This is a simple C++ runner for doing Advent-of-Code challenges. It's designed to have one executable which is given parameters for the day and part, and then it will run that particular solution.
It's designed for C++ under Linux. I've tested it with CentOS 8 and GCC version 8.3.1.
After checking out the project, it's just 2 commands to build
./mkdirs.sh
make
mkdirs.sh will create the bin and build directories and their subdirectories. It only needs to be run the first time.
The starter comes with a Day 0 implementation, which is based on Advent of Code 2018, day 1, part 1. The part 1 solution matches the problem, and the part 2 solution the same as part 1, except it gives the negation of the part 1 result for a solution. Looking at aoc_day_0.h, /aoc_day_0.cpp, and aoc_days.cpp will show all the coding needed to hook in a new day.
Each day's solution will inherit from the AocDay class (include/solutions/aoc_day.h) as is done in include/solutions/aoc_day_0.h .
There are two functions to override in the child class:
virtual string part1(string filename, vector<string> extra_args);
virtual string part2(string filename, vector<string> extra_args);
The base AocDay class has default implementations of these functions, so you don't even have to define a part2 function until after part1 is done, if you'd prefer.
Each function takes two parameters - the filename for the input file and a vector of extra arguments, which I'll describe below.
The return value is a string for the solution. "Why a string and not an int/long?" you might ask. Although Advent of code 2019 had only numeric solutions, I found that in 2018 there were times when the solution was non-numeric (day 7 part 1, day 13 parts 1 and 2). So, this function returns a string back to the driver program.
The parameter of extra_args is useful to prevent code changes for hard-coding limits or other constants. For example, let's say the test cases presented in the problem description show a result after 10 iterations, but then the actual answer is supposed to run 1000000 times. You can pass the 10 or 1000000 value as an extra arguent instead of having to change a constant and recompile. These are also passed in as strings, which can then be converted to ints/longs/whatever as needed.
I'll preface this by saying that I don't like dealing with Makefiles. I'm sure I could make some of this prettier, but it is what it is.
The Makefile for this compiles a couple of libraries - librunner.a for some of the basic test runner functionality, and libsolutions.a for the daily solutions. As things get more complex, I'm sure more libraries will get added. If you want to see what I mean, there's a more complex example in my intCode repo (https://github.com/bcooperstl/intCode).
For each day's solution, the Makefile file will need to be modified to add in a .o file target that corresponds to the .cpp file for that day. That .o file must also be added to libsolutions.a .
The final executable is built as bin/aoc
There are some helper functions in the FileUtils class to read in and parse an input file.
bool read_as_list_of_strings(string filename, vector<string> & lines) - Read the file given by filename and return the lines of the file in the lines vector.
bool read_as_list_of_split_strings(string filename, vector<vector<string>> & split_strings, char delimiter, char quote_char, char comment_char) - Read the file given by filename, return it as a list of list of strings in the split_strings vector. delimiter identfies a delimiter to identify how to split the strings. quote_char gives a way to allow the delimiter to appear in a string by quoting around it. comment_char allows for a line to be skipped from the output if it starts with this character. This probably sounds like overkill, but I use a bunch of this with my test file format. See testing.md for an example and more details.
bool read_as_list_of_split_longs(string filename, vector<vector<long>> & split_longs, char delimiter, char quote_char, char comment_char) - Same as read_as_list_of_split_strings but with longs. Think of it as an easy way to read in an intcode program from 2019, if you will.
There are three modes to run this program as shown in the usage. This section describes how to run one input file through. The testing.md file describes the two other modes.
The command line is:
bin/aoc -d day -p part -f filename [extra_args...]
Everything should be straight forward - give it the day, part (1 or 2), input file, and optionally any extra arguments. It'll spit out the result or tell you if there's an error.
For example:
[brian@dev1 advent-of-code-cpp-starter]$ bin/aoc -d 0 -p 1 -f data/sample/day0_input.txt
***Day 0 Part 1 for file data/sample/day0_input.txt has result 569