Askalono Versions Save

Changed

• Updated SPDX dataset

Fixed

• Fix silent failures in file I/O when using the CLI

Changed

• Updated SPDX dataset
• Bumped zstd up to 0.11; thanks @mellowagain

Changed

• Updated SPDX dataset
• Tons of code cleanup; thanks @hdhoang
• zstd update to 0.8; thanks @decathorpe
• Replace failure with anyhow; thanks @hdhoang

Added

• Re-introduced gzip cache compression as an optional feature. The default is still zstd.
• Store gained get_original to fetch the canonical license text, instead of needing to access the text via a match.

Changed

• Updated SPDX dataset

Changed

• Performance improvements in text pre-processing (#48, thanks @AnthonyMikh)
• Updated SPDX dataset

Note: If you're using the binary downloaded from GitHub, there's effectively no change between this version and 0.4.0.

Fixed

• Removed some extraneous files from the core askalono packaged crate.
• Include LICENSE, NOTICE, and README.md in askalono-cli crate.

Added

• askalono will attempt to ignore license "title lines" (e.g. "The MIT License") that occasionally appear on licenses.

Changed

• (Breaking) Switch to Rust 2018 edition.
• (Breaking) Drop the previously-deprecated aliases field. Aliases can instead be queried with the aliases function on a Store.
• (Breaking) analyze (on Store) no longer returns a Result. There was no reasonable case where it could fail.
• (Breaking) with_view, white_out, optimize_bounds on TextData also no longer return Result, as they never had an expected failure path outside of programming errors. A panic may occur if an out-of-bounds view is used -- this is intentional. See commit 8d11161c.
• Stores are now compressed with zstd instead of gzip, which provides better compression and performance particularly for the dataset used by askalono.
• Fresh SPDX dataset.
• URLs in licenses will be "black-boxed" in case modified/re-hosted.

Fixed

• The lcs_removal preprocessor has been fixed to be less aggressive on certain repeated statements (#42).
• Fixed CLI help text strings (#34)

NOTE: askalono.linux-static in this release contains code that is slightly newer than 0.3.0. Specifically, it's from commit 45276591150295cda177f0935d1bda6802276edf -- consider it a preview release for some upcoming features. In general, you should use the regular Linux build, but in some cases that's not possible.

Added

• Scanning strategies have been added in ScanStrategy. These include some common high-level logic that can be used for diving deeper into askalono's results.
• Store gained methods to add your own licenses to its dataset at runtime. These are add_license and add_variant.
• Store lets you retrieve and set the list of aliases for a given license.
• Store can also tell you which licenses are present with licenses.
• TextData's with_view is now public. It can be used to change the region of lines in a text you're interested in without needing to wholly re-compute data for the underlying string.
• TextData also learned white_out, which is the inverse of with_view. It erases the active view's lines. This can be useful when a match has been found and you'd like to "hide" it to perform further analysis.
• The command-line application learned how to output JSON for better machine parsing.
• EXPERIMENTAL: askalono is WebAssembly-capable. Loading caches is not yet supported due to missing upstream dependencies. This likely to be available in the near future.

Changed

• optimize_bounds now returns a Result. Previously it would have paniced if self was missing text.

Deprecated

• The aliases field has been deprecated and will be removed in a future release. Prefer looking up aliases in the store when needed.

Fixed

• optimize_bounds now respects existing line views/windows on TextData structs.

Added

• Full documentation for the public API.
• The CLI can now read a list of filenames from stdin (batch mode).
• The CLI is capable of crawling a directory tree, looking for license-like files and identifying them.
• The library and CLI can now optimize a matched license to tell you where in your source the match was identified.

Changed

• Revised API. API changes are likely to be much more rare from here on out.
• Many performance improvements. analyze is now able to run in 3-5 ms on commodity hardware!
• Text storage and normalization improvements.
• Duplicate SPDX entries are now stored as aliases.
• SPDX definitions have been updated.

Removed

• The "diff" option is now only available if compiled with the "diagnostics" feature (off by default). This was intended for debugging and had no practical use in the binary.

Fixed

• Resolved a potential panic for short/empty license files (a divide-by-zero was involved)