tegdb 0.5.0

The name TegridyDB (short for TegDB) is inspired by the Tegridy Farm in South Park and tries to correct some of the wrong database implementations, such as null support, implicit conversion support, etc.
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91

use std::fs::OpenOptions;
use std::io::{Seek, SeekFrom, Write};

use tegdb::log::{
    LENGTH_FIELD_BYTES, STORAGE_FORMAT_VERSION, STORAGE_HEADER_SIZE, STORAGE_MAGIC,
    TX_COMMIT_MARKER,
};
use tegdb::tgstream::parser::find_last_commit_offset;

fn write_header(file: &mut std::fs::File, valid_data_end: u64) {
    let mut header = vec![0u8; STORAGE_HEADER_SIZE];
    header[0..STORAGE_MAGIC.len()].copy_from_slice(STORAGE_MAGIC);
    header[6..8].copy_from_slice(&STORAGE_FORMAT_VERSION.to_be_bytes());
    header[8..12].copy_from_slice(&0u32.to_be_bytes());
    // max_key/max_val are not validated by parser, fill with defaults
    header[12..16].copy_from_slice(&(1024u32).to_be_bytes());
    header[16..20].copy_from_slice(&(256 * 1024u32).to_be_bytes());
    header[20] = 1u8; // BE
                      // Record valid_data_end [21..29)
    header[21..29].copy_from_slice(&valid_data_end.to_be_bytes());
    file.write_all(&header).unwrap();
}

fn write_record(file: &mut std::fs::File, key: &[u8], value: &[u8]) {
    let key_len = (key.len() as u32).to_be_bytes();
    let val_len = (value.len() as u32).to_be_bytes();
    file.write_all(&key_len).unwrap();
    file.write_all(&val_len).unwrap();
    file.write_all(key).unwrap();
    file.write_all(value).unwrap();
}

#[test]
fn test_find_last_commit_offset_simple() {
    let dir = tempfile::tempdir().unwrap();
    let path = dir.path().join("parser_test.teg");
    let mut f = OpenOptions::new()
        .create(true)
        .truncate(true)
        .read(true)
        .write(true)
        .open(&path)
        .unwrap();

    // Write placeholder header first
    write_header(&mut f, STORAGE_HEADER_SIZE as u64);

    // Entry 1
    write_record(&mut f, b"k1", b"v1");
    // Commit 1
    write_record(&mut f, TX_COMMIT_MARKER, b"");
    // Entry 2 (after commit)
    write_record(&mut f, b"k2", b"v2");

    // Get current position as valid_data_end
    let valid_data_end = f.stream_position().unwrap();

    // Incomplete trailing record (key_len only) – should be ignored (outside valid_data_end)
    f.write_all(&(4u32.to_be_bytes())).unwrap();

    f.flush().unwrap();

    // Update header with correct valid_data_end
    f.seek(SeekFrom::Start(0)).unwrap();
    write_header(&mut f, valid_data_end);
    f.flush().unwrap();

    let mut rf = OpenOptions::new().read(true).open(&path).unwrap();
    let off = find_last_commit_offset(&mut rf).unwrap();

    // We expect offset to be right after the commit record written
    // The commit record layout: 4 + 4 + key_len + value_len = 8 + 12 + 0 = 20 for key "__TX_COMMIT__"
    // Compute expected position by re-parsing using the same logic: header + k1 + commit
    // Simpler: ensure it is at least after header and larger than first entry
    assert!(off > STORAGE_HEADER_SIZE as u64);

    // Seek to the offset and ensure we can continue reading the next full record (k2)
    rf.seek(SeekFrom::Start(off)).unwrap();
    use std::io::Read;
    let mut len_buf = [0u8; LENGTH_FIELD_BYTES];
    rf.read_exact(&mut len_buf).unwrap();
    let klen = u32::from_be_bytes(len_buf) as usize;
    rf.read_exact(&mut len_buf).unwrap();
    let vlen = u32::from_be_bytes(len_buf) as usize;
    let mut k = vec![0u8; klen];
    rf.read_exact(&mut k).unwrap();
    assert_eq!(&k, b"k2");
    let mut v = vec![0u8; vlen];
    rf.read_exact(&mut v).unwrap();
    assert_eq!(&v, b"v2");
}