lazysphere/command/ed.c

#include "command.h"
#include "lslib.h"

#include <stdlib.h>
#include <string.h>

#define INPUT_LEN 1024

static char** lines = NULL;
static unsigned long line_capacity;
static unsigned long line_count;
static unsigned long line_current;
static bool pending_writes;
static char* default_filename = NULL;
static re_t last_regex = NULL;

enum LineAddressType {
    INDEX,
    RANGE,
    SET,
    FREE
};

struct LineAddress {
    enum LineAddressType type;
    union {
        struct {
            long int i;
        } index;
        struct {
            long int a;
            long int b;
        } range;
        struct {
            long int* b;
            unsigned long c;
            unsigned long s;
        } set;
    } data;
    bool empty;
};

enum RegexDirection {
    BEFORE,
    AFTER,
    ALL
};

static bool read_regex(char** end, re_t* regex, enum RegexDirection dir) {
    char c;
    char* index = *end;
    char* regex_str = index;
    while(true) {
        c = *(index++);
        if (c == '\0') {
            error_s("missing regex after %c\n", dir == BEFORE ? '?' : '/');
            return false;
        }
        if (c == (dir == BEFORE ? '?' : '/')) {
            *(index - 1) = '\0';
            break;
        }
    }
    *regex = re_compile(regex_str);
    last_regex = *regex;
    *end = index;
    return true;
}

static bool parse_regex(char** end, struct LineAddress* address, enum RegexDirection dir) {
    re_t regex;
    unsigned long cap, siz, i, until;
    long int* buf;

    if (!read_regex(end, &regex, dir)) return false;

    cap = 8;
    siz = 0;
    buf = xalloc(cap * sizeof(long int));

    i = (dir == ALL ? 0 : line_current);
    until = (dir == BEFORE ? 0 : line_count - 1);
    for (; (dir == BEFORE ? i >= until : i < until); dir == BEFORE ? i-- : i++) {
        int len;
        if (re_matchp(regex, lines[i], &len) == -1) {
            if (dir == BEFORE && i == 0) break;
            continue;
        }
        if (cap == siz) {
            cap *= 2;
            buf = xrealloc(buf, cap * sizeof(long int));
        }
        buf[siz] = i;
        siz++;
        if (dir == BEFORE && i == 0) break;
    }

    address->empty = false;
    address->type = SET;
    address->data.set.s = siz;
    address->data.set.c = cap;
    address->data.set.b = buf;
    return true;
}

static void free_address (struct LineAddress address) {
    if (address.type != SET) return;
    address.type = FREE;
    free(address.data.set.b);
}

static void expand_buffer(long int** buf, unsigned long* cap, unsigned long* size) {
    if (*cap == *size) {
        *cap *= 2;
        *buf = xrealloc(*buf, sizeof(long int) * *cap);
    }
}

static bool parse_regex_lines(char** end, struct LineAddress* address) {
    re_t regex;
    unsigned long cap, siz;
    long int* buf;
    int len;
    struct LineAddress addr;

    if (!read_regex(end, &regex, ALL)) return false;

    cap = 8;
    siz = 0;
    buf = xalloc(cap * sizeof(long int));

    addr = *address;
    if (addr.type == INDEX) {
        if (re_matchp(regex, lines[addr.data.index.i], &len) != -1) {
            buf[0] = addr.data.index.i;
            siz = 1;
        }
    } else if (addr.type == RANGE) {
        long int i;
        for (i = addr.data.range.a; i < addr.data.range.b; i++) {
            if (re_matchp(regex, lines[i], &len) == -1) continue;
            expand_buffer(&buf, &cap, &siz);
            buf[siz] = i;
            siz++;
        }
    } else if (addr.type == SET) {
        unsigned long i;
        for (i = 0; i < addr.data.set.s; i++) {
            if (re_matchp(regex, lines[addr.data.set.b[i]], &len) == -1) continue;
            expand_buffer(&buf, &cap, &siz);
            buf[siz] = addr.data.set.b[i];
            siz++;
        }
    }

    free_address(*address);
    address->empty = false;
    address->type = SET;
    address->data.set.s = siz;
    address->data.set.c = cap;
    address->data.set.b = buf;
    return true;
}

static bool read_address(char** command, bool whitespace, struct LineAddress* a) {
    char* index = *command;
    struct LineAddress address;

    char* end_pre;
    long int n_pre;
    char pre;

    memset(&address, 0, sizeof(struct LineAddress));
    
    address.empty = false;
    if (strlen(*command) < 1) {
        address.type = INDEX;
        address.data.index.i = line_current + 1;
        if (line_current >= line_count) line_current = line_count - 1;
        *a = address;
        return true;
    }

    n_pre = strtol(index, &end_pre, 10) - 1;
    if (end_pre == index) {
        n_pre = -1;
    } else {
        if (n_pre < 0) {
            error_s("input cannot be negative\n");
            return false;
        }
        index = end_pre;
    }
    
    pre = *(index++);
    switch (pre) {
        case '.':
            address.type = INDEX;
            address.data.index.i = line_current;
            break;
        case '$':
            address.type = INDEX;
            address.data.index.i = line_count - 1;
            break;
        case '-':
        case '^': {
            char* end;
            long int n;

            address.type = INDEX;
            n = strtol(index, &end, 10) - 1;

            if (n < 0) {
                error_s("input cannot be negative\n");
                return false;
            }

            if (index == end) {
                address.data.index.i = line_current - 1;
            } else {
                address.data.index.i = line_current - n;
            }

            if (address.data.index.i < 0) {
                error_s("line number %ld does not exist\n", address.data.index.i + 1);
                return false;
            }

            break;
        }
        case '+': {
            char* end;
            long int n;

            address.type = INDEX;
            n = strtol(index, &end, 10) - 1;

            if (n < 0) {
                error_s("input cannot be negative\n");
                return false;
            }
            if (index == end) {
                address.data.index.i = line_current + 1;
            } else {
                address.data.index.i = line_current + n;
            }
            if (address.data.index.i >= (long int) line_count) {
                error_s("line number %ld does not exist\n", address.data.index.i + 1);
                return false;
            }
            break;
        }
        case '%':
            address.type = RANGE;
            address.data.range.a = 0;
            address.data.range.b = line_count - 1;
            break;
        case ';':
            address.type = RANGE;
            address.data.range.a = line_current;
            address.data.range.b = line_count - 1;
            break;
        case '/': 
            if (!parse_regex(&index, &address, AFTER)) return false;
            break;
        case '?':
            if (!parse_regex(&index, &address, BEFORE)) return false;
            break;
        default: {
            index--;
            if (n_pre == -1) {
                address.type = INDEX;
                address.data.index.i = line_current;
                address.empty = true;
                break;
            } else if (whitespace) {
                address.type = INDEX;
                address.data.index.i = line_current + n_pre;
            } else {
                address.type = INDEX;
                address.data.index.i = n_pre;
            }
            if (address.data.index.i < 0 || address.data.index.i >= (long int) line_count) {
                error_s("line number %ld does not exist\n", address.data.index.i + 1);
                return false;
            }
        }
    }
    *command = index;
    *a = address;
    return true;
}

static void free_data(bool all) {
    if (lines != NULL) {
        unsigned long i;
        for (i = 0; i < line_count; i++) {
            free(lines[i]);
        }
        free(lines);
        lines = NULL;
    }
    if (all && default_filename != NULL) {
        free(default_filename);
        default_filename = NULL;
    }
}

static void load_empty(void) {
    free_data(false);

    line_capacity = 8;
    lines = xalloc(sizeof(char*) * line_capacity);

    line_count = 0;
    line_current = 0;
    pending_writes = false;
}

static void get_input(FILE* file, char*** buffer, unsigned long* capacity, unsigned long* size) {
    unsigned long cap = 8;
    unsigned long siz = 0;
    char** buf = xalloc(sizeof(char*) * cap);

    char* line = NULL;
    size_t offset = 0;
    
    clearerr(stdin);
    while (getline(&line, &offset, file) != -1) {
        if (cap == siz) {
            cap *= 2;
            buf = xrealloc(buf, sizeof(char*) * cap);
        }
        buf[siz] = line;
        siz++;
        line = NULL;
    }

    free(line);

    *buffer = buf;
    *capacity = cap;
    *size = siz;
}

int ed_getline(char *buf, size_t size) {
    size_t i = 0;
    int ch;
    clearerr(stdin);
    while ((ch = getchar()) != EOF) {  /* Read until EOF ... */
        if (i + 1 < size) {
            buf[i++] = ch;
        }
        if (ch == '\n') {  /* ... or end of line */
            break;  
        }
    } 
    buf[i] = '\0';
    if (i == 0) { 
        return EOF;
    }
    return i;
}

static void load_file(FILE* file) {
    free_data(false);
    line_current = 0;
    get_input(file, &lines, &line_capacity, &line_count);
    if (file != stdin)
        fclose(file);
    
}

static bool check_if_sure(char* prompt) {
    char buf[INPUT_LEN];

    if (!pending_writes) {
        return true;
    }

    printf("%s", prompt);
    fflush(stdout);

    if (ed_getline(buf, INPUT_LEN) == EOF) {
        putchar('\n');
        return false;
    }

    return prefix("y", buf);
}

static bool skip_whitespace(char** index) {
    char c;
    bool w = false;
    while (c = **index, c == ' ' || c == '\t') { (*index)++; w = true; }
    return w;
}

static void expand(unsigned long count) {
    if (count < line_capacity) return;
    line_capacity *= 2;
    if (count > line_capacity) line_capacity = count;
    lines = xrealloc(lines, line_capacity * sizeof(char*));
}

static void append_lines(unsigned long index, char** new, unsigned long new_len) {
    if (new_len < 1) return;
    pending_writes = true;
    expand(line_count + new_len);
    if (index + 1 <= line_count)
        memmove(&lines[index+new_len], &lines[index], sizeof(char*) * (line_count - index));
    memcpy(&lines[index], new, sizeof(char*) * new_len);
    line_count += new_len;
}

static void delete_lines(unsigned long a, unsigned long b) {
    unsigned long i;

    if (b < a) return;
    pending_writes = true;
    
    for (i = a; i <= b; i++) {
        free(lines[i]);
    }
    if (b == line_count - 1) {
        line_count = a;
        return;
    }
    memmove(&lines[a], &lines[b+1], sizeof(char*) * (line_count - (b + 1)));
    line_count -= (b - a) + 1;
    line_current = a;
    if (line_current >= line_count) line_current = line_count - 1;
}

static bool handle_append(struct LineAddress* address) {
    char** buf;
    unsigned long cap, size;

    if (address->type != INDEX) {
        error_s("append command requires index addressing\n");
        return false;
    }
    
    if (line_count == 0) {
        address->data.index.i = -1;
    }
    
    get_input(stdin, &buf, &cap, &size);
    
    if (size > 0) {
        append_lines(address->data.index.i + 1, buf, size);
        printf("ed: appened %lu lines\n", size);
    }
    
    line_current += size;
    free(buf);
    return true;
}

static bool handle_delete(struct LineAddress* address) {
    if (address->empty && address->data.index.i >= (long int) line_count) {
        error_s("line number %ld does not exist\n", address->data.index.i + 1);
        return false;
    }

    if (address->type == INDEX) {
        delete_lines(address->data.index.i, address->data.index.i);
        output("deleted line %ld\n", address->data.index.i+1);
    } else if (address->type == RANGE) {
        delete_lines(address->data.range.a, address->data.range.b);
        output("deleted lines %ld-%ld\n", address->data.range.a+1, address->data.range.b+1);
    } else if (address->type == SET) {
        unsigned long i;
        for (i = 0; i < address->data.set.s; i++) {
            delete_lines(address->data.set.b[i], address->data.set.b[i]);
        }
        output("deleted %lu lines\n", address->data.set.s);
    }
    
    return true;
}

static bool get_file_name(char** filename) {
    size_t len = strlen(*filename);

    if (len < 1 || (len == 1 && **filename == '\n')) {
        if (default_filename == NULL) {
            error_s("no default filename specified\n");
            return false;
        }
        *filename = default_filename;
    } else {
        if ((*filename)[len - 1] == '\n') {
            (*filename)[len - 1] = '\0';
            len--;
        }
        if (default_filename != NULL) {
            default_filename = xrealloc(default_filename, len + 1);
        } else {
            default_filename = xalloc(len + 1);
        }
        memcpy(default_filename, *filename, len + 1);
        *filename = default_filename;
    }
    return true;
}

static void write_file(char* filename, struct LineAddress* address, char* type) {
    FILE* file;
    int wrote;

    if (line_count < 1) {
        error_s("cannot write empty file\n");
        return;
    }
    
    if (!get_file_name(&filename)) return;
    file = get_file_s(filename, type);
    if (file == NULL) return;
    
    wrote = 0;
    
    if (address->empty) {
        unsigned long i;
        for (i = 0; i < line_count; i++) {
            fprintf(file, "%s", lines[i]); 
            wrote++;
        }
    } else if (address->type == INDEX) {
        long int i = address->data.index.i;
        fprintf(file, "%s", lines[i]); 
        wrote++;
    } else if (address->type == RANGE) {
        long int i;
        for (i = address->data.range.a; i < address->data.range.b; i++) {
            fprintf(file, "%s", lines[i]); 
            wrote++;
        }
    } else if (address->type == SET) {
        unsigned long i;
        for (i = 0; i < address->data.set.s; i++) {
            fprintf(file, "%s", lines[address->data.set.b[i]]);
            wrote++;
        }
    }

    pending_writes = false;
    fclose(file);
    output("wrote %d lines from %s\n", wrote, filename);
}

static void read_file(char* filename) {
    FILE* file;
    char** buf;
    long int line;
    unsigned long capacity, size;

    if (!get_file_name(&filename)) return;
    file = get_file_s(filename, "r");
    if (file == NULL) return;

    capacity = 8;
    size = 0;
    buf = xalloc(capacity * sizeof(char*));
    get_input(file, &buf, &capacity, &size);

    if (size < 1) {
        free(buf);
        error_s("attempted to read a empty file\n");
        return;
    }

    line = -1;
    if (line_count > 0) {
        line = line_count - 1;
    }

    append_lines(line, buf, size);
    free(buf);
    output("read and appended %lu lines from %s\n", size, filename);
}

static void expand_string(char** buf, int* capacity, int* size, char* text, int len) {
    if (*size + len >= *capacity) {
        *capacity *= 2;
        if (*capacity < *size + len) *capacity = *size + len;
        *buf = xrealloc(*buf, *capacity);
    }
    memcpy(*buf + *size, text, len);
    *size += len;
}

static int substute_string(long int index, long int matches, re_t regex, char* sub, int sub_len) {
    int capacity = 8;
    int size = 0;
    char* buf = xalloc(sizeof(char) * capacity);
    long int left;

    int offset = 0;
    int matches_found = 0;
    while(true) {
        int distance, len;
        
        if (lines[index][offset] == '\0') break;
        
        if (matches_found >= matches && matches > 0) break;
        
        if ((distance = re_matchp(regex, &lines[index][offset], &len)) == -1) {
            break;
        }
        
        if (distance > 0) {
            expand_string(&buf, &capacity, &size, &lines[index][offset], distance);
        }

        expand_string(&buf, &capacity, &size, sub, sub_len);
        offset += len + distance;
        matches_found++;
    }

    left = strlen(lines[index] + offset);
    expand_string(&buf, &capacity, &size, &lines[index][offset], left + 1);

    free(lines[index]);
    lines[index] = buf;
    return matches_found;
}

static void prompt(void) {
    char buf[INPUT_LEN];
    char* index;
    char cmd;
    bool whitespace, linenumbers;
    struct LineAddress address;
    
    printf(": ");
    fflush(stdout);

    if (ed_getline(buf, INPUT_LEN) == EOF) { putchar('\n'); return; }
    if (buf[0] == '\0') { putchar('\n'); return; }

    index = &buf[0];
    whitespace = skip_whitespace(&index);

    if (!read_address(&index, whitespace, &address)) return;

    cmd = *(index++);
    
    if (cmd == ',') {
        struct LineAddress address2;
        
        if (address.type != INDEX) {
            error_s("comma range addressing requires two index addresses\n");
            free_address(address);
            return;
        }
        
        whitespace = skip_whitespace(&index);
        
        if (!read_address(&index, whitespace, &address2)) {
            free_address(address);
            return;
        }
        
        if (address2.type != INDEX) {
            error_s("comma range addressing requires two index addresses\n");
            free_address(address);
            free_address(address2);
            return;
        }
        
        address.type = RANGE;
        address.data.range.a = address.data.index.i; /* cursed */
        address.data.range.b = address2.data.index.i;

        cmd = *(index++);
    }

    if (address.type == RANGE && address.data.range.a > address.data.range.b) {
        error_s("range addressing must be in ascending order\n");
        free_address(address);
        return;
    }

    linenumbers = false;

test:
    switch (cmd) {
        case '\0':
        case '\n':
            if (address.empty) {
                if (line_current == line_count) {
                    error_s("line number %ld does not exist\n", line_current + 1);
                    break;
                } else if (line_current + 1 == line_count) {
                    error_s("line number %ld does not exist\n", line_current + 2);
                    break;
                } else {
                    line_current++;
                }
                printf("%s", lines[line_current]);
                break;
            }
            if (address.type == INDEX) {
                line_current = address.data.index.i;
            } else if (address.type == RANGE) {
                line_current = address.data.range.b;
            } else if (address.type == SET) {
                error_s("unexpected range addressing\n");
                break;
            }
            printf("%s", lines[line_current]);
            break;
        case 'a':
            handle_append(&address);
            break;
        case 'd':
            handle_delete(&address);
            break;
        case 'c':
            if (!handle_delete(&address)) { break; }
            address.type = INDEX;
            address.data.index.i = line_current - 1;
            handle_append(&address);
            break;
        case 'n':
            linenumbers = true;
            __attribute__((fallthrough));
        case 'p':
            if (address.empty && address.data.index.i >= (long int) line_count) {
                error_s("line number %ld does not exist\n", address.data.index.i + 1);
                break;
            }
            if (address.type == INDEX) {
                if (linenumbers) printf("%ld\t", address.data.index.i + 1);
                printf("%s", lines[address.data.index.i]);
            } else if (address.type == RANGE) {
                long int i;
                for (i = address.data.range.a; i <= address.data.range.b; i++) {
                    if (linenumbers) printf("%ld\t", i + 1);
                    printf("%s", lines[i]);
                }
            } else if (address.type == SET) {
                unsigned long i;
                for (i = 0; i < address.data.set.s; i++) {
                    if (linenumbers) printf("%ld\t", address.data.set.b[i] +1);
                    printf("%s", lines[address.data.set.b[i]]);
                }
            }
            break;
        case 'q':
            if (!check_if_sure("Quit for sure? ")) break;
            __attribute__((fallthrough));
        case 'Q':
            free_address(address);
            free_data(true);
            exit(EXIT_SUCCESS);
            break;
        case 'g':
            skip_whitespace(&index);
            free_address(address);
            if (*(index++) != '/') {
                error_s("unexpected character at start of regex\n");
                break;
            }
            if (!parse_regex(&index, &address, ALL)) { return; }
            skip_whitespace(&index);
            if (*index == '\n' || *index == '\0') {
                cmd = 'p';
            } else {
                cmd = *(index++);
            }
            goto test;
            break;
        case 's': {
            char* replace = index;
            long int matches, sub_len, matches_found;
            unsigned long i;
            
            skip_whitespace(&index);
            if (*(index++) != '/') {
                error_s("unexpected character at start of regex\n");
                break;
            }
            if (!parse_regex_lines(&index, &address)) { return; }
            while(*index != '\0' && *index != '/') index++;
            if (*index != '/') {
                error_s("/ missing after %c\n", *index);
                break;
            }
            if (address.data.set.s < 1) {
                error_s("no matches found\n");
                break;
            }
            *(index++) = '\0';
            if (*index == '\0' || *index == '\n') {
                matches = 1;
            } else if (*index == 'g') {
                matches = -1;
            } else {
                char* end;
                matches = strtol(index, &end, 10);
                if (end == index) {
                    error_s("invalid number: %s\n", index);
                    break;
                }
                if (matches < 1) {
                    error_s("matches cannot be less than 1\n");
                    break;
                }
            }
            sub_len = strlen(replace);
            matches_found = 0;
            
            for (i = 0; i < address.data.set.s; i++) {
                matches_found += substute_string(address.data.set.b[i], matches, last_regex, replace, sub_len);
            }
            output("replaced %ld matches over %ld lines\n", matches_found, address.data.set.s);
            pending_writes = true;
            break;
        }
        case 'w': {
            bool quit = false;
            if (*index == 'q') {
                index++;
                quit = true;
            }
            skip_whitespace(&index);
            write_file(index, &address, "w");
            if (quit) {
                free_address(address);
                free_data(true);
                exit(EXIT_SUCCESS);
            }
            break;
        }
        case 'r': {
            skip_whitespace(&index);
            read_file(index);
            break;
        }
        case 'e': 
            if (!check_if_sure("Load new file for sure? ")) break;
            __attribute__((fallthrough));
        case 'E': {
            char* filename;
            FILE* file;

            skip_whitespace(&index);
            
            filename = index;
            if(!get_file_name(&filename)) break;
            
            file = get_file_s(filename, "r");
            if (file == NULL) break;
            
            load_file(file);
            break;
        }
        case 'W':
            skip_whitespace(&index);
            write_file(index, &address, "a");
            break;
        case '=':
            printf("%ld\n", line_current + 1);
            break;
        default:
            error_s("unimplemented command\n");
            break;
    }

    free_address(address);

}

static void prompt_loop(void) {
    while (true) {
        prompt();
    }
}

static void help(void) {
    printf("Usage: ed [FILE]\n\n");
    printf("Edit a given [FILE] or create a new text file\n\n");
    printf("\t(.,.)\tnewline\t\tgo to address line and print\n");
    printf("\t(.)\ta\t\tappend new data after given line\n");
    printf("\t(.,.)\tc\t\treplace given lines with new data\n");
    printf("\t(.,.)\td\t\tdelete given lines\n");
    printf("\t\te file\t\tdelete current buffer and edit file\n");
    printf("\t\tE file\t\tdelete current buffer and edit file unconditionally\n");
    printf("\t\tg/re/command\tgrep all lines with regex and run command on matches\n");
    printf("\t(.,.)\tn\t\tprint given lines along with their line numbers\n");
    printf("\t(.,.)\tp\t\tprint given lines\n");
    printf("\t\tq\t\tquit file\n");
    printf("\t\tQ\t\tquit file unconditionally\n");
    printf("\t($)\tr file\t\tread file and append to end of buffer\n");
    printf("\t(.,.)\ts/re/replace/\treplace the first match on each matching line\n");
    printf("\t(.,.)\ts/re/replace/g\treplace all matches on each matching line\n");
    printf("\t(.,.)\ts/re/replace/n\treplace n matches on each matching line\n");
    printf("\t(.,.)\tw file\t\twrite contents of lines to file\n");
    printf("\t(.,.)\twq file\t\twrite contents of lines to file then quit\n");
    printf("\t(.,.)\tW file\t\tappend contents of line to file\n");
    printf("\t\t=\t\tprint current line number\n");
}

COMMAND(ed_main) {

    parse_help(argc, argv, help);

    if (argc < 1) {
        load_empty();
        prompt_loop();
    } else {
        FILE* file = get_file(argv[0], "r");
        load_file(file);
        get_file_name(&argv[0]);
        prompt_loop();
    }
    return EXIT_SUCCESS;
}