The decode grouping of functions (decode, decode_to, and decode_into) perform the task of doing bulk decoding from an input of code_units to the encoding's code_point type.
There are 3 named functions for this behavior, and each function comes with several function overloads. Each named function produces increasingly more information, letting you opt into just how much information and control you'd like over the algorithm and behavior. The first one simply returns a container with the transformation applied, discarding much of the operation's result information. This is useful for quick, one-off conversions where you do not care about any errors and would rather let it be handled by the error handler. The second _to suffixed functions return a container within a result type that contains additional information. The final _into suffixed functions take an output range to write into, letting you explicitly control just how much space there is to write into as well as returning a detailed result type.
The return type for these function calls is one of:
- the desired output container (highest level);
ztd::text::decode_result </api/decode_result>orztd::text::stateless_decode_result </api/decode_result>with the desired output container embedded as the .output parameter (mid level); or,ztd::text::decode_result </api/decode_result>orztd::text::stateless_decode_result </api/decode_result>returning just the input and output ranges (lowest level).
This is the highest level bulk function.
This set of function overloads takes the provided input, encoding, handler and state and produces an output container type. The default container will either be a std::basic_string of the code_point type, or a std::vector if it is not a known "character" type.
The container type can be specified by passing it as an explicit template parameter to this function, such as ztd::text::decode<std::vector<char32_t>>("bark", ztd::text::ascii{});. The output container is default constructed.
It will either call push_back/insert directly on the target container to fill it up, or serialize data to a temporary buffer (controlled by ZTD_TEXT_INTERMEDIATE_TRANSCODE_BUFFER_BYTE_SIZE <config-ZTD_TEXT_INTERMEDIATE_TRANSCODE_BUFFER_BYTE_SIZE>) before then copying it into the desired output container through any available means (bulk .insert, repeated .push_back, or repeated single .insert with the .cend() iterator in that order).
This is the "fire and forget" version of the decode function, returning only the container and not returning any of the result or state information used to construct it.
This is the mid level bulk function.
This set of function overloads takes the provided input, encoding, handler and state and produces an output container type that is embedded within a ztd::text::decode_result </api/decode_result>, or a ztd::text::stateless_decode_result </api/stateless_decode_result>, depending on whether or not you called the version which takes a ztd::text::decode_state_t\<Encoding\> </api/decode_state> as a parameter or if it had to create one on the stack internally and discard it after the operation was finished.
The container type can be specified by passing it as an explicit template parameter to this function, such as ztd::text::decode_to<std::u32string>("meow", ztd::text::ascii{});. The output container is default constructed.
It will either call push_back/insert directly on the target container to fill it up, or serialize data to a temporary buffer (controlled by ZTD_TEXT_INTERMEDIATE_TRANSCODE_BUFFER_BYTE_SIZE <config-ZTD_TEXT_INTERMEDIATE_TRANSCODE_BUFFER_BYTE_SIZE>) before then copying it into the desired output container through any available means (bulk .insert, repeated .push_back, or repeated single .insert with the .cend() iterator in that order).
If nothing goes wrong or the error handler lets the algorithm continue, .input on the result should be empty.
This is the lowest level bulk function.
This set of function overloads takes the provided input, encoding, output, handler, and state and writes data into the output range specified by output. The result is a ztd::text::decode_result </api/decode_result>, or a ztd::text::stateless_decode_result </api/stateless_decode_result>, depending on whether or not you called the version which takes a ztd::text::decode_state_t\<Encoding\> </api/decode_state> as a parameter or if it had to create one on the stack internally and discard it after the operation was finished.
It is up to the end-user to provide a suitably-sized output range for output, otherwise this operation may return with ztd::text::encoding_error::insufficient_output </api/encoding_error>. for the result's error_code member. The amount of space consumed can be determined by checking the std::distance between the .begin() of the original output parameter and the .begin() of the returned .output member. The result also has error information and an .input member for checking how much input was consumed.
If nothing goes wrong or the error handler lets the algorithm continue, .input on the result should be empty.
All named functions have 4 overloads. Each of the "higher level" functions, at the end of their overload call chain, will call the lower-level decode_into to perform the work. The final decode_into call uses the following ordering of extension points into calling the base implementation:
text_decode_into(input, encoding, output, handler, state)- An internal, implementation-defined customization point.
basic_decode_into
The base function call, basic_decode_into, simply performs the core decode loop </design/converting/decode> using the Lucky 7 </design/lucky 7> design. This design also means minimal stack space is used, keeping the core algorithm suitable for resource-constrained devices.
Note
👉 This means that if you implement none of the extension points whatsoever, implementing the basic decode_one function on your Encoding Object type will guarantee a proper, working implementation.
Note
👉 If you need to call the "basic" form of this function that takes no secret implementation shortcuts or user-defined extension points, then call basic_decode_into directly. This can be useful to stop infinity loops when your extension points cannot handle certain inputs and thereby needs to "delegate" to the basic case.
ztd_text_decode