Microsoft Technology Licensing, LLC

Deborah Chen

Jordi Mola

Jackson Michael Davis

Patrick Nelson

Thomas Lai

Del Myers

Juan Carlos Arevalo Baeza

Andrew R Sterland

Mark Alistair Wilson-Thomas

Simon Calvert

Kenneth W Sykes

Kesavan Shanmugam

James M Pinkerton

John Sells Tilford

Leslie Yvette Richardson

Marc Manuel Johannes Brockschmidt

Miltiadis Allamanis

Srivatsn Narayanan

Guillermo Serrato Castilla

Noah Falk

Joshua Bates Stevens

Based on replay of a thread, one implementation observes an influx of a value of a memory cell comprising an interaction between the thread and the value of the memory cell at an execution time point in the replaying, and determines whether the value of the memory cell observed from the influx is inconsistent with a prior value of the memory cell as known by the thread at the execution time point. If so, this implementation initiates an indication of a data inconsistency. Based on replay of a plurality of threads, another implementation identifies a memory cell that was accessed by a first thread while a thread synchronization mechanism was active on the first thread. Then, if there was another access to the memory cell by a second thread without use of the thread synchronization mechanism, this implementation initiates an indication of a potential data contention.

Identifying data inconsistencies and data contention based on historic debugging traces

Identifying and reporting potential data inconsistencies and/or potential data contentions based on historic debugging traces. Based on replay of a thread, some implementations observe an influx of a value to a memory cell, and determine whether the value of the memory cell observed from the influx is inconsistent with a prior value of the memory cell as known by the thread. If so, these implementations can initiate an indication of a data inconsistency. Based on replay of a plurality of threads, other implementations identify a memory cell that was accessed by a first thread while a thread synchronization mechanism was active on the first thread. Then, if there was another access to the memory cell by a second thread without use of the thread synchronization mechanism, these implementations might initiate an indication of a potential data contention.

Improved techniques for asynchronously displaying the results of a codebase analysis service are provided herein. Initially, machine learning is applied to a corpus of model code. In doing so, a machine learning model is generated, where the model identifies coding practices that are included in the corpus of model code. After this model is generated, then the model is applied to a current codebase by comparing coding practices of the current codebase to the identified coding practices that were extracted, or rather identified, from the corpus of model code. Then, in response to detecting one or more differences between the current codebase's coding practices and the identified coding practices, where the differences satisfy a pre-determined difference threshold, a user interface is caused to display one or more insights. These insights beneficially provide additional detailed information describing the differences.

Displaying results of a codebase analysis service

Described technologies aid execution control during replays of traced program behavior. Cross-level mapping correlates source code, an intermediate representation, and native instructions in a trace. The trace includes a record of native code instructions which were executed by a runtime-managed program. The trace does not include any executing instance of the runtime. Breakpoints are set to align trace locations with source code expressions or statements, and to skip over garbage collection and other code unlikely to interest a developer. A live debugging environment is adapted to support trace-based reverse execution. An execution controller in a debugger or other tool may utilize breakpoint ranges, cross-level mappings, backward step-out support, and other items to control a replay execution of the trace. Aspects of familiar compilers or familiar runtimes may be re-purposed for innovative execution control which replays previously generated native code, as opposed to their established purpose of generating native code.

Execution control with cross-level trace mapping

Described technologies extend the information available from an execution trace of a program by providing heuristically-derived values for memory contents when the trace does not include data expressly showing the value of a memory cell at a particular execution time. Various heuristics are described. The heuristics may use information about the memory cell at other times to produce the derived value. Some heuristics use other trace data, such as whether the memory cell is in a stack, whether there are gaps in the trace, or whether garbage collection or compilation occurred near the time in question. Grounds for the derived value are reported along with the derived value. A time-travel debugger or other program analysis tool can then present the derived values to users, or make other use of the derived values and grounds to assist debugging and other efforts to improve the functioning of a computing system.

Level-crossing memory trace inspection queries

Improving how a codebase, which may include source code, related databases, test files, code history, and/or changes, is drafted, edited, debugged, or otherwise developed. Machine learning is performed on a model codebase to establish a machine learning model. When a change to a codebase occurs, the machine learning model may be applied to evaluate that change. A change context providing context for this change is accessed. An analyzer then analyzes the change using the machine learning model and at least a part of the change context to generate an analysis result. Some information about the result is rendered. After rendering that information, a determination regarding how a user responded to the information is performed, and a subsequent analysis is then modified based on the user's response.

Code development using continued machine learnings

Growing community of inventors

Seattle, WA, United States of America

Deborah Chen