Microsoft Technology Licensing, LLC

Thierry Miceli

Mark Lee Hall

Richard Thomas Russo

Jason J Shirk

Stephan Thomas Lavavej

Tanveer Ahmed Gani

Mitchell Warren Slep

Declarations from an input source code are serialized into a stream of tokens produced by following each branch of a preprocessor conditional directive statement that interrupts a declaration. Tokens are labeled with a parsing path indicator corresponding to a parsing path induced by branches of a preprocessor conditional directive. The declarations that are formed along the different parsing paths are serialized by fetching the tokens that belong to the first parsing path in a first pass, and passing the tokens on to a next phase of a compiler. The pointer that marks the next token is repositioned to return to the start of the declaration. The declaration may be serialized again through the second parsing path in a second pass. The operation may be repeated until each of the parsing paths induced by the presence of branches of the preprocessor conditional directives in the source code is exhausted.

Parsing of declarations in all branches of preprocessor conditionals

Many integrated development environments provide 'intelligent feedback' to developers. Current 'intelligent feedback' techniques may not work well with source code files comprising context sensitive programming languages because edits to one place in the code may change the meaning of arbitrary other places in the code, thus requiring additional full semantic parses to maintain correctness of the system. As provided herein, context-free &#x201c;name&#x201d; parsing may be performed to support entity candidate recursive semantic parsing in determining the meaning of the entity. A name parse may be performed to build and/or update a name table and a range table comprising names of entities and their positions within a source code file. Upon receiving a request to determine the meaning of a selected entity within the source code file, the tables may be queried to determine entity names and locations that are to be semantically parsed to determine a selected entity meaning.

Efficient incremental parsing of context sensitive programming languages

Declarations from an input source code or tokenized source code are serialized into a stream of tokens produced by following each branch of a preprocessor conditional directive statement that interrupts a declaration. Tokens are labeled with a parsing path indicator corresponding to a parsing path induced by branches of a preprocessor conditional directive. The declarations that are formed along the different parsing paths are serialized by fetching the tokens that belong to the first parsing path in a first pass, and passing the tokens on to a next phase of a compiler. The pointer that marks the next token is repositioned to return to the start of the declaration. The declaration may be serialized again through the second parsing path in a second pass. The operation may be repeated until each of the parsing paths induced by the presence of branches of the preprocessor conditional directives in the source code is exhausted.

Growing community of inventors

Seattle, WA, United States of America

Thierry Miceli