include "convcs.m";
convcs := load Convcs Convcs->PATH;

Btos: module {
	init: fn(arg: string): string;
	btos: fn(s: Convcs->State, b: array of byte, nchars: int)
			: (Convcs->State, string, int);
};

Stob: module {
	init: fn (arg: string): string;
	stob: fn(s: Convcs->State, str: string)
			: (Convcs->State, array of byte);
};

Convcs: module {
	State: type string;
	Startstate: con "";

	init: fn(csfile: string): string;
	getbtos: fn(cs: string): (Btos, string);
	getstob: fn(cs: string): (Stob, string);
	enumcs: fn(): list of (string, string, int);
	aliases: fn(cs: string): (string, list of string);
};

The Convcs module provides an entry point to the suite, mapping character set names and aliases to their associated converter implementation.

The Convcs module

init(csfile)

Init should be called once to initialise the internal state of Convcs. The csfile argument specifies the path of the converter mapping file. If this argument is nil, the default mapping file /lib/convcs/charsets is used.

getbtos(cs)

Getbtos returns an initialised Btos module for converting from the requested encoding.

The return value is a tuple, holding the module reference and an error string. If any errors were encountered in locating, loading or initialising the requested converter, the module reference will be nil and the string will contain an explanation.

The character set name, cs, is normalised by mapping all upper-case latin1 characters to lower-case before comparison with character set names and aliases from the charsets file.

getstob(cs)

Getstob returns an initialised Stob module for converting from strings to the requested encoding. Apart from the different module type, the return value and semantics are the same as for getbtos().

enumcs()

Returns a list describing the set of available converters. The list is comprised of (name, desc, mode) tuples.

Name is the standard name of the character set.

Desc is a more friendly description taken directly from the desc attribute given in the charsets file. If the attribute is not given then desc is set to name.

Mode is a bitmap that details the converters available for the given character set. Valid mode bits are Convcs->BTOS and Convcs->STOB. For convenience, Convcs->BOTH is also defined.

aliases(cs)

Returns the aliases for the character set name cs. The return value is the tuple (desc, aliases).

Desc is the descriptive text for the character set, as returned by enumcs().

Aliases lists all the known aliases for cs. The first name in the list is the default character set name - the name that all the other aliases refer to. If the aliases list is nil then there was an error in looking up the character set name and desc will detail the error.

Using a Btos converter

The Btos module returned by getbtos() is already initialised and is ready to start the conversion. Conversions can be made on a individual basis, or in a `streamed' mode.

converter->btos(s, b, nchars)

Converts raw byte codes of the character set encoding to a Limbo string.

The argument s is a converter state as returned from the previous call to btos on the same input stream. The first call to btos on a particular input stream should give Convcs->Startstate (or nil) as the value for s. The argument b is the bytes to be converted. The argument nchars is the maximal length of the string to be returned. If this argument is -1 then as much of b will be consumed as possible. A value of 0 indicates to the converter that there is no more data and that any pending state should be flushed.

The return value of btos is the tuple (state, str, nbytes) where state is the new state of the converter, str is the converted string, and nbytes is the number of bytes from b consumed by the conversion.

The same converter module can be used for multiple conversion streams by maintaining a separate state variable for each stream.

Using an Stob converter

The Stob module returned by getstob() is already initialised and is ready to start the conversion.

converter->stob(s, str)

Converts the limbo string str to the raw byte codes of the character set encoding. The argument s represents the converter state and is treated in the same way as for converter->btos() described above. To terminate a conversion stob should be called with an emtpy string in order to flush the converter state.

The return value of stob is the tuple (state, bytes) where state is the new state of the converter and bytes is the result of the conversion.

Conversion errors

When using converter->btos() to convert data to Limbo strings, any byte sequences that are not valid for the specific character encoding scheme will be converted to the Unicode error character 16rFFFD.

When using converter->stob() to convert Limbo strings, any Unicode characters that can not be mapped into the character set will normally be substituted by the US-ASCII code for `?'. Note that this may be inappropriate for certain conversions, such converters will use a suitable error character for their particular character set and encoding scheme.

Charset file format

The file /lib/convcs/charsets provides the mapping between character set names and their implementation modules. The file format conforms to that supported by cfg(2). The following description relies on terms defined in the cfg(2) manual page.

Each record name defines a character set name. If the primary value of the record is non-empty then the name is an alias, the value being the real name. An alias record must point to an actual converter record, not to another alias, as Convcs only follows one level of aliasing.

Each converter record consists of a set of tuples with the following primary attributes:

desc

A more descriptive name for the character set encoding. This attribute is used for the description returned by enumcs().

btos

The path to the Btos module implementation of this converter.

stob

The path to the Stob module implementation of this converter.

Both the btos and stob tuples can have an optional arg attribute which is passed to the init() function of the converter when initialised by Convcs. If a converter record has neither an stob nor a btos tuple, then it is ignored.

The following example is an extract from the standard Inferno charsets file:

cp866=ibm866
866=ibm866
ibm866=
    desc='Russian MS-DOS CP 866'
    stob=/dis/lib/convcs/cp_stob.dis    arg=/lib/convcs/ibm866.cp
    btos=/dis/lib/convcs/cp_btos.dis    arg=/lib/convcs/ibm866.cp

This entry defines Stob and Btos converters for the character set called ibm866. The converters are actually the generic codepage converters cp_stob and cp_btos paramaterized with a codepage file. The entry also defines the aliases cp866 and 866 for the name ibm866.

/lib/convcs/charsets

The default mapping between character set names and their implementation modules.

/lib/convcs/csname.cp

Codepage files for use by the generic codepage converters cp_stob and cp_btos. Each file consists of 256 unicode runes mapping codepage byte values to unicode by their index. The runes are stored in the utf-8 encoding.

/appl/lib/convcs/convcs.b

Implementation of the Convcs module.

/appl/lib/convcs/cp_btos.b

Generic Btos codepage converter.

/appl/lib/convcs/cp_stob.b

Generic Stob codepage converter.