- Support of Python 3. Compatibility with python2 is preserved for this release.
- Add MSR-3 validating repertoire.
- Create a "Tag" management page.
- Create a function to assign tags/WLE to code points.
- Allow choosing reference id when adding reference in editor.
- Add tags in code point list view.
- Improve display of LGR validation when rebuild is valid.
- Catch exceptions generated when input file is not a UTF-8 encoded file.
- Add required file encoding on label fields in forms.
- Update harmonization tool.
- When adding an out-of-repertoire variant, automatically add it to the repertoire + add the mapping to the code point.
- Improve wording of landing page.
- Improve handling of output for comparison tools for large LGRs.
- Update variant type to blocked for sample French LGR.
- Add notice that import function is permissive.
- Remove length limitation on label validation. Process will be run asynchronously if label can generate too many variants.
- Export testing of variant symmetry and transitivity in AJAX view.
- Add combined form of sequences in HTML output.
- Sort resource/repertoire file lists.
- Remove script parameter for harmonization tools.
- Display rule name(s) when a code point does not comply with a contextual rule.
- Better notification for invalid language tags.
- Better consistency on forms' labels.
- Fix redirection when adding a code point fails.
- Fix link to label in "Validate label" view.
- Remove references from repertoire on deletion.
- Replace number of variants by number of variant mappings in LGR validation statistics.
- Add proper error message when adding a tag to a sequence.
- Add support for missing language exception.
- Fix script retrievals from repertoires.
- Do not display useless error message when editing existing references.
- Fix creation of references from code point/variant edition screen.
- Fix deletion of action.
- Display code points in hexadecimal in HTML output.
- Add identity mapping for newly added variants not present in repertoire.
- Fix header hiding page content.
- Catch error in HTML output on ill-formed LGR's WLE.
New features:
- New interface to view all forms of a label: display the U-label, A-label and code point sequence.
- Display the combined form of sequences.
- New tool to check that multiple LGRs are harmonized.
- New function to populate the missing symmetric and transitive variants.
- New function to add code point from a script.
- When adding a variant from another script, automatically add it to the repertoire if not present. Fixes:
- Improve handling of large LGR.
- Fix variant type that could be be changed from "blocked".
- Fix label validation with ranges.
- Always display failed rule when validating a label.
- Improve display for summary (now renamed to LGR validation).
- Fix tag count in HTML output.
- Fix variant member count in HTML output.
- Use hyperlinks for the table of references in HTML output.
- Reduce number of rows in the variant set table in HTML output.
- Add number of mappings on variant set table in HTML output.
- Add link to save HTML output.
Fixes:
- Remove script/mimetype from merged description when all descriptions use the text/html mimetype.
- Fix cross-script variants tool: output variants composed of code point which are in other scripts than the one(s) defined in the LGR.
Fixes:
- Merge LGR containing the same code point.
- Improve HTML formatting for description in HTML output.
- Improve merge for metadata's description element: if all descriptions are using HTML, then output will be in HTML.
- Include references in WLE table in HTML output.
- Better display of regex for WLE in HTML output.
- Order metadata section per language then per script.
- Fix size of variant set in HTML output.
- Download result links are now ordered with newest on top.
- Emphasize that the "Allocated set labels" file is optional.
- Fix typo on homepage.
- Update links to published RFC.
- Allow import of an LGR part of a set as a standalone LGR.
- Improve cross-script variants tool: generate the variants using the merged LGR, and list element LGRs used by the variants' code points.
- Use redis as default Celery's broker for better stability.
New features:
- Handling of LGR sets: import multiple LGR documents to create a set (merged LGR).
- Update difference and annotation tools to process LGR sets.
- New tool to list cross-script variants for LGR sets.
- HTML output of LGR: render LGR (or LGR set) in an HTML page.
Fixes:
- Update dispositions to RFC: default for variants is "blocked", update default rules.
- Improve design of homepage.
- 500 error when clicking on a code point's variant that is not in LGR.
- Fix Dockerfile: follow redirects for ICU downloads, add SHA-sum checks.
Fixes:
- Fix disposition handling in collision and diff tools
- Improve tools error description provided in e-mail
- Improve tools output
- Manage bidirectional text in tools output
- Improve tools performance
New features:
- Add cancel button on metadata screen
- Add a list of current and imported LGRs into import/new panels
- Add supported Unicode version in about page
- Add an annotation tool
- Add a checkbox to limit rules output in tools
- Save tools output on server and display download link in session
- Add a management command to clean storage
- Some comparison leads to invalid LGR, return a download link for user to fix it
Fixes:
- Fix manual import of RFC3743
- Fix error field position in "Import CP from file" screen
- Improve code point details view
- Various fixes in union and intersection
- Various fixes in variant generation
- Various display improvements
- Limit logging output
- Update RNG file
- Various code factorization
New features:
- Add documentation for core API and web-application global design.
- New tool for computing label collisions with two versions of an LGR. Add asynchronous interface for users to submit their requests and get result by email.
Fixes:
- Update RNG file to last version.
- Enable definition of when/not-when attribute on range.
Fixes:
- Fix processing of look-around when anchor code point was repeated.
- Fix processing of infinite look-behind rules.
- Fix range expansion references.
- Correctly process variant label in context rule.
- Catch regex exceptions.
- Fix WLE validation.
- Handle empty set/pattern in WLE processing.
- Ensure label is in LGR when computing variants when dealing with sequences.
- When/Not-when attributes are mutually exclusive.
- Fix count attributes on char match operator.
- Fix class combinations for intersection and symetric-difference.
- Update RNG file.
- Make the title more clear to be clickable as « home »
- Correctly include Docker configuration.
New features:
- WLE rules editor.
- Tag edition on code points.
- Context rules selection on code points.
- Expand ranges to single code point.
- New LGR comparison application:
- Union of 2 LGRs.
- Intersection of 2 LGRs.
- Diff (text output) of 2 LGRs.
- Save summary to file.
- Delete LGR from working session.
- Add Dockerfile.
- Update namespace: imported LGRs will be converted to new namespace on export.
Fixes:
- Disable external entities when parsing XML (security review report).
- Properly handle IDNA errors in label validation.
- Import variants when manually importing from RFC file.
- Fix label variants generation when context rules are used.