Replace yaml_parser_initialize with safe alternative

src/api.rs

@@ -4,11 +4,12 @@ use alloc::vec::Vec;
 use crate::yaml::{YamlEventData, YamlNodeData};
 use crate::{
     libc, yaml_break_t, yaml_document_t, yaml_emitter_t, yaml_encoding_t, yaml_event_t,
-    yaml_mapping_style_t, yaml_mark_t, yaml_node_pair_t, yaml_node_t, yaml_parser_t,
-    yaml_scalar_style_t, yaml_sequence_style_t, yaml_tag_directive_t, yaml_token_t,
+    yaml_mapping_style_t, yaml_mark_t, yaml_node_pair_t, yaml_node_t, yaml_parser_state_t,
+    yaml_parser_t, yaml_scalar_style_t, yaml_sequence_style_t, yaml_tag_directive_t, yaml_token_t,
     yaml_version_directive_t, YAML_ANY_ENCODING, YAML_UTF8_ENCODING,
 };
 use core::ptr;
+use std::collections::VecDeque;
 
 pub(crate) const INPUT_RAW_BUFFER_SIZE: usize = 16384;
 pub(crate) const INPUT_BUFFER_SIZE: usize = INPUT_RAW_BUFFER_SIZE;
@@ -18,19 +19,33 @@ pub(crate) const OUTPUT_BUFFER_SIZE: usize = 16384;
 ///
 /// This function creates a new parser object. An application is responsible
 /// for destroying the object using the yaml_parser_delete() function.
-pub unsafe fn yaml_parser_initialize(parser: *mut yaml_parser_t) -> Result<(), ()> {
-    __assert!(!parser.is_null());
-    core::ptr::write(parser, yaml_parser_t::default());
-    let parser = &mut *parser;
-    parser.raw_buffer.reserve(INPUT_RAW_BUFFER_SIZE);
-    parser.buffer.reserve(INPUT_BUFFER_SIZE);
-    parser.tokens.reserve(16);
-    parser.indents.reserve(16);
-    parser.simple_keys.reserve(16);
-    parser.states.reserve(16);
-    parser.marks.reserve(16);
-    parser.tag_directives.reserve(16);
-    Ok(())
+pub fn yaml_parser_new<'r>() -> yaml_parser_t<'r> {
+    yaml_parser_t {
+        read_handler: None,
+        input: Default::default(),
+        eof: false,
+        buffer: VecDeque::with_capacity(INPUT_BUFFER_SIZE),
+        unread: 0,
+        raw_buffer: VecDeque::with_capacity(INPUT_RAW_BUFFER_SIZE),
+        encoding: YAML_ANY_ENCODING,
+        offset: 0,
+        mark: yaml_mark_t::default(),
+        stream_start_produced: false,
+        stream_end_produced: false,
+        flow_level: 0,
+        tokens: VecDeque::with_capacity(16),
+        tokens_parsed: 0,
+        token_available: false,
+        indents: Vec::with_capacity(16),
+        indent: 0,
+        simple_key_allowed: false,
+        simple_keys: Vec::with_capacity(16),
+        states: Vec::with_capacity(16),
+        state: yaml_parser_state_t::default(),
+        marks: Vec::with_capacity(16),
+        tag_directives: Vec::with_capacity(16),
+        aliases: Vec::new(),
+    }
 }
 
 /// Destroy a parser.

src/bin/run-parser-test-suite.rs

@@ -12,27 +12,22 @@
 )]
 
 use libyaml_safer::{
-    yaml_parser_delete, yaml_parser_initialize, yaml_parser_parse, yaml_parser_set_input,
-    yaml_parser_t, YamlEventData, YAML_DOUBLE_QUOTED_SCALAR_STYLE, YAML_FOLDED_SCALAR_STYLE,
-    YAML_LITERAL_SCALAR_STYLE, YAML_PLAIN_SCALAR_STYLE, YAML_SINGLE_QUOTED_SCALAR_STYLE,
+    yaml_parser_delete, yaml_parser_new, yaml_parser_parse, yaml_parser_set_input, YamlEventData,
+    YAML_DOUBLE_QUOTED_SCALAR_STYLE, YAML_FOLDED_SCALAR_STYLE, YAML_LITERAL_SCALAR_STYLE,
+    YAML_PLAIN_SCALAR_STYLE, YAML_SINGLE_QUOTED_SCALAR_STYLE,
 };
 use std::env;
 use std::error::Error;
 use std::fs::File;
 use std::io::{self, Read, Write};
-use std::mem::MaybeUninit;
 use std::process::{self, ExitCode};
 use std::slice;
 
-pub(crate) unsafe fn unsafe_main(
+pub(crate) fn test_main(
     stdin: &mut dyn Read,
     stdout: &mut dyn Write,
 ) -> Result<(), Box<dyn Error>> {
-    let mut parser = MaybeUninit::<yaml_parser_t>::uninit();
-    if yaml_parser_initialize(parser.as_mut_ptr()).is_err() {
-        return Err("Could not initialize the parser object".into());
-    }
-    let mut parser = parser.assume_init();
+    let mut parser = yaml_parser_new();
 
     yaml_parser_set_input(&mut parser, stdin);
 
@@ -137,7 +132,7 @@ pub(crate) unsafe fn unsafe_main(
     Ok(())
 }
 
-unsafe fn print_escaped(stdout: &mut dyn Write, s: &str) {
+fn print_escaped(stdout: &mut dyn Write, s: &str) {
     for ch in s.bytes() {
         let repr = match &ch {
             b'\\' => b"\\\\",
@@ -161,7 +156,7 @@ fn main() -> ExitCode {
     for arg in args {
         let mut stdin = File::open(arg).unwrap();
         let mut stdout = io::stdout();
-        let result = unsafe { unsafe_main(&mut stdin, &mut stdout) };
+        let result = test_main(&mut stdin, &mut stdout);
         if let Err(err) = result {
             let _ = writeln!(io::stderr(), "{}", err);
             return ExitCode::FAILURE;

src/lib.rs

@@ -116,8 +116,8 @@ pub use crate::api::{
     yaml_emitter_set_indent, yaml_emitter_set_output, yaml_emitter_set_output_string,
     yaml_emitter_set_unicode, yaml_emitter_set_width, yaml_event_delete,
     yaml_mapping_end_event_initialize, yaml_mapping_start_event_initialize, yaml_parser_delete,
-    yaml_parser_initialize, yaml_parser_set_encoding, yaml_parser_set_input,
-    yaml_parser_set_input_string, yaml_scalar_event_initialize, yaml_sequence_end_event_initialize,
+    yaml_parser_new, yaml_parser_set_encoding, yaml_parser_set_input, yaml_parser_set_input_string,
+    yaml_scalar_event_initialize, yaml_sequence_end_event_initialize,
     yaml_sequence_start_event_initialize, yaml_stream_end_event_initialize,
     yaml_stream_start_event_initialize, yaml_token_delete,
 };
@@ -150,9 +150,7 @@ mod tests {
     #[test]
     fn sanity() {
         unsafe {
-            let mut parser = core::mem::MaybeUninit::uninit();
-            yaml_parser_initialize(parser.as_mut_ptr()).unwrap();
-            let mut parser = parser.assume_init();
+            let mut parser = yaml_parser_new();
             // const SANITY_INPUT: &'static str =
             //     "Mark McGwire:\n  hr: 65\n  avg: 0.278\nSammy Sosa:\n  hr: 63\n  avg: 0.288\n";
             const SANITY_INPUT: &'static str = r#"
@@ -199,7 +197,7 @@ foo: bar
 
     #[test]
     fn test_case() {
-        let mut parser = parser_new();
+        let mut parser = yaml_parser_new();
         let mut input = TEST_CASE_QF4Y.as_bytes();
         yaml_parser_set_input_string(&mut parser, &mut input);
         let mut doc = yaml_document_t::default();
@@ -298,12 +296,4 @@ foo: bar
             emitter.assume_init()
         }
     }
-
-    fn parser_new<'w>() -> yaml_parser_t<'w> {
-        unsafe {
-            let mut emitter = core::mem::MaybeUninit::uninit();
-            yaml_parser_initialize(emitter.as_mut_ptr()).unwrap();
-            emitter.assume_init()
-        }
-    }
 }

src/yaml.rs

@@ -2,7 +2,7 @@ use alloc::collections::VecDeque;
 use alloc::string::String;
 use alloc::vec::Vec;
 
-use crate::libc;
+use crate::{api::yaml_parser_new, libc};
 use core::ptr;
 
 pub use self::yaml_encoding_t::*;
@@ -688,32 +688,7 @@ pub struct yaml_parser_t<'r> {
 
 impl<'r> Default for yaml_parser_t<'r> {
     fn default() -> Self {
-        Self {
-            read_handler: None,
-            input: Default::default(),
-            eof: Default::default(),
-            buffer: Default::default(),
-            unread: Default::default(),
-            raw_buffer: Default::default(),
-            encoding: Default::default(),
-            offset: Default::default(),
-            mark: Default::default(),
-            stream_start_produced: Default::default(),
-            stream_end_produced: Default::default(),
-            flow_level: Default::default(),
-            tokens: Default::default(),
-            tokens_parsed: Default::default(),
-            token_available: Default::default(),
-            indents: Default::default(),
-            indent: Default::default(),
-            simple_key_allowed: Default::default(),
-            simple_keys: Default::default(),
-            states: Default::default(),
-            state: Default::default(),
-            marks: Default::default(),
-            tag_directives: Default::default(),
-            aliases: Default::default(),
-        }
+        yaml_parser_new()
     }
 }