Create a new character set mapping instance using based on a trie table. Charmap
instances map strings to
other character sets. The charsets can be of any type, single-byte, multi-byte,
shifting, etc.
All mappings are done to or from Unicode in the UTF-16 encoding, which is the base
character set and encoding used by Javascript itself. In order to convert
between two non-Unicode character sets, you must chain two charmap instances together
to first map to Unicode and then back to the second charset.
The options parameter controls which mapping is constructed and its behaviours. The
current list of supported options are:
- charset - the name of the native charset to map to or from. This can be
given as an Charset instance or as a string that contains any commonly used name
for the character set, which is normalized to a standard IANA name.
If a name is not given, this class will default to the Western European character
set called ISO-8859-15.
- missing - specify what to do if a mapping is missing for a particular
character. For example, if you are mapping Unicode characters to a particular native
character set that does not support particular Unicode characters, the mapper will
follow the behaviour specified in this property. Valid values are:
- skip - skip any characters that do not exist in the target charset
- placeholder - put a static placeholder character in the output string
wherever there is an unknown character in the input string. Use the placeholder
parameter to specify which character to use in this case
- escape - use an escape sequence to represent the unknown character
The default value for the missing property if not otherwise specified is "escape"
so that information is not lost.
- placeholder - specify the placeholder character to use when the
mapper cannot map a particular input character to the output string. If this
option is not specified, then the '?' (question mark) character is used where
possible.
- escapeStyle - what style of escape sequences should be used to
escape unknown characters in the input when mapping to native, and what
style of espcae sequences should be parsed when mapping to Unicode. Valid
values are:
- html - Escape the characters as HTML entities. This would use
the standard HTML 5.0 (or later) entity names where possible, and numeric
entities in all other cases. Eg. an "e" with an acute accent would be
"é"
- js - Use the Javascript escape style. Eg. an "e" with an acute
accent would be "\u00E9". This can also be specified as "c#" as
it uses a similar escape syntax.
- c - Use the C/C++ escape style, which is similar to the the
Javascript style, but uses an "x" in place of the "u". Eg. an "e" with an
acute accent would be "\x00E9". This can also be specified as "c++".
- java - Use the Java escape style. This is very similar to the
the Javascript style, but the backslash has to be escaped twice. Eg. an
"e" with an acute accent would be "\\u00E9". This can also be specified
as "ruby", as Ruby uses a similar escape syntax with double backslashes.
- perl - Use the Perl escape style. Eg. an "e" with an acute
accent would be "\N{U+00E9}"
The default if this style is not specified is "js" for Javascript.
- onLoad - a callback function to call when this object is fully
loaded. When the onLoad option is given, this class will attempt to
load any missing data using the ilib loader callback.
When the constructor is done (even if the data is already preassembled), the
onLoad function is called with the current instance as a parameter, so this
callback can be used with preassembled or dynamic loading or a mix of the two.
- sync - tell whether to load any missing data synchronously or
asynchronously. If this option is given as "false", then the "onLoad"
callback must be given, because the instance returned from this constructor will
not be usable for a while.
- loadParams - an object containing parameters to pass to the
loader callback function when data is missing. The parameters are not
interpretted or modified in any way. They are simply passed along. The object
may contain any property/value pairs as long as the calling code is in
agreement with the loader callback function as to what those parameters mean.
If this copy of ilib is pre-assembled and all the data is already available,
or if the data was already previously loaded, then this constructor will call
the onLoad callback immediately when the initialization is done.
If the onLoad option is not given, this class will only attempt to load any
missing data synchronously.