Skip to content

Latest commit

 

History

History
87 lines (71 loc) · 3.26 KB

CHANGELOG.md

File metadata and controls

87 lines (71 loc) · 3.26 KB
Raw

Changelog

HEAD

  • Bump packaged FIT SDK version to 21.141.00 (lingepumpe)

v0.7.0

  • Bump packaged FIT SDK version to 21.133.00 (robinkrahl)
  • Remove dead code files in generate-fit-profile (robinkrahl)
  • Rework profile generation code to leverage the quote crate

v0.6.1

  • Bump packaged FIT SDK version to 21.115.00
  • Fix chrono deprecations
  • Update dependencies

v0.6.0

  • Restructured project to drop the calamine dependency for end users, profile generation is now done in a separate private crate
  • Update calamine dependency (evensolberg)
  • Refactor profile decoding module to be functional and not create large "info" structures
  • Implement new decoding options to cover additional use cases
  • Improve component expansion to use a nom-like call chain to avoid loops and clones
  • Bump packaged FIT SDK version to 21.105.00

v0.5.1

  • Bump packaged FIT SDK version to 21.89.00

v0.5.0

  • Use Arc instead of Rc to make FitObjects fully thread-safe (danielalvsaaker)
  • Bump packaged FIT SDK version to 21.78.00

v0.4.3

  • Upgrade nom parser to version 7.0
  • Bump packaged FIT SDK version to 21.67.00
  • Clippy/Typo cleanup (by danielalvsaaker, xehpuk)
  • Add doc comments to field types and messages (by xehpuk)
  • Allow CRC validation to be skipped.
  • Properly handle FIT files where the header CRC is zero (credit ddboline)
  • Check that field size and base type are an even multiple of each other

v0.4.2

  • Bump packaged FIT SDK version to 21.54.01

v0.4.1

  • Bump packaged FIT SDK version to 21.53.00

v0.4.0

  • Improve parser error handling so that the user gets the byte position where parsing failed instead of the number of bytes remaining
  • Expand public API to expose more low level deserialization objects.
  • Use an Rc<...> to hold definition messages to avoid unnecessary copying.
  • Provide an interface that allows a continuous stream of data in addition to batch processing (see examples/streaming.rs).
  • Implement CRC validation of both the header and data sections.
  • Bump packaged FIT SDK version to 21.47.00

v0.3.0

  • Refactor the internal de-serialization and decoding logic.
  • Return better errors when data buffer is incomplete
  • Add the FIT Profile version to the profile module as a constant
  • Add a script to regenerate the profile module from a new SDK
  • Bump packaged FIT SDK version to 21.40.00

v0.2.0

  • Improve the API to allow manipulating the data with less cloning
  • Expose the actual MesgNum enum value to the user in the kind field of each FitDataRecord instead of its name.
  • Upated FitDataRecord to store fields as a Vec, sorted by definition number. Each field contains the name, number, value and units (if defined).
  • Rewrote parser logic to more closely resemble how serde deserializers are typically implemented.
  • Added formal error handling so the code should no longer panic anywhere
  • Removed the FitFile struct, instead the parser returns a vector of FitDataRecord objects.
  • Moved the fit_dump executable to examples and renamed it fit_to_json, this was done to remove the library's dependency on serde_json and structopt.
  • Reimplement the parse top level function as from_reader using the new return type.

v0.1.1

Fixed a parsing error when handling string data fields that caused the parser to panic.

v0.1.0

Initial release on crates.io.