发明名称 State capture after execution in dependent sequences
摘要 The execution of a dynamic code segment sequence that includes at least two code segments in sequence. The first code segment is first executed and the corresponding state of the environment is captured. The second code segment is then executed. When the second code segment is later re-executed, the first code segment is not re-executed again. Rather, the environmental state is set to be the captured state that existed when the first code segment originally executed. Then, the second code segment may be executed without spending the resources required to re-run the first code segment. This may be employed at authoring time, or after deployment time.
申请公布号 US9047407(B2) 申请公布日期 2015.06.02
申请号 US201113329130 申请日期 2011.12.16
申请人 Microsoft Technology Licensing, LLC 发明人 Jazdzewski Charles P.;Fanning Michael C.
分类号 G06F9/44;G06F11/30;G06F11/36 主分类号 G06F9/44
代理机构 代理人 Sullivan Kevin;Hoff Aaron;Minhas Micky
主权项 1. A method by which a computing system executes a dynamic code segment sequence in an environment, the code segment sequence comprising a plurality of sequential code segments and the method comprising the following acts: executing a first code segment; capturing a first environmental state of a first instance of an execution context, the first environmental state representing a state of at least a portion of the first instance of the execution context after the execution of the first code segment is stopped; executing a second code segment; writing a second environmental state of a second instance of the execution context, wherein writing the second environmental state of the second instance of the execution context comprises performing the following for at least one context component in the first instance of the execution context: detecting that a read is to be performed on the at least one context component, and in response creating a proxy context component for the second environmental state of the second instance of the execution context, the proxy context component referencing the at least one context component as captured in the first environmental state of the first instance of the execution context;detecting that a write is to be performed on the at least one context component, and in response: copying values from the at least one context component as captured in the first environmental state of the first instance of the execution context to the proxy context component in the second environmental state of the second instance of the execution context;declassifying the proxy context component in the second environmental state of the second instance of the execution context as a proxy for said at least one context component; and performing the write on the copied values copied to the declassified proxy context component.
地址 Redmond WA US