v2.0.0-beta.5

[pelletier]

While this release focused on adding back missing encoder and decoder features, it also comes with a slew of bug fixes, thanks to @moorereason's awesome differential fuzzing work! This is the last planned pre-release dedicated to the core library. Check out the updated readme for upgrade instructions.

The next milestone is to port the existing tooling (tomljson, tomllint, binary releases) to go-toml v2.

Thank you, and happy holidays!

What's Changed

What's new

• API: Encoder and Decoder options are chainable by @pelletier in API: Encoder and Decoder options are chainable #670
• Decoder: show struct field in type mismatch errors by @pelletier in Decoder: show struct field in type mismatch errors #684
• Encoder: omitempty flag by @pelletier in Encoder: omitempty flag #692
• Decoder: allow commas in tags by @pelletier in Decoder: allow commas in tags #693
• Decoder: time allows extra precision by @pelletier in Decoder: time allows extra precision #710
• Encode: add comment struct tag by @pelletier in Encode: add comment struct tag #711

Performance

• Decoder: use lookup table to validate ASCII by @pelletier in Decoder: use lookup table to validate ASCII #654
• Decode: code cleanup for struct cache by @pelletier in Decode: code cleanup for struct cache #659

Fixed bugs

• Decode: error on underscore before exponent in floats by @pelletier in Decode: error on underscore before exponent in floats #647
• Decode: add missing checks for LocalTime by @pelletier in Decode: add missing checks for LocalTime #650
• Decoder: flag invalid carriage returns in strings by @pelletier in Decoder: flag invalid carriage returns in strings #652
• Decode: correctly handle carriage return in comments by @pelletier in Decode: correctly handle carriage return in comments #656
• Decode: prevent comments that look like dates to be accepted by @pelletier in Decode: prevent comments that look like dates to be accepted #657
• Decoder: validate inline tables don't redefine keys by @pelletier in Decoder: validate inline tables don't redefine keys #660
• Decoder: validate inline tables in arrays don't redefine keys by @pelletier in Decoder: validate inline tables in arrays don't redefine keys #663
• Decode: accept E as exponent notation by @moorereason in Decode: accept E as exponent notation #665
• Decoder: prevent duplicates of inline tables by @pelletier in Decoder: prevent duplicates of inline tables #667
• Decoder: flag invalid carriage returns in literal strings by @moorereason in Decoder: flag invalid carriage returns in literal strings #673
• Encoder: try to use pointer type TextMarshaler by @pelletier in Encoder: try to use pointer type TextMarshaler #679
• Decoder: validate bounds of day and month in dates by @pelletier in Decoder: validate bounds of day and month in dates #680
• Decoder: fail on unescaped \r not followed by \n by @pelletier in Decoder: fail on unescaped \r not followed by \n #681
• Decoder: fix panic on table array behind a pointer by @pelletier in Decoder: fix panic on table array behind a pointer #682
• Decoder: flag bad \r in literal multiline strings by @pelletier in Decoder: flag bad \r in literal multiline strings #687
• Decoder: check timezones start with +,-,z,Z by @pelletier in Decoder: check timezones start with +,-,z,Z #688
• Decoder: check tz has : between hours and minutes by @pelletier in Decoder: check tz has : between hours and minutes #691
• Decoder: check for invalid characters in timezone by @pelletier in Decoder: check for invalid characters in timezone #695
• Decode: restrict timezone offset values by @moorereason in Decode: restrict timezone offset values #696
• Decode: ensure signed numbers don't start with an underscore by @moorereason in Decode: ensure signed numbers don't start with an underscore #698
• Decode: allow maximum seconds value of 60 by @moorereason in Decode: allow maximum seconds value of 60 #700
• Decode: ensure signed exponents don't start with an underscore by @moorereason in Decode: ensure signed exponents don't start with an underscore #699
• Decoder: prevent modification of inline tables by @pelletier in Decoder: prevent modification of inline tables #702
• Decoder: disallow modification of existing table by @pelletier in Decoder: disallow modification of existing table #704
• Decode: fix maximum time offset values by @moorereason in Decode: fix maximum time offset values #706
• Decode: allow CRLF as whitespace before line ending backslash by @pelletier in Decode: allow CRLF as whitespace before line ending backslash #709
• Decode: fix index panic on integers with just a sign by @moorereason in Decode: fix index panic on integers with just a sign #716
• Decode: check buffer length before parsing simple key by @moorereason in Decode: check buffer length before parsing simple key #717

Documentation

• Encoder: mention "-" to prevent encoding field in doc by @pelletier in Encoder: mention "-" to prevent encoding field in doc #683
• Readme: document more differences with v1 by @pelletier in Readme: document more differences with v1 #712
• Doc: clarify errors returned by Decode by @pelletier in Doc: clarify errors returned by Decode #713

Full Changelog: v2.0.0-beta.4...v2.0.0-beta.5

---

This discussion was created from the release v2.0.0-beta.5.

[moorereason]

I agree. I think I've unearthed the low-hanging fruit as far as fuzzing goes.

FWIW: I've primarily been doing differential fuzzing against toml.dart since it seems to be a solid parser (I've yet to find any bugs in it). The other projects I've tried that have a toml-test decoder (BurntSushi, tomlc99, toml11, and toml-f) all have outstanding bugs that need to be resolved before I can continue fuzzing with them.