This file is empty

Member resip::BranchParameter::setSigcompCompartment (const Data &)

The encoding here could be more efficient.

Member resip::ConnectionBase::transport () const

should be reference

It might be worthwhile examining the heap usage of this class in the context of using realloc everywhere appropriate. (realloc is defined in ANSI C, SVID, and the OpenGroup "Single Unix Specification").

Member resip::Data::charEncoded () const

This method should be removed

Member resip::Data::charUnencoded () const

This method should be removed

Member resip::Data::Data (ShareEnum, const Data &staticData)

This implementation has some confusing and conflicting comments. (e.g. is Borrow actually okay? Can there be some way to use it with Take as long as you play with mShareEnum correctly?)

Member resip::Data::truncate (size_type len)

Remove this at some point.

Member resip::DomainMatcher::addDomain (const Data &domain)=0

Make this case-insensitive.

Class resip::ExpiresCategory

Seems like this and UInt32Category (numeric with a comment) could be renamed to something a little more self-explanatory.

Class resip::FiniteFifo< Msg >

remove class

Class resip::MessageWaitingContents::AccountHeader

This is an awkward attempt to make accessing the MessageAccount Uri operate like accessing the other MessageWaiting headers even though the user of these calls has to treat them differently to being with since the two header calls return different classes. ugh. .mjf.

Class resip::ParserCategory

Maybe a better name? IHaveParams? ElemWithParams?

Member resip::ParserContainer< T >::reverse () const

!bwc! optimize this (we are copying each ParserContainer twice)

Class resip::ProducerFifoBuffer< T >

Refactor so we can use this with AbstractFifo<T>? This is presently not the case because AbstractFifo does not expose its API publicly.

Member resip::Random::getVersion4UuidUrn ()

This is something of a suboptimal hack. Ideally, we would encapsulate UUID as its own class, with options to create any of versions 1 through 5 (and Nil UUIDs). This class would have various access methods to pull the result out as a bitstring, as raw data, as a URN, etc. For what I need to do right now, however, version 4 UUIDs get me where I need to go. <abr>

Member resip::SipMessage::clone () const

.dlb. sure would be nice to have overloaded return value here..

Member resip::SipMessage::defineHeader (RetryAfter, "Retry-After", UInt32Category, "RFC 3261")

!dlb! this one is not quite right – can have (comment) after field value

Member resip::SipMessage::SipMessage (const SipMessage &message)

.dlb. public, allows pass by value to compile.

Member resip::SipStack::setContentLengthChecking (bool check)

should this be fixed to work with other applicable transports? []

Member resip::SipStack::statisticsManagerEnabled ()

is this documented correctly? [!]

Member resip::TimeLimitFifo< Msg >::setCountDepthTolerance (unsigned int max)

getCountDepthTolerance returns a size_t and this takes an unsigned int, f.setCountDepthTolerance(f.getCountDepthTolerance()) will fail to compile on many systems [!]

Member resip::TimeLimitFifo< Msg >::setTimeDepthTolerance (unsigned int maxSecs)

examine parallelism in units [!]

Member resip::TransactionUser::addDomain (const Data &domain)

Make this case-insensitive.

Member resip::Transport::send (std::unique_ptr< SendData > data)=0

!bwc! What we do with a SendData is flexible. It might make a copy, or send synchronously, or convert to another type, etc.

!bch! Should this be protected and not public? !bwc! TransportSelector uses this directly for retransmissions.

Member resip::Transport::shutdown ()

?dcm? pure virtual protected method to enforce this?

Member resip::Transport::startOwnProcessing ()=0

?dcm? avoid the received a message but haven't added a transport to the TransportSelector race, but this might not be necessary.

Member resip::Transport::Transport (Fifo< TransactionMessage > &rxFifo, int portNum, IpVersion version, const Data &interfaceObj, const Data &tlsDomain=Data::Empty, AfterSocketCreationFuncPtr socketFunc=nullptr, Compression &compression=Compression::Disabled, unsigned transportFlags=0, const Data &netNs=Data::Empty)

Note that because of InternalTransport's constructor, tlsDomain is always set to Data::Empty at construction time, in practice.

Member resip::Tuple::isV4 () const

!dcm! – should deprecate asap