#include <kmime_codecs.h>
Inherited by KMime::Base64Decoder, KMime::IdentityEnDecoder, KMime::QuotedPrintableDecoder, and KMime::UUDecoder.
Public Member Functions | |
| virtual bool | decode (const char *&scursor, const char *const send, char *&dcursor, const char *const dend)=0 |
| virtual bool | finish (char *&dcursor, const char *const dend)=0 |
Protected Member Functions | |
| Decoder (bool withCRLF=false) | |
Protected Attributes | |
| const bool | mWithCRLF |
Friends | |
| class | Codec |
Detailed Description
Stateful decoder class, modelled after TQTextDecoder.
Overview
KMime decoders are designed to be able to process encoded data in chunks of arbitrary size and to work with output buffers of also arbitrary size. They maintain any state necessary to go on where the previous call left off.
The class consists of only two methods of interest: see decode, which decodes an input block and finalize, which flushes any remaining data to the output stream.
Typically, you will create a decoder instance, call decode as often as necessary, then call finalize (most often a single call suffices, but it might be that during that call the output buffer is filled, so you should be prepared to call finalize as often as necessary, ie. until it returns true).
Values
Both methods return true to indicate that they've finished their job. For decode, a return value of true means that the current input block has been finished (false most often means that the output buffer is full, but that isn't required behavior. The decode call is free to return at arbitrary times during processing).
For finalize, a return value of true means that all data implicitly or explicitly stored in the decoder instance has been flushed to the output buffer. A false return value should be interpreted as "check if the output buffer is full and call me
again", just as with decode.
Pattern
Since the decoder maintains state, you can only use it once. After a sequence of input blocks has been processed, you finalize the output and then delete the decoder instance. If you want to process another input block sequence, you create a new instance.
Typical usage (in contains the (base64-encoded) input data), taking into account all the conventions detailed above:
KMime::Codec * codec = KMime::Codec::codecForName( "base64" ); kdFatal( !codec ) << "No codec found for base64!" << endl; KMime::Decoder * dec = codec->makeDecoder(); assert( dec ); // should not happen TQByteArray out( 256 ); // small buffer is enough ;-) TQByteArray::Iterator iit = in.begin(); TQByteArray::Iterator oit = out.begin(); // decode the chunk while ( !dec->decode( iit, in.end(), oit, out.end() ) ) if ( oit == out.end() ) { // output buffer full, process contents do_something_with( out ); oit = out.begin(); } // repeat while loop for each input block // ... // finish (flush remaining data from decoder): while ( !dec->finish( oit, out.end() ) ) if ( oit == out.end() ) { // output buffer full, process contents do_something_with( out ); oit = out.begin(); } // now process last chunk: out.resize( oit - out.begin() ); do_something_with( out ); // _delete_ the decoder, but not the codec: delete dec;
Stateful CTE decoder class
Definition at line 268 of file kmime_codecs.h.
Constructor & Destructor Documentation
◆ Decoder()
|
inlineprotected |
Protected constructor.
Use KMime::Codec::makeDecoder to create an instance. The bool parameter determines whether lines end with CRLF (true) or LF (false, default).
Definition at line 276 of file kmime_codecs.h.
Member Function Documentation
◆ decode()
|
pure virtual |
Decode a chunk of data, maintaining state information between calls.
See class decumentation for calling conventions.
◆ finish()
|
pure virtual |
Call this method to finalize the output stream.
Writes all remaining data and resets the decoder. See KMime::Codec for calling conventions.
The documentation for this class was generated from the following file: