Feynman learning method (symbol table)

Symbol table is a data structure that collects identifier information during compilation , Generally, there are arrays 、 Linked list 、 Hashtable 、 Binary search tree, etc . The same identifiers will be organized together , Form constant table 、 Variable name table 、 Function name table and so on . The compilation phase will continue to generate symbol tables , Constantly use the symbol table to parse , This is also the main overhead of the compilation process .
There are three main functions of the symbol table ：
① Lexical analysis stage ： The compiler will recognize the identifiers one by one according to the spaces , Then in the same scope , Classify according to symbol attributes , Record it in the symbol table , Each identifier is associated with its data type 、 Scope 、 name 、 The address and other information in memory are bound together . It mainly collects the attributes of symbols .
② Syntax parsing stage ： Through the collected symbol table , Build the syntax tree , Mainly to check the legitimacy of the context . For example, in the same scope , When variables with the same name are collected , Will report a mistake , The cause of the error is the repeated definition of variables . Again for instance , Within a new scope , A new symbol table will be generated to collect information , Then suppose there is another internal scope , A new symbol table will be generated for collection , And the symbol table corresponding to this internal scope will have a pointer , Point to the symbol table corresponding to the external scope of the previous layer , When you need to resolve this identifier , If it cannot be found in the symbol table of the current scope , You need to follow the pointer to the symbol table corresponding to the scope of the previous layer , If you can't find it all the way to the symbol table of the outermost scope , You're going to report a mistake , The error message is to use undeclared variables . The main function is to test the legitimacy of the context .
③ The code is preprocessed 、 compile 、 After the compilation , Generated is the target file , stay linux Next is .O file , This file carries many symbol tables , Then you need to go through the linker , Put these .O collect files , link , Generate the final executable binary . Then this process requires a symbol table , Determine the relative address of the identifier in the final binary file and the memory layout area .