&registryMutex);
  UErrorCode ec = U_ZERO_ERROR;
  if (HAVE_REGISTRY(ec)(registry!=0 || initializeRegistry(ec))) {
      _registerInstance(adoptedPrototype);
  }
1282}

1284void Transliterator::_registerInstance(Transliterator* adoptedPrototype) {
  UErrorCode ec = U_ZERO_ERROR;
  registry->put(adoptedPrototype, TRUE1, ec);
1287}

1289void U_EXPORT2 Transliterator::registerAlias(const UnicodeString& aliasID,
                                           const UnicodeString& realID) {
  Mutex lock(&registryMutex);
  UErrorCode ec = U_ZERO_ERROR;
  if (HAVE_REGISTRY(ec)(registry!=0 || initializeRegistry(ec))) {
      _registerAlias(aliasID, realID);
  }
1296}

1298void Transliterator::_registerAlias(const UnicodeString& aliasID,
                                  const UnicodeString& realID) {
  UErrorCode ec = U_ZERO_ERROR;
  registry->put(aliasID, realID, FALSE0, TRUE1, ec);
1302}

1304/**
* Unregisters a transliterator or class.  This may be either
* a system transliterator or a user transliterator or class.
* 
* @param ID the ID of the transliterator or class
* @see #registerInstance

*/
1312void U_EXPORT2 Transliterator::unregister(const UnicodeString& ID) {
  Mutex lock(&registryMutex);
  UErrorCode ec = U_ZERO_ERROR;
  if (HAVE_REGISTRY(ec)(registry!=0 || initializeRegistry(ec))) {
      registry->remove(ID);
  }
1318}

1320/**
* == OBSOLETE - remove in ICU 3.4 ==
* Return the number of IDs currently registered with the system.
* To retrieve the actual IDs, call getAvailableID(i) with
* i from 0 to countAvailableIDs() - 1.
*/
1326int32_t U_EXPORT2 Transliterator::countAvailableIDs(void) {
  int32_t retVal = 0;
  Mutex lock(&registryMutex);
  UErrorCode ec = U_ZERO_ERROR;
  if (HAVE_REGISTRY(ec)(registry!=0 || initializeRegistry(ec))) {
      retVal = registry->countAvailableIDs();
  }
  return retVal;
1334}

1336/**
* == OBSOLETE - remove in ICU 3.4 ==
* Return the index-th available ID.  index must be between 0
* and countAvailableIDs() - 1, inclusive.  If index is out of
* range, the result of getAvailableID(0) is returned.
*/
1342const UnicodeString& U_EXPORT2 Transliterator::getAvailableID(int32_t index) {
  const UnicodeString* result = NULL__null;
1
'result' initialized to a null pointer value→
  umtx_lockumtx_lock_71(&registryMutex);
  UErrorCode ec = U_ZERO_ERROR;
  if (HAVE_REGISTRY(ec)(registry!=0 || initializeRegistry(ec))) {
2
←
Assuming 'registry' is equal to null→
3
←
Assuming the condition is false→
4
←
Taking false branch→
      result = &registry->getAvailableID(index);
  }
  umtx_unlockumtx_unlock_71(&registryMutex);
  U_ASSERT(result != NULL)(void)0; // fail if no registry
  return *result;
5
←
Returning null reference
1352}

1354StringEnumeration* U_EXPORT2 Transliterator::getAvailableIDs(UErrorCode& ec) {
  if (U_FAILURE(ec)) return NULL__null;
  StringEnumeration* result = NULL__null;
  umtx_lockumtx_lock_71(&registryMutex);
  if (HAVE_REGISTRY(ec)(registry!=0 || initializeRegistry(ec))) {
      result = registry->getAvailableIDs();
  }
  umtx_unlockumtx_unlock_71(&registryMutex);
  if (result == NULL__null) {
      ec = U_INTERNAL_TRANSLITERATOR_ERROR;
  }
  return result;
1366}

1368int32_t U_EXPORT2 Transliterator::countAvailableSources(void) {
  Mutex lock(&registryMutex);
  UErrorCode ec = U_ZERO_ERROR;
  return HAVE_REGISTRY(ec)(registry!=0 || initializeRegistry(ec)) ? _countAvailableSources() : 0;
1372}

1374UnicodeString& U_EXPORT2 Transliterator::getAvailableSource(int32_t index,
                                                UnicodeString& result) {
  Mutex lock(&registryMutex);
  UErrorCode ec = U_ZERO_ERROR;
  if (HAVE_REGISTRY(ec)(registry!=0 || initializeRegistry(ec))) {
      _getAvailableSource(index, result);
  }
  return result;
1382}

1384int32_t U_EXPORT2 Transliterator::countAvailableTargets(const UnicodeString& source) {
  Mutex lock(&registryMutex);
  UErrorCode ec = U_ZERO_ERROR;
  return HAVE_REGISTRY(ec)(registry!=0 || initializeRegistry(ec)) ? _countAvailableTargets(source) : 0;
1388}

1390UnicodeString& U_EXPORT2 Transliterator::getAvailableTarget(int32_t index,
                                                const UnicodeString& source,
                                                UnicodeString& result) {
  Mutex lock(&registryMutex);
  UErrorCode ec = U_ZERO_ERROR;
  if (HAVE_REGISTRY(ec)(registry!=0 || initializeRegistry(ec))) {
      _getAvailableTarget(index, source, result);
  }
  return result;
1399}

1401int32_t U_EXPORT2 Transliterator::countAvailableVariants(const UnicodeString& source,
                                             const UnicodeString& target) {
  Mutex lock(&registryMutex);
  UErrorCode ec = U_ZERO_ERROR;
  return HAVE_REGISTRY(ec)(registry!=0 || initializeRegistry(ec)) ? _countAvailableVariants(source, target) : 0;
1406}

1408UnicodeString& U_EXPORT2 Transliterator::getAvailableVariant(int32_t index,
                                                 const UnicodeString& source,
                                                 const UnicodeString& target,
                                                 UnicodeString& result) {
  Mutex lock(&registryMutex);
  UErrorCode ec = U_ZERO_ERROR;
  if (HAVE_REGISTRY(ec)(registry!=0 || initializeRegistry(ec))) {
      _getAvailableVariant(index, source, target, result);
  }
  return result;
1418}

1420int32_t Transliterator::_countAvailableSources(void) {
  return registry->countAvailableSources();
1422}

1424UnicodeString& Transliterator::_getAvailableSource(int32_t index,
                                                UnicodeString& result) {
  return registry->getAvailableSource(index, result);
1427}

1429int32_t Transliterator::_countAvailableTargets(const UnicodeString& source) {
  return registry->countAvailableTargets(source);
1431}

1433UnicodeString& Transliterator::_getAvailableTarget(int32_t index,
                                                const UnicodeString& source,
                                                UnicodeString& result) {
  return registry->getAvailableTarget(index, source, result);
1437}

1439int32_t Transliterator::_countAvailableVariants(const UnicodeString& source,
                                             const UnicodeString& target) {
  return registry->countAvailableVariants(source, target);
1442}

1444UnicodeString& Transliterator::_getAvailableVariant(int32_t index,
                                                 const UnicodeString& source,
                                                 const UnicodeString& target,
                                                 UnicodeString& result) {
  return registry->getAvailableVariant(index, source, target, result);
1449}

1451#ifdef U_USE_DEPRECATED_TRANSLITERATOR_API

1453/**
* Method for subclasses to use to obtain a character in the given
* string, with filtering.
* @deprecated the new architecture provides filtering at the top
* level.  This method will be removed Dec 31 2001.
*/
1459UChar Transliterator::filteredCharAt(const Replaceable& text, int32_t i) const {
  UChar c;
  const UnicodeFilter* localFilter = getFilter();
  return (localFilter == 0) ? text.charAt(i) :
      (localFilter->contains(c = text.charAt(i)) ? c : (UChar)0xFFFE);
1464}

1466#endif

1468/**
* If the registry is initialized, return TRUE.  If not, initialize it
* and return TRUE.  If the registry cannot be initialized, return
* FALSE (rare).
*
* IMPORTANT: Upon entry, registryMutex must be LOCKED.  The entire
* initialization is done with the lock held.  There is NO REASON to
* unlock, since no other thread that is waiting on the registryMutex
* cannot itself proceed until the registry is initialized.
*/
1478UBool Transliterator::initializeRegistry(UErrorCode &status) {
  if (registry != 0) {
      return TRUE1;
  }

  registry = new TransliteratorRegistry(status);
  if (registry == 0 || U_FAILURE(status)) {
      delete registry;
      registry = 0;
      return FALSE0; // can't create registry, no recovery
  }

  /* The following code parses the index table located in
   * icu/data/translit/root.txt.  The index is an n x 4 table
   * that follows this format:
   *  <id>{
   *      file{
   *          resource{"<resource>"}
   *          direction{"<direction>"}
   *      }
   *  }
   *  <id>{
   *      internal{
   *          resource{"<resource>"}
   *          direction{"<direction"}
   *       }
   *  }
   *  <id>{
   *      alias{"<getInstanceArg"}
   *  }
   * <id> is the ID of the system transliterator being defined.  These
   * are public IDs enumerated by Transliterator.getAvailableIDs(),
   * unless the second field is "internal".
   * 
   * <resource> is a ResourceReader resource name.  Currently these refer
   * to file names under com/ibm/text/resources.  This string is passed
   * directly to ResourceReader, together with <encoding>.
   * 
   * <direction> is either "FORWARD" or "REVERSE".
   * 
   * <getInstanceArg> is a string to be passed directly to
   * Transliterator.getInstance().  The returned Transliterator object
   * then has its ID changed to <id> and is returned.
   *
   * The extra blank field on "alias" lines is to make the array square.
   */
  //static const char translit_index[] = "translit_index";

  UErrorCode lstatus = U_ZERO_ERROR;
  UResourceBundle *bundle, *transIDs, *colBund;
  bundle = ures_openures_open_71(U_ICUDATA_TRANSLIT"icudt" "71" "l" "-" "translit", NULL__null/*open default locale*/, &lstatus);
  transIDs = ures_getByKeyures_getByKey_71(bundle, RB_RULE_BASED_IDS, 0, &lstatus);
  const UnicodeString T_PART = UNICODE_STRING_SIMPLE("-t-")icu::UnicodeString(true, u"-t-", -1);

  int32_t row, maxRows;
  if (lstatus == U_MEMORY_ALLOCATION_ERROR) {
      delete registry;
      registry = nullptr;
      status = U_MEMORY_ALLOCATION_ERROR;
      return FALSE0;
  }
  if (U_SUCCESS(lstatus)) {
      maxRows = ures_getSizeures_getSize_71(transIDs);
      for (row = 0; row < maxRows; row++) {
          colBund = ures_getByIndexures_getByIndex_71(transIDs, row, 0, &lstatus);
          if (U_SUCCESS(lstatus)) {
              UnicodeString id(ures_getKeyures_getKey_71(colBund), -1, US_INVicu::UnicodeString::kInvariant);
              if(id.indexOf(T_PART) != -1) {
                  ures_closeures_close_71(colBund);
                  continue;
              }
              UResourceBundle* res = ures_getNextResourceures_getNextResource_71(colBund, NULL__null, &lstatus);
              const char* typeStr = ures_getKeyures_getKey_71(res);
              UChar type;
              u_charsToUCharsu_charsToUChars_71(typeStr, &type, 1);

              if (U_SUCCESS(lstatus)) {
                  int32_t len = 0;
                  const UChar *resString;
                  switch (type) {
                  case 0x66: // 'f'
                  case 0x69: // 'i'
                      // 'file' or 'internal';
                      // row[2]=resource, row[3]=direction
                      {
                          
                          resString = ures_getStringByKeyures_getStringByKey_71(res, "resource", &len, &lstatus);
                          UBool visible = (type == 0x0066 /*f*/);
                          UTransDirection dir = 
                              (ures_getUnicodeStringByKey(res, "direction", &lstatus).charAt(0) ==
                               0x0046 /*F*/) ?
                              UTRANS_FORWARD : UTRANS_REVERSE;
                          registry->put(id, UnicodeString(TRUE1, resString, len), dir, TRUE1, visible, lstatus);
                      }
                      break;
                  case 0x61: // 'a'
                      // 'alias'; row[2]=createInstance argument
                      resString = ures_getStringures_getString_71(res, &len, &lstatus);
                      registry->put(id, UnicodeString(TRUE1, resString, len), TRUE1, TRUE1, lstatus);
                      break;
                  }
              }
              ures_closeures_close_71(res);
          }
          ures_closeures_close_71(colBund);
      }
  }

  ures_closeures_close_71(transIDs);
  ures_closeures_close_71(bundle);

  // Manually add prototypes that the system knows about to the
  // cache.  This is how new non-rule-based transliterators are
  // added to the system.
  
  // This is to allow for null pointer check
  NullTransliterator* tempNullTranslit = new NullTransliterator();
  LowercaseTransliterator* tempLowercaseTranslit = new LowercaseTransliterator();
  UppercaseTransliterator* tempUppercaseTranslit = new UppercaseTransliterator();
  TitlecaseTransliterator* tempTitlecaseTranslit = new TitlecaseTransliterator();
  UnicodeNameTransliterator* tempUnicodeTranslit = new UnicodeNameTransliterator();
  NameUnicodeTransliterator* tempNameUnicodeTranslit = new NameUnicodeTransliterator();
1600#if !UCONFIG_NO_BREAK_ITERATION0
   // TODO: could or should these transliterators be referenced polymorphically once constructed?
   BreakTransliterator* tempBreakTranslit         = new BreakTransliterator();
1603#endif
  // Check for null pointers
  if (tempNullTranslit == NULL__null || tempLowercaseTranslit == NULL__null || tempUppercaseTranslit == NULL__null ||
      tempTitlecaseTranslit == NULL__null || tempUnicodeTranslit == NULL__null || 
1607#if !UCONFIG_NO_BREAK_ITERATION0
      tempBreakTranslit == NULL__null ||
1609#endif
      tempNameUnicodeTranslit == NULL__null )
  {
      delete tempNullTranslit;
      delete tempLowercaseTranslit;
      delete tempUppercaseTranslit;
      delete tempTitlecaseTranslit;
      delete tempUnicodeTranslit;
      delete tempNameUnicodeTranslit;
1618#if !UCONFIG_NO_BREAK_ITERATION0
      delete tempBreakTranslit;
1620#endif
      // Since there was an error, remove registry
      delete registry;
      registry = NULL__null;

      status = U_MEMORY_ALLOCATION_ERROR;
      return 0;
  }

  registry->put(tempNullTranslit, TRUE1, status);
  registry->put(tempLowercaseTranslit, TRUE1, status);
  registry->put(tempUppercaseTranslit, TRUE1, status);
  registry->put(tempTitlecaseTranslit, TRUE1, status);
  registry->put(tempUnicodeTranslit, TRUE1, status);
  registry->put(tempNameUnicodeTranslit, TRUE1, status);
1635#if !UCONFIG_NO_BREAK_ITERATION0
  registry->put(tempBreakTranslit, FALSE0, status);   // FALSE means invisible.
1637#endif

  RemoveTransliterator::registerIDs(); // Must be within mutex
  EscapeTransliterator::registerIDs();
  UnescapeTransliterator::registerIDs();
  NormalizationTransliterator::registerIDs();
  AnyTransliterator::registerIDs();

  _registerSpecialInverse(UNICODE_STRING_SIMPLE("Null")icu::UnicodeString(true, u"Null", -1),
                          UNICODE_STRING_SIMPLE("Null")icu::UnicodeString(true, u"Null", -1), FALSE0);
  _registerSpecialInverse(UNICODE_STRING_SIMPLE("Upper")icu::UnicodeString(true, u"Upper", -1),
                          UNICODE_STRING_SIMPLE("Lower")icu::UnicodeString(true, u"Lower", -1), TRUE1);
  _registerSpecialInverse(UNICODE_STRING_SIMPLE("Title")icu::UnicodeString(true, u"Title", -1),
                          UNICODE_STRING_SIMPLE("Lower")icu::UnicodeString(true, u"Lower", -1), FALSE0);

  ucln_i18n_registerCleanupucln_i18n_registerCleanup_71(UCLN_I18N_TRANSLITERATOR, utrans_transliterator_cleanuputrans_transliterator_cleanup_71);

  return TRUE1;
1655}

1657U_NAMESPACE_END}

1659// Defined in transreg.h:

1661/**
* Release all static memory held by transliterator.  This will
* necessarily invalidate any rule-based transliterators held by the
* user, because RBTs hold pointers to common data objects.
*/
1666U_CFUNCextern "C" UBool utrans_transliterator_cleanuputrans_transliterator_cleanup_71(void) {
  U_NAMESPACE_USEusing namespace icu_71;
  TransliteratorIDParser::cleanup();
  if (registry) {
      delete registry;
      registry = NULL__null;
  }
  return TRUE1;
1674}

1676#endif /* #if !UCONFIG_NO_TRANSLITERATION */

1678//eof