In LoadComplete 4.60, the following bugs reported by our customers have been fixed:
Fixed: Rendezvous points worked incorrectly in tests that simulated a great number of virtual users and in tests that simulated virtual users on several load stations.
Fixed: LoadComplete could not record any traffic if the “Bypass proxy server for local addresses” option was enabled in the system proxy settings.
Fixed: Load test reports could show the start time of virtual user simulation incorrectly.
Fixed: LoadComplete did not distribute stepwise load among virtual user groups.
Fixed: LoadComplete could not record traffic if the name of the tested web site matched the name of one of the servers specified in the list of exceptions in the Internet Explorer proxy settings.
Fixed: LoadComplete ignored the list of exceptions in the system proxy settings and always simulated traffic via a proxy server.
Fixed: During traffic recording, LoadComplete could freeze if the HTTP client was using HEAD requests. The client could freeze as well.
Fixed: When extracting data from Rich Internet Application responses, instead of working with the text version of the response contents, LoadComplete could try to parse the contents and failed if the data format was not supported.
Fixed: LoadComplete failed to stop traffic recording correctly if the Connection header in the responses from the tested web server had a wrong format.
Fixed: LoadComplete could throw the “ADOQuery is not a valid component name” exception when configuring database variables.
Fixed: The Canada (ca-central-1) and London (eu-west-2) Amazon EC2 regions were not supported.
Fixed: LoadComplete could launch an incorrect number of Amazon cloud computers when simulating traffic from the cloud.