apache_avro/
util.rs

1// Licensed to the Apache Software Foundation (ASF) under one
2// or more contributor license agreements.  See the NOTICE file
3// distributed with this work for additional information
4// regarding copyright ownership.  The ASF licenses this file
5// to you under the Apache License, Version 2.0 (the
6// "License"); you may not use this file except in compliance
7// with the License.  You may obtain a copy of the License at
8//
9//   http://www.apache.org/licenses/LICENSE-2.0
10//
11// Unless required by applicable law or agreed to in writing,
12// software distributed under the License is distributed on an
13// "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
14// KIND, either express or implied.  See the License for the
15// specific language governing permissions and limitations
16// under the License.
17
18use crate::{schema::Documentation, AvroResult, Error};
19use serde_json::{Map, Value};
20use std::{
21    io::Read,
22    sync::{
23        atomic::{AtomicBool, AtomicUsize, Ordering},
24        Once,
25    },
26};
27
28/// Maximum number of bytes that can be allocated when decoding
29/// Avro-encoded values. This is a protection against ill-formed
30/// data, whose length field might be interpreted as enormous.
31/// See max_allocation_bytes to change this limit.
32pub const DEFAULT_MAX_ALLOCATION_BYTES: usize = 512 * 1024 * 1024;
33static MAX_ALLOCATION_BYTES: AtomicUsize = AtomicUsize::new(DEFAULT_MAX_ALLOCATION_BYTES);
34static MAX_ALLOCATION_BYTES_ONCE: Once = Once::new();
35
36/// Whether to set serialization & deserialization traits
37/// as `human_readable` or not.
38/// See [set_serde_human_readable] to change this value.
39// crate-visible for testing
40pub(crate) static SERDE_HUMAN_READABLE: AtomicBool = AtomicBool::new(true);
41static SERDE_HUMAN_READABLE_ONCE: Once = Once::new();
42
43pub trait MapHelper {
44    fn string(&self, key: &str) -> Option<String>;
45
46    fn name(&self) -> Option<String> {
47        self.string("name")
48    }
49
50    fn doc(&self) -> Documentation {
51        self.string("doc")
52    }
53
54    fn aliases(&self) -> Option<Vec<String>>;
55}
56
57impl MapHelper for Map<String, Value> {
58    fn string(&self, key: &str) -> Option<String> {
59        self.get(key)
60            .and_then(|v| v.as_str())
61            .map(|v| v.to_string())
62    }
63
64    fn aliases(&self) -> Option<Vec<String>> {
65        // FIXME no warning when aliases aren't a json array of json strings
66        self.get("aliases")
67            .and_then(|aliases| aliases.as_array())
68            .and_then(|aliases| {
69                aliases
70                    .iter()
71                    .map(|alias| alias.as_str())
72                    .map(|alias| alias.map(|a| a.to_string()))
73                    .collect::<Option<_>>()
74            })
75    }
76}
77
78pub fn read_long<R: Read>(reader: &mut R) -> AvroResult<i64> {
79    zag_i64(reader)
80}
81
82pub fn zig_i32(n: i32, buffer: &mut Vec<u8>) {
83    zig_i64(n as i64, buffer)
84}
85
86pub fn zig_i64(n: i64, buffer: &mut Vec<u8>) {
87    encode_variable(((n << 1) ^ (n >> 63)) as u64, buffer)
88}
89
90pub fn zag_i32<R: Read>(reader: &mut R) -> AvroResult<i32> {
91    let i = zag_i64(reader)?;
92    i32::try_from(i).map_err(|e| Error::ZagI32(e, i))
93}
94
95pub fn zag_i64<R: Read>(reader: &mut R) -> AvroResult<i64> {
96    let z = decode_variable(reader)?;
97    Ok(if z & 0x1 == 0 {
98        (z >> 1) as i64
99    } else {
100        !(z >> 1) as i64
101    })
102}
103
104fn encode_variable(mut z: u64, buffer: &mut Vec<u8>) {
105    loop {
106        if z <= 0x7F {
107            buffer.push((z & 0x7F) as u8);
108            break;
109        } else {
110            buffer.push((0x80 | (z & 0x7F)) as u8);
111            z >>= 7;
112        }
113    }
114}
115
116fn decode_variable<R: Read>(reader: &mut R) -> AvroResult<u64> {
117    let mut i = 0u64;
118    let mut buf = [0u8; 1];
119
120    let mut j = 0;
121    loop {
122        if j > 9 {
123            // if j * 7 > 64
124            return Err(Error::IntegerOverflow);
125        }
126        reader
127            .read_exact(&mut buf[..])
128            .map_err(Error::ReadVariableIntegerBytes)?;
129        i |= (u64::from(buf[0] & 0x7F)) << (j * 7);
130        if (buf[0] >> 7) == 0 {
131            break;
132        } else {
133            j += 1;
134        }
135    }
136
137    Ok(i)
138}
139
140/// Set a new maximum number of bytes that can be allocated when decoding data.
141/// Once called, the limit cannot be changed.
142///
143/// **NOTE** This function must be called before decoding **any** data. The
144/// library leverages [`std::sync::Once`](https://doc.rust-lang.org/std/sync/struct.Once.html)
145/// to set the limit either when calling this method, or when decoding for
146/// the first time.
147pub fn max_allocation_bytes(num_bytes: usize) -> usize {
148    MAX_ALLOCATION_BYTES_ONCE.call_once(|| {
149        MAX_ALLOCATION_BYTES.store(num_bytes, Ordering::Release);
150    });
151    MAX_ALLOCATION_BYTES.load(Ordering::Acquire)
152}
153
154pub fn safe_len(len: usize) -> AvroResult<usize> {
155    let max_bytes = max_allocation_bytes(DEFAULT_MAX_ALLOCATION_BYTES);
156
157    if len <= max_bytes {
158        Ok(len)
159    } else {
160        Err(Error::MemoryAllocation {
161            desired: len,
162            maximum: max_bytes,
163        })
164    }
165}
166
167/// Set whether serializing/deserializing is marked as human readable in serde traits.
168/// This will adjust the return value of `is_human_readable()` for both.
169/// Once called, the value cannot be changed.
170///
171/// **NOTE** This function must be called before serializing/deserializing **any** data. The
172/// library leverages [`std::sync::Once`](https://doc.rust-lang.org/std/sync/struct.Once.html)
173/// to set the limit either when calling this method, or when decoding for
174/// the first time.
175pub fn set_serde_human_readable(human_readable: bool) {
176    SERDE_HUMAN_READABLE_ONCE.call_once(|| {
177        SERDE_HUMAN_READABLE.store(human_readable, Ordering::Release);
178    });
179}
180
181pub(crate) fn is_human_readable() -> bool {
182    SERDE_HUMAN_READABLE.load(Ordering::Acquire)
183}
184
185#[cfg(test)]
186mod tests {
187    use super::*;
188    use apache_avro_test_helper::TestResult;
189    use pretty_assertions::assert_eq;
190
191    #[test]
192    fn test_zigzag() {
193        let mut a = Vec::new();
194        let mut b = Vec::new();
195        zig_i32(42i32, &mut a);
196        zig_i64(42i64, &mut b);
197        assert_eq!(a, b);
198    }
199
200    #[test]
201    fn test_zig_i64() {
202        let mut s = Vec::new();
203
204        zig_i64(0, &mut s);
205        assert_eq!(s, [0]);
206
207        s.clear();
208        zig_i64(-1, &mut s);
209        assert_eq!(s, [1]);
210
211        s.clear();
212        zig_i64(1, &mut s);
213        assert_eq!(s, [2]);
214
215        s.clear();
216        zig_i64(-64, &mut s);
217        assert_eq!(s, [127]);
218
219        s.clear();
220        zig_i64(64, &mut s);
221        assert_eq!(s, [128, 1]);
222
223        s.clear();
224        zig_i64(i32::MAX as i64, &mut s);
225        assert_eq!(s, [254, 255, 255, 255, 15]);
226
227        s.clear();
228        zig_i64(i32::MAX as i64 + 1, &mut s);
229        assert_eq!(s, [128, 128, 128, 128, 16]);
230
231        s.clear();
232        zig_i64(i32::MIN as i64, &mut s);
233        assert_eq!(s, [255, 255, 255, 255, 15]);
234
235        s.clear();
236        zig_i64(i32::MIN as i64 - 1, &mut s);
237        assert_eq!(s, [129, 128, 128, 128, 16]);
238
239        s.clear();
240        zig_i64(i64::MAX, &mut s);
241        assert_eq!(s, [254, 255, 255, 255, 255, 255, 255, 255, 255, 1]);
242
243        s.clear();
244        zig_i64(i64::MIN, &mut s);
245        assert_eq!(s, [255, 255, 255, 255, 255, 255, 255, 255, 255, 1]);
246    }
247
248    #[test]
249    fn test_zig_i32() {
250        let mut s = Vec::new();
251        zig_i32(i32::MAX / 2, &mut s);
252        assert_eq!(s, [254, 255, 255, 255, 7]);
253
254        s.clear();
255        zig_i32(i32::MIN / 2, &mut s);
256        assert_eq!(s, [255, 255, 255, 255, 7]);
257
258        s.clear();
259        zig_i32(-(i32::MIN / 2), &mut s);
260        assert_eq!(s, [128, 128, 128, 128, 8]);
261
262        s.clear();
263        zig_i32(i32::MIN / 2 - 1, &mut s);
264        assert_eq!(s, [129, 128, 128, 128, 8]);
265
266        s.clear();
267        zig_i32(i32::MAX, &mut s);
268        assert_eq!(s, [254, 255, 255, 255, 15]);
269
270        s.clear();
271        zig_i32(i32::MIN, &mut s);
272        assert_eq!(s, [255, 255, 255, 255, 15]);
273    }
274
275    #[test]
276    fn test_overflow() {
277        let causes_left_shift_overflow: &[u8] = &[0xe1, 0xe1, 0xe1, 0xe1, 0xe1];
278        assert!(decode_variable(&mut &*causes_left_shift_overflow).is_err());
279    }
280
281    #[test]
282    fn test_safe_len() -> TestResult {
283        assert_eq!(42usize, safe_len(42usize)?);
284        assert!(safe_len(1024 * 1024 * 1024).is_err());
285
286        Ok(())
287    }
288}
apache_avro/util.rs

apache_avro/
util.rs
