Table of Contents

1. Introduction
2. Major Additions
2.1. Instance Tags
2.2. Asynchronous Private Key Generation
2.3. Extra Symmetric Key
2.4. Convert Operations
2.5. SMP, Error, and Message Event Callbacks
2.6. Fragmentation Changes
3. Required Changes
3.1. OtrlMessageAppOps Callbacks
3.1.1. Removed Operations
3.1.2. Added Operations
3.2. Instance Tags
3.3. Fragmentation Changes
3.4. Asynchronous Private Key Generation
3.5. Library Initialization

1. Introduction

This file contains information about the changes between the 3.2.0 and
the 4.0.0 APIs for libotr.  Note that applications compiled against
previous versions of OTR will not work with libotr 4.0.0.

2. Major Additions

This section describes the new features in OTR 4.0.0 along with a short
history or motivation for each.

2.1. Instance Tags

Clients generate instance tags that are intended to be persistent. If
the same client is logged into the same account from multiple locations,
the intention is that he or she will have different instance tags at
each location. OTR wire messages (fragmented and unfragmented) include
the source and destination instance tags. If a client receives a message
that lists a destination instance tag different from his own, the client
will discard it (and issue a callback notifying the application of the
event).

This avoids an issue on IM networks that always relay all messages to
all sessions of a client who is logged in multiple times. In this
situation, OTR clients can attempt to establish an OTR session
indefinitely if there are interleaving messages from each of the
sessions.

2.2. Asynchronous Private Key Generation

Key generation can happen in a separate thread without blocking an
application.

2.3. Extra Symmetric Key

An extra symmetric key is kept synchronized during a conversation with a
buddy.  Either side can send a signal that they wish to use this key for
some external purpose (e.g. things like a file transfer, in some other
channel of communication).

2.4. Convert Operations

There is now a callback that is made immediately before a message is
encrypted and immediately after a message is decrypted. This callback
can tweak the plaintext message as needed. For example, this could allow
an application to convert formatting on a message if this would normally
be done on the plaintext by some other entity while the message is in
transit.

2.5. SMP, Error, and Message Event Callbacks

To avoid hard-coded English phrases in libotr, operations which used to
pass back strings are replaced by operations that pass back event codes.

2.6. Fragmentation Changes

In libotr version 3.2.0, you would need to call otrl_message_sending()
to create an encrypted message, and then call fragment_and_send() to get
libotr to fragment and inject that message. In libotr 4.0.0, the
functionality of fragment_and_send() has been integrated into
otrl_message_sending().

3. Required Changes

3.1. OtrlMessageAppOps Callbacks

3.1.1. Removed Operations

/* Display a notification message for a particular accountname /
 * protocol / username conversation. */
void (*notify)(void *opdata, OtrlNotifyLevel level,
	const char *accountname, const char *protocol,
	const char *username, const char *title,
	const char *primary, const char *secondary);

The notify() operation was removed since it was used to pass in
hardcoded English strings. This has been replaced by error and message
event callbacks, described below, which pass event codes rather than
hardcoded strings.


/* Display an OTR control message for a particular accountname /
 * protocol / username conversation.  Return 0 if you are able to
 * successfully display it.  If you return non-0 (or if this
 * function is NULL), the control message will be displayed inline,
 * as a received message, or else by using the above notify()
 * callback. */
int (*display_otr_message)(void *opdata, const char *accountname,
	const char *protocol, const char *username, const char *msg);

The display_otr_message() operation was removed for the same reasons as
above for the notify() operation.


/* Return a newly allocated string containing a human-friendly name
 * for the given protocol id */
const char *(*protocol_name)(void *opdata, const char *protocol);

/* Deallocate a string allocated by protocol_name */
void (*protocol_name_free)(void *opdata, const char *protocol_name);

The above operations are no longer required, as they were used when
preparing messages shown to users.


/* Log a message.  The passed message will end in "\n". */
void (*log_message)(void *opdata, const char *message);

The log_message() operation was also replaced by message event
callbacks.

3.1.2. Added Operations

/* We received a request from the buddy to use the current "extra"
 * symmetric key.  The key will be passed in symkey, of length
 * OTRL_EXTRAKEY_BYTES.  The requested use, as well as use-specific
 * data will be passed so that the applications can communicate other
 * information (some id for the data transfer, for example). */
void (*received_symkey)(void *opdata, ConnContext *context,
	unsigned int use, const unsigned char *usedata,
	size_t usedatalen, const unsigned char *symkey);

This is called when a remote buddy has specified a use for the current
symmetric key. If your application does not use the extra symmetric key
it does not need to provide an implementation for this operation.


/* Return a string according to the error event. This string will then
 * be concatenated to an OTR header to produce an OTR protocol error
 * message. The following are the possible error events:
 * - OTRL_ERRCODE_ENCRYPTION_ERROR
 * 		occured while encrypting a message
 * - OTRL_ERRCODE_MSG_NOT_IN_PRIVATE
 * 		sent encrypted message to somebody who is not in
 * 		a mutual OTR session
 * - OTRL_ERRCODE_MSG_UNREADABLE
 *		sent an unreadable encrypted message
 * - OTRL_ERRCODE_MSG_MALFORMED
 * 		message sent is malformed */
const char *(*otr_error_message)(void *opdata, ConnContext *context,
	OtrlErrorCode err_code);

/* Deallocate a string returned by otr_error_message */
void (*otr_error_message_free)(void *opdata, const char *err_msg);

These methods are for producing human-readable error message that will
be sent to the remote buddy when one of these error conditions occurs.
They will be appended to the string "?OTR Error: ". Implementing this
operation is not required, but depending on your application it may be a
good idea.


/* Return a string that will be prefixed to any resent message. If this
 * function is not provided by the application then the default prefix,
 * "[resent]", will be used.
 * */
const char *(*resent_msg_prefix)(void *opdata, ConnContext *context);

/* Deallocate a string returned by resent_msg_prefix */
void (*resent_msg_prefix_free)(void *opdata, const char *prefix);

These operations give the option of chosing an alternative to the
English string "[resent]", when a message is resent.


/* Update the authentication UI with respect to SMP events
 * These are the possible events:
 * - OTRL_SMPEVENT_ASK_FOR_SECRET
 *      prompt the user to enter a shared secret. The sender application
 *      should call otrl_message_initiate_smp, passing NULL as the question.
 *      When the receiver application resumes the SM protocol by calling
 *      otrl_message_respond_smp with the secret answer.
 * - OTRL_SMPEVENT_ASK_FOR_ANSWER
 *      (same as OTRL_SMPEVENT_ASK_FOR_SECRET but sender calls
 *      otrl_message_initiate_smp_q instead)
 * - OTRL_SMPEVENT_CHEATED
 *      abort the current auth and update the auth progress dialog
 *      with progress_percent. otrl_message_abort_smp should be called to
 *      stop the SM protocol.
 * - OTRL_SMPEVENT_INPROGRESS 	and
 *   OTRL_SMPEVENT_SUCCESS 	and
 *   OTRL_SMPEVENT_FAILURE    	and
 *   OTRL_SMPEVENT_ABORT
 *      update the auth progress dialog with progress_percent
 * - OTRL_SMPEVENT_ERROR
 *      (same as OTRL_SMPEVENT_CHEATED)
 * */
void (*handle_smp_event)(void *opdata, OtrlSMPEvent smp_event,
	ConnContext *context, unsigned short progress_percent,
	char *question);

These SMP events are initiated by otrl_message_receiving() when it has
received an SMP TLV from a remote buddy. If you application is
implementing support for SMP authentication it should handle these
events appropriately.

Previously applications had to manually check, upon receiving messages,
whether the message contained any SMP TLVs that are relevant to the
current SMP state.

/* Handle and send the appropriate message(s) to the sender/recipient
 * depending on the message events. All the events only require an opdata,
 * the event, and the context. The message and err will be NULL except for
 * some events (see below). The possible events are:
 * - OTRL_MSGEVENT_ENCRYPTION_REQUIRED
 *      Our policy requires encryption but we are trying to send
 *      an unencrypted message out.
 * - OTRL_MSGEVENT_ENCRYPTION_ERROR
 *      An error occured while encrypting a message and the message
 *      was not sent.
 * - OTRL_MSGEVENT_CONNECTION_ENDED
 *      Message has not been sent because our buddy has ended the
 *      private conversation. We should either close the connection,
 *      or refresh it.
 * - OTRL_MSGEVENT_SETUP_ERROR
 *      A private conversation could not be set up. A gcry_error_t
 *      will be passed.
 * - OTRL_MSGEVENT_MSG_REFLECTED
 *      Received our own OTR messages.
 * - OTRL_MSGEVENT_MSG_RESENT
 *      The previous message was resent.
 * - OTRL_MSGEVENT_RCVDMSG_NOT_IN_PRIVATE
 *      Received an encrypted message but cannot read
 *      it because no private connection is established yet.
 * - OTRL_MSGEVENT_RCVDMSG_UNREADABLE
 *      Cannot read the received message.
 * - OTRL_MSGEVENT_RCVDMSG_MALFORMED
 *      The message received contains malformed data.
 * - OTRL_MSGEVENT_LOG_HEARTBEAT_RCVD
 *      Received a heartbeat.
 * - OTRL_MSGEVENT_LOG_HEARTBEAT_SENT
 *      Sent a heartbeat.
 * - OTRL_MSGEVENT_RCVDMSG_GENERAL_ERR
 *      Received a general OTR error. The argument 'message' will
 *      also be passed and it will contain the OTR error message.
 * - OTRL_MSGEVENT_RCVDMSG_UNENCRYPTED
 *      Received an unencrypted message. The argument 'smessage' will
 *      also be passed and it will contain the plaintext message.
 * - OTRL_MSGEVENT_RCVDMSG_UNRECOGNIZED
 *      Cannot recognize the type of OTR message received.
 * - OTRL_MSGEVENT_RCVDMSG_FOR_OTHER_INSTANCE
 *      Received and discarded a message intended for another instance. */
void (*handle_msg_event)(void *opdata, OtrlMessageEvent msg_event,
	ConnContext *context, const char *message,
	gcry_error_t err);

This operation is called when some type of exceptional event has occured
that your application may want to be aware of. Your application may want
to write an event to a log file, display a message to the user, or
ignore the event.  While it is not required to implement this operation,
it is probably a good idea.


/* Create a instance tag for the given accountname/protocol if
 * desired. */
void (*create_instag)(void *opdata, const char *accountname,
	const char *protocol);

This is called when the library notices this account name and protocol
pair does not have an instance tag. Similar to create_privkey(), your
application may simply open a file handle and call:
gcry_error_t otrl_instag_generate_FILEp(OtrlUserState us, FILE *instf,
        const char *accountname, const char *protocol)

If you don't provide an implementation for this operation, a new
non-persistent instance tag will be randomly generated. One benefit to
having a persisted instance tag is that if your application closes and
re-opens during a private conversation, further messages you receive
from this buddy will correctly raise the
OTRL_MSGEVENT_RCVDMSG_UNREADABLE event instead of raising
OTRL_MSGEVENT_RCVDMSG_FOR_OTHER_INSTANCE because destination instance
tag is now different from your own.


/* Called immediately before a data message is encrypted, and after a data
 * message is decrypted. The OtrlConvertType parameter has the value
 * OTRL_CONVERT_SENDING or OTRL_CONVERT_RECEIVING to differentiate these
 * cases. */
void (*convert_msg)(void *opdata, ConnContext *context,
	OtrlConvertType convert_type, char ** dest, const char *src);

/* Deallocate a string returned by convert_msg. */
void (*convert_free)(void *opdata, ConnContext *context, char *dest);

The convert_msg() operation is called immediately before a message is
encrypted and immediately after a message is decrypted. This callback
can tweak the plaintext message as needed. One use case would be for an
application to tweak formatting on the plaintext if, for example, this
is something that would normally be done on the plaintext by some other
entity while the message is in transit.

/* When timer_control is called, turn off any existing periodic
 * timer.
 * 
 * Additionally, if interval > 0, set a new periodic timer
 * to go off every interval seconds.  When that timer fires, you
 * must call otrl_message_poll(userstate, uiops, uiopdata); from the
 * main libotr thread.
 * 
 * The timing does not have to be exact; this timer is used to
 * provide forward secrecy by cleaning up stale private state that
 * may otherwise stick around in memory.  Note that the
 * timer_control callback may be invoked from otrl_message_poll
 * itself, possibly to indicate that interval == 0 (that is, that
 * there's no more periodic work to be done at this time).
 * 
 * If you set this callback to NULL, then you must ensure that your
 * application calls otrl_message_poll(userstate, uiops, uiopdata);
 * from the main libotr thread every definterval seconds (where
 * definterval can be obtained by calling
 * definterval = otrl_message_poll_get_default_interval(userstate);
 * right after creating the userstate).  The advantage of
 * implementing the timer_control callback is that the timer can be
 * turned on by libotr only when it's needed.
 * 
 * It is not a problem (except for a minor performance hit) to call
 * otrl_message_poll more often than requested, whether
 * timer_control is implemented or not.
 * 
 * If you fail to implement the timer_control callback, and also
 * fail to periodically call otrl_message_poll, then you open your
 * users to a possible forward secrecy violation: an attacker that
 * compromises the user's computer may be able to decrypt a handful
 * of long-past messages (the first messages of an OTR
 * conversation).
 */
void (*timer_control)(void *opdata, unsigned int interval);

In order to prevent a forward secrecy violation, applications using
libotr now need to be able to call otrl_message_poll on occasion.  The
simplest thing to do is just to set up a local timer that calls that
function every definterval =
otrl_message_poll_get_default_interval(userstate) seconds.  To avoid
unnecessary overhead, however, the timer_control callback is available.
If you set timer_control to non-NULL, it will be called with
instructions to turn on or off the periodic timer, and to what interval.

You must also be sure to turn off the timer before freeing your
userstate with otrl_userstate_free.

3.2. Instance Tags

If your application allows the same user to be logged in multiple times
from different locations, it should probably be aware of instance tags.
A user can maintain multiple concurrent OTR conversations with a buddy
who is logged in multiple times. Only one of the buddy's sessions can be
a client who is running OTR protocol version 2. When a user has a
conversation with a buddy who is running OTR protocol version 2, the
conversation is associated with a ConnContext that lists
"their_instance" as OTRL_INSTAG_MASTER (which has a value of 0). Each
version 3 conversation with the same buddy will have its own
ConnContext, which you can differentiate by the value in the
"their_instance" field.

In the linked list of ConnContexts, the master context for a buddy is
always listed immediately before its children. Fingerprints are only
stored with the master context. Given a ConnContext associated to a
conversation with a buddy, you can easily iterate over all the contexts
for that buddy by doing the following:

void example_something_happened(ConnContext * context) {
    ConnContext * context_iter = context->m_context;
    
    while (context_iter && context_iter->m_context == context->m_context) {
	/* Something you wish to affect all contexts of a particular buddy */
	context_iter = context_iter->next;
    }

If a user has multiple OTR sessions with the same buddy, your
application will likely want to provide some way for the user to select
which instance to send outgoing messages to. You can detect when a user
has multiple OTR sessions with the same buddy by iterating over the
ConnContexts of a buddy when a conversation has gone secure and checking
whether more than one is not in plaintext state. You specify which
instance outgoing messages are directed to in otrl_message_sending:

gcry_error_t otrl_message_sending(OtrlUserState us,
	const OtrlMessageAppOps *ops,
	void *opdata, const char *accountname, const char *protocol,
	const char *recipient, otrl_instag_t instag, const char *original_msg,
	OtrlTLV *tlvs, char **messagep, OtrlFragmentPolicy fragPolicy,
	ConnContext **contextp,
	void (*add_appdata)(void *data, ConnContext *context),
	void *data);

Instead of an actual instance tag, you can specify a meta instance tag
(e.g., if the user has not made an explicit selection). Here are the
list of meta instance tags, as defined in instag.h:

#define OTRL_INSTAG_BEST 1 /* Most secure, based on: conv status,
                            * then fingerprint status, then most recent. */
#define OTRL_INSTAG_RECENT 2 /* Most recent of the two meta instances below */
#define OTRL_INSTAG_RECENT_RECEIVED 3
#define OTRL_INSTAG_RECENT_SENT 4

OTRL_INSTAG_BEST choses the instance that has the best conv status, then
fingerprint status (in the event of a tie), then most recent (similarly
in the event of a tie). When calculating how recent an instance has been
active, OTRL_INSTAG_BEST is limited by a one second resolution.
OTRL_INSTAG_RECENT* does not have this limitation, but due to inherent
uncertainty in some networks, libotr's notion of the most recent may not
always agree with the remote network.  It is important to understand
this limitation due to the issue noted in the next paragraph.

Note that instances do add some uncertainty when dealing with networks
that only deliver messages to the most recently active session for a
buddy who is logged in multiple times. If you have a particular instance
selected, and the IM network is simply not going to deliver to that
particular instance, there isn't too much libotr can do. In this case,
you may want your application to warn when a user has selected an
instance that is not the most recent.

To explicitly specify the destination instance of a protocol version 2
conversation with a particular buddy, the instag value is
OTRL_INSTAG_MASTER.

To look up a ConnContext associated with a particular instance (or meta-
instance), specify the instance in otrl_context_find():

ConnContext * otrl_context_find(OtrlUserState us, const char *user,
	const char *accountname, const char *protocol,
	otrl_instag_t their_instance, int add_if_missing, int *addedp,
	void (*add_app_data)(void *data, ConnContext *context), void *data)

If your application persists instance tags, when it starts up, it should
call one the following functions to read the persisted instance tags:

gcry_error_t otrl_instag_read(OtrlUserState us, const char *filename);
gcry_error_t otrl_instag_read_FILEp(OtrlUserState us, FILE *instf);

It would make sense to do this immediately after your application has
read stored privkeys and fingerprints.

3.3. Fragmentation Changes

In libotr version 3.2.0, you would need to call otrl_message_sending()
to create an encrypted message, and then call fragment_and_send() to
get libotr to fragment and inject that message. In libotr 4.0.0, the
functionality of fragment_and_send() has been integrated into
otrl_message_sending(). Simply specify an OtrlFragmentPolicy to
otrl_message_sending(). The fragmentation policies are the same as
before, and an addition policy "OTRL_FRAGMENT_SEND_SKIP" has been added
for cases when fragmentation is not desired.

3.4. Asynchronous Private Key Generation

An application that wants to begin asynchronous key generation calls the
following method:

/* Begin a private key generation that will potentially take place in
 * a background thread.  This routine must be called from the main
 * thread.  It will set *newkeyp, which you can pass to
 * otrl_privkey_generate_calculate in a background thread.  If it
 * returns gcry_error(GPG_ERR_EEXIST), then a privkey creation for
 * this accountname/protocol is already in progress, and *newkeyp will
 * be set to NULL. */
gcry_error_t otrl_privkey_generate_start(OtrlUserState us,
        const char *accountname, const char *protocol, void **newkeyp)

A background thread can call the following method with the structure
that was passed into "newkeyp" above:

/* Do the private key generation calculation.  You may call this from a
 * background thread.  When it completes, call
 * otrl_privkey_generate_finish from the _main_ thread. */
gcry_error_t otrl_privkey_generate_calculate(void *newkey)


Upon completion the application would call:

/* Call this from the main thread only.  It will write the newly created
 * private key into the given file and store it in the OtrlUserState. */
gcry_error_t otrl_privkey_generate_finish(OtrlUserState us,
        void *newkey, const char *filename)

If the privkey generation was cancelled, the application should call:

/* Call this from the main thread only, in the event that the background
 * thread generating the key is cancelled.  The newkey is deallocated,
 * and must not be used further. */
void otrl_privkey_generate_cancelled(OtrlUserState us, void *newkey)


3.5. Library Initialization

If you currently initialize libotr with the recommended OTRL_INIT;
macro, you do not need to change anything.

If you call otrl_init(ver_major, ver_minor, ver_sub) directly, then know
that this function no longer returns void.  Previously, if the
application requested version numbers incompatible with those of the
library, the library would exit(1).  Now, the otrl_init call will return
a non-zero error code.  You must check the return value of otrl_init (a
gcry_error_t), and if it is non-zero, your application's expected
API/ABI does not match the installed libotr, and libotr cannot be used.