Simply put your desired script in "src/script.txt" then compile the code g++ -o main src/main.cpp after that execute it with ./main run it on tinyLanguageParser folder and not inside src. you will get the results either in run prompt or log.txt if it is a valid you get the constructed AST in both JSON format and BMP format for image representation otherwise in case it is an invalid program then you get the error.
In case you want to run Tests you must first read how I'm handling tests, then just compile the code and run it. Basically the tests are methods that have try catch inside, and you must expect the resulting output ( string error ) to see if it passes or no.
I assumed that program D is not valid as we have a buffer overflow, we try to access the 10th element and assign it which does not exist as we start from 0, we already have return 0 and i<0, so we have to assume that 0 is present as digit, and we start from 0.
Main method does not have parameters as described it is function and has the fixed declaration “int main().”
I wrote a Lexer and Parser from scratch depending on the provided EBNF grammar for this Tiny Programming Language. The Lexer class first lex the given script in form of tokens, the whitespaces and comments are skipped then we parse those tokens one by one using Top Down Parsing technique to go from top to bottom while building the AST. For the AST construction I defined multiple Nodes which may be updated for more precise construction of AST, declaration of function, single, multiple variables, if, else, while, assignment, return, function call, and more.. which can be found in astNode class. JSON format convertor is based simply on those node and each node has its own toJSON function that does the work. From JSON to BMP image I have implemented a class imageConvertor that takes the JSON then converts it simply to BMP format which can make it very useful and informative BMP image still needs updates.
Float values works with . instead of , --> 10.2 is valid float while 10,2 is not a valid float Int values Char Boolean Arrays of any type
I assumed that if statement only works if else is present, the logic I built was that "if" have true and false branch, returning a null pointer for false branch is not a good idea in C++ especially when building the AST nodes.
I assumed that characters works with '' instead of this ‘’ , example : c = ‘a’; ---> Invalid c = 'a'; ---> Valid
I wanted to handle tests differently without including any library or framework for testing only for the sack of portability and ease of use I made 2 classes one for valid and the other for invalid syntax as described in project, each class contains multiple functions that have different scripts with one main function that call those tests, in a simple way I test with this script if the parseAST throws an error or no, if yes then print what it expected to do, example : if a give an invalid script in the function then when it catches then the test passed because it was expected to catch, I also print the exception.
A parser for a fictional language called Tiny that is similar to C Lexical Analysis: We must first tokenize the script, then we start parsing token by token, the parsing will fail in the following cases: -Declarations outside functions then the parsing fails as global variables are not allowed. -Assignment for undeclared variables, after we check them in the current scope of symbol table. -Another declaration after assignment, as the tiny has first declarations part then assignment part. -Syntax error, like finishing a statement with semicolon ; or not following the rules for function declaration, parenthesis and other tokens.
Syntax Analysis: We match each token with our grammar that describe Tiny Language, if a token doesn't match a specific rule then we output syntax error. there is a case where we match multiple rules because of the ambiguity of the grammar, we also have to choose the ideal rule.
Semantic Analysis: If the syntax analysis pass then we can start our semantic check, if for example we have "l"+1 then we output a semantic error
Overflow Analysis: If an operation either addition,subtraction, multiplication, division or modulo exceeds the max int or min int, also same for buffers then we output overflow error with the details of the line and variables. I assumed that we have only unsigned numbers, basically 16 bits is equal to a max_int of 65536. After getting the AST we try to go node by node and whenever we have an binary operation node then we check the right value and left value with the given operation to check if it overflows or no.
Output: If all analysis pass then we output valid program with the AST.
CMAKE is not well maintained and currently the project works only with normal g++ compile.
GIT has been used to control the versions and get the work-log, the repo is currently private and cannot be accessed only by me.
There is a Dockerfile that you can use to build the image otherwise docker pull dzimo16/tinylanguageparser:v1 you can pull this in your local
C++ 17 CMAKE 3.25 G++ 13.2
LAST VERSION