Class CompoundTransliterator

java.lang.Object
com.ibm.icu.text.Transliterator
com.ibm.icu.text.CompoundTransliterator
All Implemented Interfaces:
StringTransform, Transform<String,String>

class CompoundTransliterator extends Transliterator
A transliterator that is composed of two or more other transliterator objects linked together. For example, if one transliterator transliterates from script A to script B, and another transliterates from script B to script C, the two may be combined to form a new transliterator from A to C.

Composed transliterators may not behave as expected. For example, inverses may not combine to form the identity transliterator. See the class documentation for Transliterator for details.

Copyright © IBM Corporation 1999. All rights reserved.

  • Field Details

    • trans

      private Transliterator[] trans
    • numAnonymousRBTs

      private int numAnonymousRBTs
  • Constructor Details

    • CompoundTransliterator

      CompoundTransliterator(List<Transliterator> list)
      Package private constructor for Transliterator from a vector of transliterators. The caller is responsible for fixing up the ID.
    • CompoundTransliterator

      CompoundTransliterator(List<Transliterator> list, int numAnonymousRBTs)
    • CompoundTransliterator

      CompoundTransliterator(String id, UnicodeFilter filter2, Transliterator[] trans2, int numAnonymousRBTs2)
      Internal method for safeClone...
      Parameters:
      id -
      filter2 -
      trans2 -
      numAnonymousRBTs2 -
  • Method Details

    • init

      private void init(List<Transliterator> list, int direction, boolean fixReverseID)
      Finish constructing a transliterator: only to be called by constructors. Before calling init(), set trans and filter to NULL.
      Parameters:
      list - a vector of transliterator objects to be adopted. It should NOT be empty. The list should be in declared order. That is, it should be in the FORWARD order; if direction is REVERSE then the list order will be reversed.
      direction - either FORWARD or REVERSE
      fixReverseID - if true, then reconstruct the ID of reverse entries by calling getID() of component entries. Some constructors do not require this because they apply a facade ID anyway.
    • getCount

      public int getCount()
      Returns the number of transliterators in this chain.
      Returns:
      number of transliterators in this chain.
    • getTransliterator

      public Transliterator getTransliterator(int index)
      Returns the transliterator at the given index in this chain.
      Parameters:
      index - index into chain, from 0 to getCount() - 1
      Returns:
      transliterator at the given index
    • _smartAppend

      private static void _smartAppend(StringBuilder buf, char c)
      Append c to buf, unless buf is empty or buf already ends in c.
    • toRules

      public String toRules(boolean escapeUnprintable)
      Override Transliterator: Create a rule string that can be passed to createFromRules() to recreate this transliterator.
      Overrides:
      toRules in class Transliterator
      Parameters:
      escapeUnprintable - if true then convert unprintable character to their hex escape representations, \\uxxxx or \\Uxxxxxxxx. Unprintable characters are those other than U+000A, U+0020..U+007E.
      Returns:
      the rule string
    • addSourceTargetSet

      public void addSourceTargetSet(UnicodeSet filter, UnicodeSet sourceSet, UnicodeSet targetSet)
      Description copied from class: Transliterator
      Returns the set of all characters that may be generated as replacement text by this transliterator, filtered by BOTH the input filter, and the current getFilter().

      SHOULD BE OVERRIDDEN BY SUBCLASSES. It is probably an error for any transliterator to NOT override this, but we can't force them to for backwards compatibility.

      Other methods vector through this.

      When gathering the information on source and target, the compound transliterator makes things complicated. For example, suppose we have:

       Global FILTER = [ax]
       a > b;
       :: NULL;
       b > c;
       x > d;
       
      While the filter just allows a and x, b is an intermediate result, which could produce c. So the source and target sets cannot be gathered independently. What we have to do is filter the sources for the first transliterator according to the global filter, intersect that transliterator's filter. Based on that we get the target. The next transliterator gets as a global filter (global + last target). And so on.

      There is another complication:

       Global FILTER = [ax]
       a >|b;
       b >c;
       
      Even though b would be filtered from the input, whenever we have a backup, it could be part of the input. So ideally we will change the global filter as we go.
      Overrides:
      addSourceTargetSet in class Transliterator
      targetSet - TODO
      See Also:
    • handleTransliterate

      protected void handleTransliterate(Replaceable text, Transliterator.Position index, boolean incremental)
      Specified by:
      handleTransliterate in class Transliterator
      Parameters:
      text - the buffer holding transliterated and untransliterated text
      index - the indices indicating the start, limit, context start, and context limit of the text.
      incremental - if true, assume more text may be inserted at pos.limit and act accordingly. Otherwise, transliterate all text between pos.start and pos.limit and move pos.start up to pos.limit.
      See Also:
    • computeMaximumContextLength

      private void computeMaximumContextLength()
      Compute and set the length of the longest context required by this transliterator. This is preceding context.
    • safeClone

      public Transliterator safeClone()
      Temporary hack for registry problem. Needs to be replaced by better architecture.