Struct JitDumpReader

pub struct JitDumpReader<R: Read> { /* private fields */ }

Parses a jitdump file and allows iterating over records.

This reader works with complete jitdump files as well as with partial files which are still being written to. This makes it useful in live-profiling settings.

The records refer to memory owned by the reader, to minimize copies.

Implementations

impl<R: Read> JitDumpReader<R>

pub fn new(reader: R) -> Result<Self, JitDumpError>

Create a new JitDumpReader. JitDumpReader does its own buffering so there is no need to wrap a File into a BufReader.

pub fn new_with_buffer_size( reader: R, buffer_size: usize, ) -> Result<Self, JitDumpError>

Create a new JitDumpReader, with a manually-specified buffer chunk size.

pub fn header(&self) -> &JitDumpHeader

The file header.

pub fn endian(&self) -> Endianness

The file endian.

pub fn next_record_header( &mut self, ) -> Result<Option<JitDumpRecordHeader>, Error>

Returns the header of the next record.

pub fn next_record_timestamp(&mut self) -> Result<Option<u64>, Error>

Returns the timestamp of the next record.

When operating on partial files, None means that not enough bytes for the header of the next record are available. Some means that we have enough bytes for the header but we may not have enough bytes to get the entire record.

If next_record_timestamp returns Ok(Some(...)), the next call to next_record() can still return None!

pub fn next_record_type(&mut self) -> Result<Option<JitDumpRecordType>, Error>

Returns the record type of the next record.

pub fn next_record_offset(&self) -> u64

Returns the file offset at which the next record (specifically its record header) starts.

pub fn next_record(&mut self) -> Result<Option<JitDumpRawRecord<'_>>, Error>

Returns the next record.

When operating on partial files, this will return Ok(None) if the entire record is not available yet. Future calls to next_record may return Ok(Some) if the data has become available in the meantime, because they will call read on R again.

impl<R: Read + Seek> JitDumpReader<R>

pub fn skip_next_record(&mut self) -> Result<bool, Error>

Skip the upcoming record. If this returns true, the record has been skipped. If false is returned, it means the file could not be seeked far enough to skip the entire record (for example because this is a partial file which has not been fully written), and the next record remains unchanged from before the call to skip_next_record.

You may want to call this if you’ve called next_record_type and have determined that you’re not interested in the upcoming record. It saves having to read the full record into a contiguous slice of memory.

Trait Implementations

impl<R: Clone + Read> Clone for JitDumpReader<R>

fn clone(&self) -> JitDumpReader<R>

Returns a duplicate of the value.

const fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl<R: Debug + Read> Debug for JitDumpReader<R>

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.