rust/compiler/rustc_serialize/src, branch 1.71.0 https://github.com/rust-lang/rust http://git.dreamy.place/mirrors/rust/atom?h=1.71.0 2023-05-14T22:59:11+00:00 Fix the `FileEncoder` buffer size. 2023-05-14T22:59:11+00:00 Nicholas Nethercote n.nethercote@gmail.com 2023-05-02T03:19:32+00:00 urn:sha1:f2df861c7f35d8fe25ff9a2e9188ad90f8262188 It allows a variable size, but in practice we always use the default of 8192 bytes. This commit fixes it to that size, which makes things slightly faster because the size can be hard-wired in generated code. The commit also: - Rearranges some buffer capacity checks so they're all in the same form (`x > BUFSIZE`). - Removes some buffer capacity assertions and comments about them. With an 8192 byte buffer, we're not in any danger of overflowing a `usize`. Factor out more repeated code in `{write,read}_leb128!`. 2023-05-04T03:52:14+00:00 Nicholas Nethercote n.nethercote@gmail.com 2023-05-04T03:10:22+00:00 urn:sha1:723ca2a33df3774914abc68a8ab79238904ef067 Make them generate the entire function, not just the function body. Rename `file_encoder_write_leb128!`. 2023-05-04T03:51:20+00:00 Nicholas Nethercote n.nethercote@gmail.com 2023-05-04T02:55:01+00:00 urn:sha1:4ac959a3c083dcd3c6536c26513b26ff891abfe6 `MemEncoder` was recently removed, leaving `FileEncoder` as the only encoder. So this prefix is no longer needed, and `write_leb128!` matches the existing `read_leb128!`. Reorder some `MemDecoder` methods. 2023-05-04T03:11:51+00:00 Nicholas Nethercote n.nethercote@gmail.com 2023-05-01T06:06:43+00:00 urn:sha1:58002faca0ac45589c6092030a5459eae9720e3d So they match the order in the `Decoder` trait. Remove a low value comment. 2023-05-04T00:42:42+00:00 Nicholas Nethercote n.nethercote@gmail.com 2023-05-01T06:40:27+00:00 urn:sha1:b71ce293e843f17e27bd2868981dead23691019e Remove `MemEncoder`. 2023-05-02T02:02:32+00:00 Nicholas Nethercote n.nethercote@gmail.com 2023-05-01T08:51:05+00:00 urn:sha1:ebee3f8515c6f5189b69ae56919ab5bba934aabe It's only used in tests. Which is bad, because it means that `FileEncoder` is used in the compiler but isn't used in tests! `tests/opaque.rs` now tests encoding/decoding round-trips via file. Because this is slower than memory, this commit also adjusts the `u16`/`i16` tests so they are more like the `u32`/`i32` tests, i.e. they don't test every possible value. Remove `MemDecoder::read_byte`. 2023-04-28T08:34:55+00:00 Nicholas Nethercote n.nethercote@gmail.com 2023-04-27T23:16:31+00:00 urn:sha1:a676dfa888b3d14abfa5d9b9a1045f8c1bde6793 It's just a synonym for `read_u8`. Add some provided methods to `Encoder`/`Decoder`. 2023-04-28T08:34:54+00:00 Nicholas Nethercote n.nethercote@gmail.com 2023-04-27T23:06:57+00:00 urn:sha1:7a16d25365b5f0aa815948237c46fb1843386d7a The methods for `i8`, `bool`, `char`, `str` are the same for all impls, because they layered on top of other methods. Remove a low-value assertion. 2023-04-28T08:34:52+00:00 Nicholas Nethercote n.nethercote@gmail.com 2023-04-27T23:01:00+00:00 urn:sha1:fa133f5354ac29096d1577d5ba9c1400c2ad3b0f Checking that `read_raw_bytes(len)` changes the position by `len` is a reasonable thing for a test, but isn't much use in just one of the zillion `Decodable` impls. Add a comment explaining the lack of `Decoder::read_enum_variant`. 2023-04-27T23:51:00+00:00 Nicholas Nethercote n.nethercote@gmail.com 2023-04-27T22:53:13+00:00 urn:sha1:37c9e45186e00c197902a7c7349c18383aa0abf7 Because I was wondering about it, and this may save a future person from also wondering.