refactor

README.md

@@ -0,0 +1,21 @@
+```
+// Generates a random.txt file with parameters specified, and exits
+// parameters besides N, are only used internally to specify how big are the chunks written
+<main> -g -N 100000 -b 10 -n 101 -o <filename>
+
+// Sort the input file, using the temporary dir, and the parameters specified, save to the output file
+<main> -i <input-file> -o <output-file> -d <tmp dir> -N 100000 -b 10 -n 101
+
+// If no option is specified, system will prompt for parameters, and use the input data from stdin
+<main>
+```
+
+- option `-i` implies the program is supposed to sort something, and thus is incompatible with `-g`
+- when a parameter is not specified the default values are used:
+  - i = "1.in"
+  - o = "1.out"
+  - d = "tmp"
+  - g isn't set (false)
+  - N = 100000
+  - b = 10
+  - n = 101

buffer.c

@@ -0,0 +1,43 @@
+#include "buffer.h"
+#include "record.h"
+#include "status_codes.h"
+
+buffer create_buffer(int capacity)
+{
+    buffer out;
+    out.length = 0;
+    out.capacity = capacity;
+    out.location = malloc(sizeof(record) * capacity);
+    return out;
+}
+
+int read_buffer(buffer* buff, FILE* in)
+{
+    int status;
+    record* current = buff->location;
+    buff->length = 0;
+    for (int i = 0; i < buff->capacity; i++) {
+        status = read_record(in, current);
+        if (status != SUCCESS) {
+            return status;
+        }
+        buff->length++;
+        current++;
+    }
+    return SUCCESS;
+}
+
+int write_buffer(buffer* buff, FILE* out)
+{
+    int status;
+    record* current = buff->location;
+    for (int i = 0; i < buff->length; i++) {
+        status = print_record(out, current);
+        if (status < 0) {
+            return status;
+        }
+        fprintf(out, "\n");
+        current++;
+    }
+    return SUCCESS;
+}

buffer.h

@@ -0,0 +1,12 @@
+#pragma once
+#include "record.h"
+
+typedef struct {
+    record* location;
+    int length;
+    int capacity;
+} buffer;
+
+buffer create_buffer(int capacity);
+int read_buffer(buffer* buff, FILE* in);
+int write_buffer(buffer* buff, FILE* in);

config.c

@@ -0,0 +1,52 @@
+#include "config.h"
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <unistd.h>
+
+int process_args(int argc, char** argv, Configuration* opts)
+{
+    opts->input_file = DEFAULT_IN;
+    opts->output_file = DEFAULT_OUT;
+    opts->directory = DEFAULT_DIR;
+    opts->generate_data = false;
+    opts->N = DEFAULT_NN;
+    opts->n = DEFAULT_N;
+    opts->b = DEFAULT_B;
+    int opt;
+    bool input_set = false;
+    while ((opt = getopt(argc, argv, "i:o:d:gN:n:b:")) != -1) {
+        switch (opt) {
+        case 'i':
+            opts->input_file = optarg;
+            input_set = true;
+            break;
+        case 'o':
+            opts->output_file = optarg;
+            break;
+        case 'd':
+            opts->directory = optarg;
+            break;
+        case 'g':
+            opts->generate_data = true;
+            break;
+        case 'N':
+            opts->N = atoi(optarg);
+            break;
+        case 'n':
+            opts->n = atoi(optarg);
+            break;
+        case 'b':
+            opts->b = atoi(optarg);
+            break;
+        case '?':
+            printf("wrong argument\n");
+            return ERROR;
+        }
+    }
+    if (opts->generate_data && input_set) {
+        printf("incompatible arguments\n");
+        return ERROR;
+    }
+    return SUCCESS;
+}

config.h

@@ -0,0 +1,23 @@
+#pragma once
+#include "status_codes.h"
+#include <stdbool.h>
+
+#define DEFAULT_IN "1.in"
+#define DEFAULT_OUT "1.in"
+#define DEFAULT_DIR "tmp"
+#define DEFAULT_NN 100000
+#define DEFAULT_N 101
+#define DEFAULT_B 10
+
+typedef struct
+{
+    const char* input_file;
+    const char* output_file;
+    const char* directory;
+    bool generate_data;
+    int N; // number of records in a file
+    int b; // blocking factor
+    int n; // number of buffers
+} Configuration;
+
+int process_args(int argc, char** argv, Configuration* opts);

main.c

@@ -1,167 +1,62 @@
+#include "buffer.h"
+#include "config.h"
 #include "record.h"
-#include "tape.h"
-#include <stdbool.h>
-#include <stdio.h>
-#include <stdlib.h>
+#include "status_codes.h"
 #include <sys/stat.h>
 #include <sys/types.h>
 #include <time.h>
 #include <unistd.h>
 
-#define DEFAULT_FILE "tests/1.in"
-#define DEFAULT_DIR "tmp"
-#define DEFAULT_NN 100000
-#define DEFAULT_N 101
-#define DEFAULT_B 10
-#define END_PROGRAM -1
-#define SUCCESS 0
-#define ERROR -2
-
-void print_usage()
+int generate_file(int N, const char* filename)
 {
-    printf("Options:\n");
-    printf("  -f [filename]\tRead/Generate records from [filename], default %s\n", DEFAULT_FILE);
-    printf("  -d [dirname]\tSet the directory used for temporary files, default %s\n", DEFAULT_DIR);
-    printf("  -N [count]\tSet total number of records, default %d\n", DEFAULT_NN);
-    printf("  -n [count]\tSet blocking factor, default %d\n", DEFAULT_N);
-    printf("  -b [count]\tSet number of buffers, default %d\n", DEFAULT_B);
-    printf("  -r\t randomly generate N records\n");
-    printf("  -h\tShow this help message\n");
-}
+    srand(time(0));
+    FILE* f = fopen(filename, "w");
+    record rec;
 
-typedef struct
-{
-    FILE* input;
-    char* filename;
-    char* dirname;
-    bool generate_data;
-    bool manual_input;
-    int N; // number of records in a file
-    int b; // blocking factor
-    int n; // number of buffers
-} options;
-
-// reads args into opts
-int process_args(int argc, char** argv, options* opts)
-{
-    if (opts == NULL) {
-        return ERROR;
-    }
-    opts->input = stdin;
-    opts->filename = DEFAULT_FILE;
-    opts->dirname = DEFAULT_DIR;
-    opts->generate_data = false;
-    opts->manual_input = false;
-    opts->N = DEFAULT_NN;
-    opts->n = DEFAULT_N;
-    opts->b = DEFAULT_B;
-    int opt;
-    while ((opt = getopt(argc, argv, "f:d:N:n:b:rhm")) != -1) {
-        switch (opt) {
-        case 'f':
-            opts->filename = optarg;
-            break;
-        case 'd':
-            opts->dirname = optarg;
-            break;
-        case 'r':
-            opts->generate_data = true;
-            break;
-        case 'N':
-            opts->N = atoi(optarg);
-            break;
-        case 'n':
-            opts->n = atoi(optarg);
-            break;
-        case 'b':
-            opts->b = atoi(optarg);
-            break;
-        case 'm':
-            opts->manual_input = true;
-            break;
-        case 'h':
-            print_usage();
-            return END_PROGRAM;
-        case '?':
-            fprintf(stderr, "invalid option, terminating\n");
-            return ERROR;
-        }
-    }
-    if (opts->manual_input == true && opts->generate_data == true) {
-        fprintf(stderr, "Incompatible arguments: mr, aborting\n");
-        return ERROR;
-    }
-    opts->input = fopen(opts->filename, "r+");
-    if (!opts->input) {
-        opts->input = fopen(opts->filename, "w+");
-    }
-    if (!opts->input) {
-        fprintf(stderr, "Error: Cannot open file '%s'\n", opts->filename);
-        return ERROR;
-    }
-    struct stat st = { 0 };
-    if (stat(opts->dirname, &st) == -1) {
-        mkdir(opts->dirname, 0777);
+    for (int i = 0; i < N; i++) {
+        random_record(&rec);
+        print_record(f, &rec);
+        fprintf(f, "\n");
     }
     return SUCCESS;
 }
 
-void generate_file(options* opts)
+int sort_file(Configuration* opts)
 {
-    srand(time(0));
-    record rec;
-    for (int i = 0; i < opts->N; i++) {
-        random_record(&rec);
-        print_record(opts->input, &rec);
-        fprintf(opts->input, "\n");
-    }
-    fseek(opts->input, 0, SEEK_SET);
-}
+    // create buffer
+    buffer buff = create_buffer(opts->b * opts->n);
+    // open file
+    FILE* in = fopen(opts->input_file, "r");
+    // <stage 1>
+    // while !eof:
+    for (int i = 0; true; i++) {
+        //     read buffer
+        int status = read_buffer(&buff, in);
+        printf("buffer i:%d, status: %d\n", i, status);
+        //     sort buffer
 
-void write_file(options* opts)
-{
+        //     write run
+        write_buffer(&buff, stdout);
+        if (status != SUCCESS) {
+            break;
         }
-
-int sort_block(tape* t, buffer* buff)
-{
-    static int counter = 0;
-    int status = read_buffer(buff, t);
-    if (status != SUCCESS && status != EOF) {
-        fprintf(stderr, "Error reading file: %d\n", status);
-        return status;
     }
-
-    sort_buffer(buff);
-    printf("sorted the buffer %d, the results:\n", counter);
-    print_buffer(buff);
-    counter++;
-    return status;
-}
-
-int sort_file(options* opts)
-{
-    tape a = { 0, 0, opts->input };
-    buffer buff;
-    buff.capacity = opts->b * opts->n;
-    buff.location = malloc(sizeof(record) * buff.capacity);
-    int status = 0;
-    while (status == SUCCESS) {
-        status = sort_block(&a, &buff);
-    }
-    if (status != EOF) {
+    // <stage 2>
+    // while tmp:
+    //     merge b-1 runs
+    //
+    // close file
+    // free buffer
+    fclose(in);
     free(buff.location);
-        fclose(a.file);
-        return status;
-    }
-    free(buff.location);
-    fclose(a.file);
     return SUCCESS;
 }
 
 int main(int argc, char** argv)
 {
-    options opts;
+    Configuration opts;
     if (argc <= 1) {
+        // #todo: stdin input
         fprintf(stderr, "please specify input args\n");
         return ERROR;
     }
@@ -170,12 +65,42 @@ int main(int argc, char** argv)
         return result;
     }
     if (opts.generate_data) {
-        generate_file(&opts);
-        return SUCCESS;
-    }
-    if (opts.manual_input) {
-        write_file(&opts);
+        return generate_file(opts.N, opts.output_file);
     }
     sort_file(&opts);
-    fclose(opts.input);
 }
+// int sort_block(tape* t, buffer* buff)
+// {
+//     static int counter = 0;
+//     int status = read_buffer(buff, t);
+//     if (status != SUCCESS && status != EOF) {
+//         fprintf(stderr, "Error reading file: %d\n", status);
+//         return status;
+//     }
+
+//     sort_buffer(buff);
+//     printf("sorted the buffer %d, the results:\n", counter);
+//     print_buffer(buff);
+//     counter++;
+//     return status;
+// }
+
+// int sort_file(options* opts)
+// {
+//     tape a = { 0, 0, opts->input };
+//     buffer buff;
+//     buff.capacity = opts->b * opts->n;
+//     buff.location = malloc(sizeof(record) * buff.capacity);
+//     int status = 0;
+//     while (status == SUCCESS) {
+//         status = sort_block(&a, &buff);
+//     }
+//     if (status != EOF) {
+//         free(buff.location);
+//         fclose(a.file);
+//         return status;
+//     }
+//     free(buff.location);
+//     fclose(a.file);
+//     return SUCCESS;
+// }

makefile

@@ -1,30 +1,33 @@
 CFLAGS = -g
 
-build/main: build/main.o build/tape.o build
-	clang $(CFLAGS) build/main.o build/tape.o -o build/main
+build/main: build/main.o build/config.o build/buffer.o
+	clang $(CFLAGS) build/main.o  build/config.o build/buffer.o -o build/main
 
 build/main.o: main.c build
 	clang $(CFLAGS) -c main.c -o build/main.o
 
-build/tape.o: tape.c build
-	clang $(CFLAGS) -c tape.c -o build/tape.o
+build/config.o: config.c build
+	clang $(CFLAGS) -c config.c -o build/config.o
+
+build/buffer.o: buffer.c build
+	clang $(CFLAGS) -c buffer.c -o build/buffer.o
 
 build:
 	mkdir -p build
 
-tests:
-	mkdir -p tests
+build/run: build
+	mkdir -p build/run
 
 clean:
-	rm -rf build tests tmp
+	rm -rf build
 
-run: build/main tests/1.in
-	build/main -f tests/1.in | less
+build/run/1.in: build/run build/main
+	build/main -g -o build/run/1.in
+
+run: build/main build/run/1.in
+	build/main -i build/run/1.in | less
 
 debug: build/main tests/1.in
-	lldb -- build/main -f tests/1.in
-
-tests/1.in: build/main tests
-	build/main -f tests/1.in -r  
+	lldb -- build/main
 
 .PHONY: clean run debug

record.h

@@ -28,9 +28,9 @@ static inline int compare_records(const void* a, const void* b)
     return result;
 }
 
-static inline void print_record(FILE* stream, const record* out)
+static inline int print_record(FILE* stream, const record* out)
 {
-    fprintf(stream, "%d %d %d %d %d %d",
+    return fprintf(stream, "%d %d %d %d %d %d",
         out->a[0], out->a[1], out->a[2], out->a[3], out->a[4], out->x);
 }
 

status_codes.h

@@ -0,0 +1,5 @@
+#pragma once
+
+#define END_PROGRAM -1
+#define SUCCESS 0
+#define ERROR -2

tape.c

@@ -1,108 +0,0 @@
-#include "record.h"
-#include "tape.h"
-
-int partition(record* arr, int start, int end)
-{
-    int i = start;
-    record* pivot = arr;
-    pivot += end;
-    for (int j = start; j < end; j++) {
-        record* current = arr;
-        current += j;
-        if (compare_records(current, pivot) > 0) {
-            continue;
-        }
-        record* i_rec = arr;
-        i_rec += i;
-        record tmp = *i_rec;
-        *i_rec = *current;
-        *current = tmp;
-        i++;
-    }
-    record* i_rec = arr;
-    i_rec += i;
-    record tmp = *i_rec;
-    *i_rec = *pivot;
-    *pivot = tmp;
-    return i;
-}
-
-void quicksort(record* arr, int start, int end, int depth)
-{
-    if (start >= end || start < 0) {
-        return;
-    }
-
-    int p = partition(arr, start, end);
-    quicksort(arr, start, p - 1, depth + 1);
-    quicksort(arr, p + 1, end, depth + 1);
-}
-
-void sort_buffer(buffer* buff)
-{
-    quicksort(buff->location, 0, buff->length - 1, 0);
-}
-
-int read_buffer(buffer* buff, tape* tape)
-{
-    int status;
-    record* destination = buff->location;
-    buff->length = 0;
-    while (buff->capacity > buff->length) {
-        status = read_record(tape->file, destination);
-        if (status != 0) {
-            return status;
-        }
-        destination++;
-        buff->length++;
-        tape->reads++;
-    }
-    return 0;
-}
-
-void print_full_buffer(buffer* buff)
-{
-    for (int i = 0; i < buff->length; i++) {
-        record* current = buff->location;
-        current += i;
-        printf("i: %d\tg: %d\t", i, g(current));
-        print_record(stdout, current);
-        printf("\n");
-    }
-}
-
-void print_buffer(buffer* buff)
-{
-    if (buff->length <= 10) {
-        print_full_buffer(buff);
-        return;
-    }
-    for (int i = 0; i < 5; i++) {
-        record* current = buff->location;
-        current += i;
-        printf("i: %d\tg: %d\t", i, g(current));
-        print_record(stdout, current);
-        printf("\n");
-    }
-    for (int i = buff->length - 5; i < buff->length; i++) {
-        record* current = buff->location;
-        current += i;
-        printf("i: %d\tg: %d\t", i, g(current));
-        print_record(stdout, current);
-        printf("\n");
-    }
-}
-
-int write_buffer(buffer* buff, tape* tape)
-{
-    record* current = buff->location;
-    record* last = current;
-    last += buff->length;
-    while (current < last) {
-        print_record(tape->file, current);
-        fprintf(tape->file, "\n");
-        current++;
-        tape->reads++;
-    }
-    return 0;
-}

tape.h

@@ -1,26 +0,0 @@
-#pragma once
-#include <stdio.h>
-
-typedef struct {
-    int reads;
-    int writes;
-    FILE* file;
-} tape;
-
-typedef struct {
-    void* location;
-    int length;
-    int capacity;
-} buffer;
-
-// sorts the entire buffer
-void sort_buffer(buffer* buff);
-
-// reads [capacity] records
-int read_buffer(buffer* buff, tape* tape);
-// writes everything
-int write_buffer(buffer* buff, tape* tape);
-// for debugging
-void print_buffer(buffer* buff);
-
-void clear_tape(tape* tape);