Given how I like to configure Antlr, the build system of choice this time is Gradle since it allowed for easier to use custom main methods and a generated-src directory.
The syntax is pretty similar to what we went over in class. There are, however, a few very important distinctions. First, unary negation exists. Second, all values are interpreted as floating point numbers rather than integers. As such, values may be given in standard decimal floating point format or in scientific notation. See src/main/antlr/CommonLexerRules` for specifics. Note that files should have a trailing newline in order to be processed properly, though this isn't a concern for the generated formats.
Due to the presence of two independent main methods, there is no generic run task for the project. Rather, each part
has its own task that can be ran. part1 takes file(s) in infix notation, executes them, then converts them to infix
and postfix format. part2 is similar, but starts with programs in prefix notation.
From the command line, use
$ ./gradlew part1to run with default arguments. If you'd like to parse your own files rather than the included test file, call it with
$ ./gradlew part1 --args="..."The command line arguments for both parts are relatively straightforward, and follow the format:
part [flag | file]*
Available flags include
--save-output: save output of modified forms to disk. These are in the forminfixed/prefixed/postfixed_file, where file refers to the input file. They are produced in the directory where the program was ran from--use-stdin: read a program from stdin rather than from files. If this flag is set, any provided files are ignored.
Any other argument not prefixed with -- will be considered as a filename to be parsed. If any other arguments starting
with -- are provided, an exception will be thrown.
Included tests can be ran with
$ ./gradlew testIf desired, extra tests can be placed in the src/test/java directory.