Review/serial and readme bunsarak

Makefile

@@ -0,0 +1,56 @@
+.POSIX:
+.SUFFIXES:
+
+# Use the system's default C compiler
+# was gnu89 reallyyyy necessary ....
+# surely gradescope machines have a compiler which can do post Y2K C ....
+
+CC		= cc
+CFLAGS	= -Wall -Wpointer-arith -Wstrict-prototypes -std=c11 -fPIC
+
+# Pathname of the pkg-config compatible utility
+# (not using this for this assignment at all)
+#PC		= pkg-config
+
+# Instructions to create an LSP db -- default is clangd with compile_flags.txt
+# e.g. To use bear you can set DB="bear -- make" which will create compile_commands.json
+DB		= echo $(LDFLAGS) $(CFLAGS) | tr ' ' '\n' > compile_flags.txt
+
+# Project Files
+BIN=test-mergesort
+OBJ=test-mergesort.o mergesort.o
+SRC=$(OBJ:%.o=%.c)
+DEP=$(OBJ:%.o=%.d)
+
+all: $(BIN)
+
+# Build binary from objects
+$(BIN): $(OBJ)
+	$(CC) $(CFLAGS) -o $@ $(OBJ) $(LDFLAGS)
+
+# Suffix rules to create .o and .d files from sources
+.SUFFIXES: .c .o
+.c.o:
+	$(CC) $(CFLAGS) $(INC) -c $<
+
+.SUFFIXES: .c .d
+.c.d:
+	$(CC) -MM $< -o $@
+
+# Unit tests
+test:
+	make -f merge_test.make
+
+# Generate LSP database on each clean
+db:
+	$(DB)
+
+clean: db
+	rm -fr $(BIN)
+	rm -fr $(OBJ)
+	rm -fr $(DEP)
+
+	make -f merge_test.make clean
+
+# For FreeBSD make use -include
+-include $(DEP)

README.md

@@ -0,0 +1,90 @@
+README.template
+
+## Project Number/Title 
+
+* Authors: Izak Baldacchino (a1830164), Bunsarak Ann (a1827385), Matthew Edmonds-Wilson(a1850372)
+* Group name: Assignment 3 Groups 190
+
+## Overview
+
+This program benchmarks the performance of the parallel merge sort algorithm.
+Given an array of random integers of size N, and a "level", it performs merge
+sort; each time the array is partitioned, a new thread is spawned to perform
+the recursive merge sort on the sub-array. This continues until the "level"
+is exceeded.
+
+## Manifest
+
+Makefile 		 -- the recipes which are used to compile the project.
+README.md 		 -- this file!
+mergesort.c 	 -- contains the source for the functions which need to be implemented.
+mergesort.h 	 -- contains the declarations for the functions and structs used.
+test-mergesort.c -- contains the source for the program which runs the algorithm.
+merge_test.make  -- tiny makefile to build the merge unit test
+
+## Building the project
+
+To build the project simply execute `make`. This assumes it is being built on a UNIX-like
+system which features some rendition of the `make` utility. In fact, any system which is
+POSIX-compliant shall be able to execute and run this makefile.
+
+From this point execute the built executable `./test-mergesort` and provide relevant arguments.
+
+macOS note: If you see a build error about <error.h>, we’ve already removed that include for macOS compatibility in this branch. If you pull a branch that still includes it, delete the line in test-mergesort.c.
+
+## Features and usage
+
+To run the program use `./test-mergesort N level seed` where `N` refers to the size of the generated
+array, `level` refers to the number of times the array is partitioned and sorted in parallel, and
+`seed` is used as the seed for generating the random array; same seed = same array.
+
+TODO: write more perhaps?
+
+## Testing
+Unit tests for merge()
+Balanced halves, unbalanced halves, left/right remainder cases, duplicates, already sorted, reverse-sorted, tiny segments (size 0/1).
+
+Integration check (temporary)
+While parallel_mergesort is incomplete, we verify my_mergesort correctness by temporarily routing parallel_mergesort to call my_mergesort (serial path) and checking the harness output. (In this branch, parallel_mergesort is intentionally a stub.)
+
+This section should detail how you tested your code. Simply stating "I ran
+it a few times and it seems to work" is not sufficient. Your testing needs
+to be detailed here.
+
+
+TODO: figure out testing scheme.
+
+## Known Bugs
+
+List known bugs that you weren't able to fix (or ran out of time to fix).
+
+TODO: haven't made any bugs yet ...
+
+## Reflection and Self Assessment
+Early failure we hit: Harness printed “sorting failed!!!!” because it always calls parallel_mergesort() and ours was a stub.
+What we learned: keep a unit test for merge() separate from threading so we can validate core correctness early.
+
+Discuss the issues you encountered during development and testing. What
+problems did you have? What did you have to research and learn on your own?
+What kinds of errors did you get? How did you fix them?
+
+What parts of the project did you find challenging? Is there anything that
+finally "clicked" for you in the process of working on this project? How well
+did the development and testing process go for you?
+
+
+TODO: haven't finished yet ...
+
+## Sources Used
+Operating Systems: Three Easy Pieces — chapters on concurrency for general pthread patterns.
+Assignment 3 specification and starter code (University of Adelaide, COMP SCI 3004–7064).
+Assistance with structuring tests and clarifying merge-edge cases provided by AI (ChatGPT).
+https://www.youtube.com/watch?v=b1n9T-Iy3SA 
+https://stackoverflow.com/questions/2550774/what-is-size-t-in-c
+
+If you used any sources outside of the textbook, you should list them here. 
+If you looked something up on stackoverflow.com or you use help from AI, and 
+fail to cite it in this section, it will be considered plagiarism and dealt 
+with accordingly. So be safe CITE!
+
+TODO: haven't used any sources yet ...

merge_test.c

@@ -0,0 +1,95 @@
+#include <assert.h>
+#include <stdio.h>
+#include <stdlib.h>
+
+#include "mergesort.h"
+
+// global state is safe and good and never goes wrong and doesn't make testing difficult!
+int* A;
+int* B;
+
+void printA(size_t start, size_t end) {
+	printf("\nA ");
+	for(size_t i = start; i < end+1; i++) {
+		printf("%d ", A[i]);
+	}
+}
+
+void printB(size_t start, size_t end) {
+	printf("\nB ");
+	for(size_t i = start; i < end+1; i++) {
+		printf("%d ", B[i]);
+	}
+}
+
+/* unit test for merge function */
+int main(void) {
+	/* contrary to popular practice, malloc should not be casted
+		as void* is casted safely automatically */
+
+	/* first test: merge two halves of the full array */
+	{
+		printf("Test 1: merge two havles of full array\n");
+		A = malloc(10 * sizeof(int));
+		B = malloc(10 * sizeof(int));
+
+		A[0] = 0; /* leftstart = 0*/
+		A[1] = 1;
+		A[2] = 2;
+		A[3] = 3;
+		A[4] = 4; /* leftend = 4*/
+
+		A[5] = 3; /* rightstart = 5 */
+		A[6] = 4;
+		A[7] = 5;
+		A[8] = 6;
+		A[9] = 7; /* rightend = 9 */
+
+		/* expected result is A=[0,1,2,3,3,4,4,5,6,7] */
+		merge(0,4,5,9);
+
+		assert(A[0] == 0);
+		assert(A[1] == 1);
+		assert(A[2] == 2);
+		assert(A[3] == 3);
+		assert(A[4] == 3);
+		assert(A[5] == 4);
+		assert(A[6] == 4);
+		assert(A[7] == 5);
+		assert(A[8] == 6);
+		assert(A[9] == 7);
+
+		printf("Test 1: Passed\n");
+	}
+
+
+	/* first test: merge two halves of the full array */
+	{
+		printf("Test 2: merge two havles of sub-array\n");
+		A = malloc(10 * sizeof(int));
+		B = malloc(10 * sizeof(int));
+
+		A[0] = 5; 
+		A[1] = 6;
+		A[2] = 7;
+		A[3] = 8;
+		A[4] = 9; 
+
+		A[5] = 10;	/* leftstart = 5 */
+		A[6] = 11;	/* leftend = 6 */
+		A[7] = 2;	/* rightstart = 7 */
+		A[8] = 3;
+		A[9] = 4;	/* rightend = 9 */
+
+		/* expected result is A=[2,3,4,10,11] */
+		merge(5,6,7,9);
+
+		assert(A[5] == 2);
+		assert(A[6] == 3);
+		assert(A[7] == 4);
+		assert(A[8] == 10);
+		assert(A[9] == 11);
+
+		printf("Test 2: Passed\n");
+	}
+}

merge_test.make

@@ -0,0 +1,50 @@
+.POSIX:
+.SUFFIXES:
+
+# Use the system's default C compiler
+# was gnu89 reallyyyy necessary ....
+# surely gradescope machines have a compiler which can do post Y2K C ....
+
+CC		= cc
+CFLAGS	= -Wall -Wpointer-arith -Wstrict-prototypes -std=c11 -fPIC
+
+# Pathname of the pkg-config compatible utility
+# (not using this for this assignment at all)
+#PC		= pkg-config
+
+# Instructions to create an LSP db -- default is clangd with compile_flags.txt
+# e.g. To use bear you can set DB="bear -- make" which will create compile_commands.json
+DB		= echo $(LDFLAGS) $(CFLAGS) | tr ' ' '\n' > compile_flags.txt
+
+# Project Files
+BIN=merge_test.test
+OBJ=merge_test.o mergesort.o
+SRC=$(OBJ:%.o=%.c)
+DEP=$(OBJ:%.o=%.d)
+
+all: $(BIN)
+
+# Build binary from objects
+$(BIN): $(OBJ)
+	$(CC) $(CFLAGS) -o $@ $(OBJ) $(LDFLAGS)
+
+# Suffix rules to create .o and .d files from sources
+.SUFFIXES: .c .o
+.c.o:
+	$(CC) $(CFLAGS) $(INC) -c $<
+
+.SUFFIXES: .c .d
+.c.d:
+	$(CC) -MM $< -o $@
+
+# Generate LSP database on each clean
+db:
+	$(DB)
+
+clean: db
+	rm -fr $(BIN)
+	rm -fr $(OBJ)
+	rm -fr $(DEP)
+
+# For FreeBSD make use -include
+-include $(DEP)