power_play/src/tar.c

#include "tar.h"
#include "bitbuff.h"
#include "string.h"
#include "util.h"

#define ARCHIVE_LOOKUP_TABLE_CAPACITY_FACTOR 2.0

/* File types:
 * '0' or (ASCII NUL) Normal file
 * '1' Hard link
 * '2' Symbolic link
 * '3' Character special
 * '4' Block special
 * '5' Directory
 * '6' FIFO
 * '7' Contiguous file
 * 'g' Global extended header with meta data(POSIX.1 - 2001)
 * 'x' Extended header with metadata for the next file in the archive(POSIX.1 - 2001)
 * 'A'-'Z' Vendor specific extensions(POSIX.1 - 1988)
 */

#define TAR_TYPE_FILE         '0'
#define TAR_TYPE_DIRECTORY    '5'
#define TAR_TYPE_PAX_HEADER_X 'x'
#define TAR_TYPE_PAX_HEADER_G 'g'

PACK(struct tar_header {
    /* Pre-posix */
    u8 file_name[100];
    u8 file_mode[8];
    u8 owner_id[8];
    u8 group_id[8];
    u8 file_size[12];
    u8 last_modified[12];
    u8 checksum[8];

    /* Both */
    u8 file_type;
    u8 linked_file_name[100];

    /* UStar */
    u8 ustar_indicator[6];
    u8 ustar_version[2];
    u8 owner_user_name[32];
    u8 owner_group_name[32];
    u8 device_major_number[8];
    u8 device_minor_number[8];
    u8 file_name_prefix[155];
    u8 padding[12];
});

INTERNAL u64 str_oct_to_u64(struct string str)
{
    u64 n = 0;
    for (u64 i = 0; i < str.len; ++i) {
        n *= 8;
        n += (u64)(str.text[i]) - '0';
    }
    return n;
}

/* `prefix` will be prepended to all file names in the archive
 *
 * NOTE: The resulting archive merely points into the supplied tar data, no
 * copying is done. Accessing the archive assumes that the data string is still valid.
 */
struct tar_archive tar_parse(struct arena *arena, struct string data, struct string prefix)
{
    __prof;

    struct tar_archive archive = ZI;
    struct bitbuff bb = bitbuff_from_string(data);
    struct bitbuff_reader br = br_from_bitbuff_no_debug(&bb);

    u64 num_files = 0;
    while (br_num_bytes_left(&br) > 1024) {

        struct tar_header header = ZI;
        br_read_bytes(&br, STRING_FROM_STRUCT(&header));

        if (!string_eq(STRING_FROM_ARRAY(header.ustar_indicator), LIT("ustar\0"))) {
            /* Invalid header */
            ASSERT(false);
            continue;
        }

        if (header.file_name_prefix[0] != 0) {
            /* Header file name prefix not supported */
            ASSERT(false);
            continue;
        }

        struct string file_size_oct_str = { .len = 11, .text = header.file_size };

        u64 file_size = str_oct_to_u64(file_size_oct_str);
        u8 *file_data_ptr = br_read_bytes_raw(&br, file_size);
        if (!file_data_ptr) {
            file_size = 0;
        }
        struct string file_data = STRING(file_size, file_data_ptr);

        /* Skip sector padding */
        u64 remaining = (512 - (file_size % 512)) % 512;
        br_seek_bytes(&br, remaining);

        b32 is_dir = header.file_type == TAR_TYPE_DIRECTORY;
        if (!is_dir && header.file_type != TAR_TYPE_FILE) {
            /* Unsupported type */
            ASSERT(header.file_type == TAR_TYPE_PAX_HEADER_X ||
                   header.file_type == TAR_TYPE_PAX_HEADER_G);
            continue;
        }

        struct string file_name_cstr = string_from_cstr_no_limit((char *)header.file_name);
        if (file_name_cstr.len >= 2) {
            /* Chop off './' prefix */
            file_name_cstr.len -= 2;
            file_name_cstr.text += 2;
        }
        struct string file_name = string_cat(arena, prefix, file_name_cstr);

        struct tar_entry *entry = arena_push(arena, struct tar_entry);
        entry->is_dir = is_dir;
        entry->file_name = file_name;
        entry->data = file_data;

        entry->next = archive.head;
        archive.head = entry;
        ++num_files;
    }

    /* Build lookup table */
    archive.lookup = fixed_dict_init(arena, (u64)((f64)num_files * ARCHIVE_LOOKUP_TABLE_CAPACITY_FACTOR));
    for (struct tar_entry *entry = archive.head; entry; entry = entry->next) {
        fixed_dict_set(arena, &archive.lookup, entry->file_name, entry);
    }

    /* Build hierarchy */
    /* NOTE: This is a separate pass because tar entry order is not guaranteed
     * (IE file entries may be encountered before their parent directory entry) */
    for (struct tar_entry *entry = archive.head; entry; entry = entry->next) {
        /* Enter into hierarchy */
        if (!entry->is_dir) {
            /* Find parent entry */
            struct tar_entry *parent_entry = NULL;
            for (struct string parent_dir_name = entry->file_name; parent_dir_name.len > 0; --parent_dir_name.len) {
                if (parent_dir_name.text[parent_dir_name.len - 1] == '/') {
                    parent_entry = fixed_dict_get(&archive.lookup, parent_dir_name);
                    break;
                }
            }
            /* Insert child into parent's list */
            if (parent_entry) {
                entry->next_child = parent_entry->next_child;
                parent_entry->next_child = entry;
            }
        }
    }

    return archive;
}

struct tar_entry *tar_get(const struct tar_archive *archive, struct string name)
{
    return fixed_dict_get(&archive->lookup, name);
}