W3MFC v1.86 / CThreadPoolServer v1.30 / CWSocket v1.25

Just to let everyone know that v1.87 of W3MFC, v1.30 of CThreadPoolServer & v1.25 of CWSocket has been released. Updates for this version include:

• Removed all legacy VC 6 code from classes. All the modules are now for VC 2005 or later only.
• Fixed a memory leak in CW3MFCClient::_TransmitFile
• All methods of CBase64 class have been made static.
• Fixed a 64 bit bug in CW3MFCDirectory::TransmitDirectory related to the 64 bit size returned from CFileFind::GetLength.

Advertisement

StartX v1.05

Just to let everyone know that v1.05 of StartX has been released. Updates for this version include:

• Updated copyright details.
• Removed VC 6 style AppWizard comments from the code
• Updated the code to clean compile on VC 2008
• The code and shipped exe has been updated to take advantage of VC 2005 specific coding techniques.
• In fact the shipped exe in the zip was compiled with VC 2005 SP1.
• CStartXApp::GetErrorMessage now uses the FORMAT_MESSAGE_IGNORE_INSERTS flag. For more information please see Raymond Chen’s blog at http://blogs.msdn.com/oldnewthing/archive/2007/11/28/6564257.aspx. Thanks to Alexey Kuznetsov for reporting this issue.

GPSCom v1.05 / IConnectionPointImplMT v1.01

Just to let everyone know that v1.05 of GPSCom and v1.01 of IConnectionPointImplMT has been released. Updates for this version include:

• Updated copyright details.
• GPSCom now internally uses v1.24 of CSerialPort.
• Fixed a bug in CGPSConnection::MonitorThread in the correct use of CT2A.
• How you implement the Fire_ methods as documented for IConnectionPointImplMP has been updated to refer to the need to serialize access to the "m_Clients" member variable which is accessed via the GetInterfaceAt method. The CProxy_IGPSConnectionEvents class in the GPSCom project and the IConnectionPointImplMT::GetInterfaceAt method have been updated to address this issue. Thanks to Michael Ford for reporting this bug.
• The actual pointers which are stored on the internal IConnectionPointImplMP::m_Clients array are now "CComGITPtr<TSinkInterface>" heap pointers rather than actual stack instances. By doing this, the code now has explicit control over when and how copy constructors and operator= methods are being called as we add a element to the array. This avoids extra overhead as the interface pointer is being added to the GIT. It also addresses an issue where the cookie value which was being returned by Advise was actually incorrect due to the fact that a copy of the CComGITPtr was being added to the array. Thanks to Michael Ford for reporting this issue.

W3MFC v1.86 / CWSocket v1.24

Just to let everyone know that v1.86 of W3MFC and v1.24 of CWSocket has been released. Updates for this version include:

• Updated copyright details.
• Fixed potential heap memory leaks in CWSocket::ReadHTTPProxyResponse.Thanks to Michal Urbanczyk for reporting this bug.
• Fixed a memory leak in CWSocket::ConnectViaSocks5
• Restructured CWSocket::ReadSocks5ConnectReply to avoid the need to allocate heap memory

CPJNSMTPConnection v2.75

Just to let everyone know that v2.75 of CPJNSMTPConnection has been released. Updates for this version include:

• Updated copyright details.
• Fixed a bug in CPJNSMTPMessage::FormDateHeader where sometimes an invalid "Date:" header is created. The bug will manifest itself for anywhere that is ahead of GMT, where the difference isn’t a multiple of one hour. According to the Windows Time Zone applet, this bug would have occured in: Tehran, Kabul, Chennai, Kolkata, Mumbai, New Delhi, Sri Jayawardenepura, Kathmandu, Yangon, Adelaide and Darwin. Thanks to Selwyn Stevens for reporting this bug and providing the fix.

CSingleInstance v1.17

Just to let everyone know that v1.17 of CSingleInstance has been released. Updates for this version include:

• Updated copyright details
• Removed VC 6 style classwizard comments from the sample apps code
• Updated ActivatePreviousInstance method to support Win64 compliant data
• ActivatePreviousInstance now takes a "dwTimeout" parameter which it now uses internally as the timeout when calling SendMessageTimeout instead of SendMessage. The code now uses SendMessageTimeout instead of SendMessage to ensure we do not hang if the previous instance itself is hung. Thanks to Paul Shore for suggesting this update.
• Updated the sample apps to clean compile on VC 2008