Noteworthy changes in this release

This release brings quite a few technical improvements. We've enhanced support for memory management in the API, added support for math characters in match rules, expanded the Python bindings to include metadata API support, and made it possible to set the log level via the LOUIS_LOGLEVEL environment variable.

There are also substantial updates to various language tables, including improvements to Portuguese 8-dot braille, Hebrew back-translation, Norwegian 8-dot support, UEB forward translation, and Hungarian braille. We've added support for Rokuten Kanji (Japanese Kanji Braille). Russian, Ukrainian, and Belarusian literary braille tables have received updates to their space handling rules, and the Danish tables have seen fixes related to typos and hyphenation.

I'd like to thank everybody for helping to bring liblouis forward.

For a detailed list of all the changes refer to the list of closed issues.

New features

• match rules now also support testing for math characters. Thanks to James Bowden for the idea and Bert Frees for the implementation.
• Add Support for "Rokuten Kanji", the other Japanese system for Kanji Braille.
• The API now contains functions to free memory that was returned by API calls. See the documentation for details. Thanks to Leonard de Ruijter.
• The Python bindings now support most of the metadata API, thanks to Leonard de Ruijter.

Bug fixes

• ucBrl in the mode parameter now correctly works also for back-translation. Thanks to Bert Frees.
• Fix a segfault when a StringBuffer is used before it is initialized. Thanks to Anthony Tibbs.
• Fix a bug in the handling of nocont in conjunction with seqdelimiter. Thanks to Bert Frees.

Braille table improvements

• Improvements to the Portuguese 8-dot Braille table: Correct angle quotation mark characters used in Portugal, add missing symbols, such as Greek alphabet and double quotation mark characters used in Brazil. Also added symbols that aren't defined in the Portuguese standard but are commonly found in texts and web pages, such as bullet points, some common mathematical symbols, etc. Thanks to Tiago M. Casal.
• Fix the back-translation of the Greek Sigma and other small improvements to the Hebrew table. Thanks to Leonard de Ruijter.
• Improvements to Norwegian 8dot thanks to Lars Bjørndal.
  • Add more characters and math symbols.
  • Fix a back-translation problem when two dot 6 are followed by a word.
  • Fix a problem where the letter "i" is followed by number.
• Improvements to UEB forward translation, in particular fixes for rarer words and some proper names. Thanks to James Bowden.
• Hungarian improvements thanks to Attila Hammer.
• Fix typos and hyphenations in the Danish tables thanks to Bue Vester-Andersen.
• Improvements to the back translation for the Russian literary Braille table: Now, if a space is intentionally input after a comma or semicolon, a double space will not occur. If no space was input, a space will still be automatically added, as before. This change also applies to Ukrainian and Belarusian literary Braille tables. Thanks to Andrey Yakuboy.

Other changes

• Hans Anton Ålien proof read the documentation and fixed quite a few typos.
• The log level of the library can now also be set via the LOUIS_LOGLEVEL environment variable.

Deprecation notice

• None

Backwards incompatible changes

Invisible changes

New, renamed or removed tables

New

• ja-rokutenkanji.utb

Renamed

None

Removed

None

Noteworthy changes in this release

This release brings significant improvements to various language tables, particularly for Classical/Biblical Hebrew where three main standards are now supported. We have a new table for Thai Grade 2 contracted braille, and substantial updates to Lithuanian braille tables. There are also improvements to Hungarian, Dutch, Norwegian, and Turkish tables. On the technical side, there's new support for building macOS universal binaries and the pre-built html and plain text documentation is no longer packaged with the released tarball.

I'd like to thank everybody for helping to bring liblouis forward.

For a detailed list of all the changes refer to the list of closed issues.

New features

• A new table for Thai Braille Grade 2, a contracted Braille system, has been added thanks to Chatupol Nootatong.

Braille table improvements

• Hungarian improvements thanks to Attila Hammer.
• Fix a bug in the Dutch table related to apostrophes after a number which are not always a minute sign. Thanks to Bert Frees.
• Fix the metadata of the Russian 6-dot computer braille table thanks to Bert Frees and Andrey Yakuboy.
• Minor improvements to the Norwegian Braille: Some small fixes regarding 6 dot, added Cyrillic alphabet and added certain new characters to 8 dot, including bullet points used by JAWS when using Liblouis for 8 dot. Thanks to Lars Bjørndal and Bert Frees.
• Major changes to the Lithuanian braille tables thanks to Rimas Kudelis
  • Renamed lt.ctb to lt-8dot.utb
  • Both tables have been reviewed and modified.
    • The 6-dot table is now closer to the national standard issued in 2019.
    • The 8-dot table is now intended to align with the 6-dot table instead of the 8-dot standard from 2011.
    • Some whitespaces in the table files have been eliminated to make them easier to read using braille terminals.
• Fixes for back-translation issues involving subscript numerals, diacritic marks, and capital letters turning into Greek letters. Makes no changes to the UEB-style cuneiform transliteration table. Thanks to Eric Harvey.
• Major Expansion and Improvements to Classical/Biblical Hebrew thanks to Eric Harvey.
  • Two new translation tables have been added so that the three main standards for Classical Hebrew are now available:
    1. The original International Hebrew Braille Code (hbo.utb)
    2. The system for full cantillation developed by Ray McAllister, Matthew Yeater, and Sarah Blake LaRose (hbo-cantillated.utb)
    3. The compact standard without dagesh or shewa represented in the Katz transcription manual (hbo-slim.utb)
  • hbo.utb is now the table for the International Hebrew Braille Code. Before it was used for the IHBC-McCallister system (now hbo-cantillated.utb). The International Hebrew Braille Code is the most-established and widely used standard for Classical Hebrew, so it seems right to make it the basic table.
  • Most of the Hebrew-specific rules are contained in .uti files, and the .utb files only add rules for uncontracted English support. Since Classical Hebrew readers usually have another first language, this allows non-English speakers to create custom translation tables with Hebrew and their first language, or for easier inclusion in multilingual tables like ancient-languages-us.utb.
  • In addition to the new tables, there are improvements to the hbo-cantillated.utb table for consistency and to make sure that braille output matches that produced by Duxbury as closely as possible:
    • Ultra long vowels now contract consistently.
    • Cantillation marks no longer interrupt the contractions for ultra-long vowels (hiriq-yod and tsere-yod).
    • Holam now always renders properly as 135 and holam waw as 246.
• Improvements to Turkish grade 1 and Turkish computer braille, thanks to Oğuz Uğur and Ozancan Karataş

Other changes

• There is a new configure option to build macOS universal binaries thanks to Bert Frees.
• Fix incorrect (non-https) link in the documentation thanks to jinwoolee.
• The pre-built html and plain text documentation is no longer packaged with the released tarball. All formats of the documentation can still be built from the source and from the released tarball as explained in the README. Thanks to Christian Egli.

Deprecation notice

• None

Backwards incompatible changes

• None

Invisible changes

• None

New, renamed or removed tables

New

• hbo-cantillated.utb (formerly hbo.utb - hbo.utb was changed to follow the IHBC)
• hbo-slim.utb
• th-g2.ctb

Renamed

• ru-compbrl.ctb -> ru-comp6.utb
• ru.ctb -> ru-comp8.utb
• lt.ctb -> lt-8dot.utb

Removed

None

Noteworthy changes in this release

This release combines braille improvements with security enhancements. We have new tables for Coptic braille, Cuneiform transliteration and for Portuguese 6-dot computer braille. There are improvements to Dutch, Hungarian, the International Phonetic Alphabet, Norwegian and UEB. On the security side a lot of people have reported and fixed buffer overflows and other memory issues.

I'd like to thank everybody for helping to bring liblouis forward.

For a detailed list of all the changes refer to the list of closed issues.

New features

• None

Bug fixes

• A large number of buffer overflow problems have been solved:
  • Fix free error in parseQuery and analyzeTable thanks to Liu Yuwei
  • Fix an overflow in passDoAction and doPassSearch thanks to Liu Yuwei
  • Fixed a segfault in finalizeCharacter thanks to Christian Egli and cx104906.
  • Fix a heap-buffer-overflow error in compileRule and _lou_extParseChars thanks to tyler92.
  • Fix a heap buffer overflow in back_passDoAction thanks to Bert Frees and Liu Yuwei.
• A doc string has been added to lou_translateString and lou_translate outlining the buffer overflow risk thanks to Christian Egli and shuangxiangkan.
• Fix the default table search path on Windows thanks to Bert Frees and Bue Vester-Andersen

Braille table improvements

• Fix the back-translation of capital sign in Dutch thanks to Bert Frees

• Major improvements to Norwegian thanks to Lars Bjørndal

• Improvements to the UEB table thanks to James Bowden. In particular the handling of words containing "somer", e.g. "Somerville" was fixed.

• Hungarian improvements (added more exceptions) thanks to Attila Hammer.

• Add a new 6-dot table for literary Coptic braille, thanks to Eric Harvey.

• Added some missing symbols (including space) to the table for the International Phonetic Alphabet, thanks to James Bowden.

• Add a table for Portuguese 6-dot computer braille, thanks to Tiago

  1. Casal and Iván Argote-Pérez. The table is based on the official

  documentation in Portuguese.

• Add a new table for Cuneiform transliteration. It differs from the existing table cuneiform-transliterated.utb in that it does not adopt UEB dot patterns for the various diacritic marks. Instead it uses bespoke diacritics inspired by some IPA symbols. Thanks to Eric Harvey.

Other changes

• None

Deprecation notice

• None

Backwards incompatible changes

• Make value in table metadata field mandatory: It makes the metadata system clearer and simpler. Empty values have not been used anyway.

Invisible changes

New, renamed or removed tables

New

• cop.utb
• cuneiform-transliterated-compact.utb
• pt-comp6.utb

Renamed

• None

Removed

• None

Noteworthy changes in this release

A lot of intense work by Matt, Paul, Leonard, Eric, Sarah and Bert went into this release. There are new tables for Biblical Language and Ancient Near Eastern Studies. We have improvements for UEB, Spanish, Turkish and Hungarian. There are also a couple of new features, for example the handling of the language metadata has been extended to handle language ranges.

I'd like to thank all the very dedicated contributors for helping to push liblouis forward.

For a detailed list of all the changes refer to the list of closed issues.

New features

• lou_translate now has an option to specify a display table. This is especially useful when using a query. Before it would just default to a Unicode display table, thanks to Christian Egli.
• Meta data for language and region can now be specified using a IETF Extended Language Range. This is particularly useful for accessing tables related to a specific region, for instance. Thanks to Bert Frees.

Bug fixes

• Fix a memory problem in the getTable function. Thanks to the oss-fuzz project for reporting and to Christian Egli for fixing it.
• Fix a build problem on Windows related to strncasecmp which was reported at the last minute by Leonard de Ruijter and fixed by Christian Egli.

Braille table improvements

• Updates to Spanish grade 1 thanks to José Manuel Delicado:
  • Removed lowered digits. This allows certain punctuation to be written again.
  • Included digits6Dots.uti. This prevents the numsign from disappearing.
  • The symbols . and , are used as numeric mode chars.
  • Added the noback operator to the apostrophe symbol, since it's not very used and conflicts with the ..
• Extensive fixes to the UEB computer code thanks to James Bowden.
• Improvements to Turkish grade 2 thanks to Oğuz Uğur and Attila Hammer.
• Hungarian improvements (added more exceptions) thanks to Attila Hammer.
• There is now an include table for symbols referencing the critical apparatus in critical texts thanks to Matt Yeater.

Other changes

• Migrate to realloc-posix from gnulib, as realloc-gnu that we used before is now deprecated. Thanks to Christian Egli.
• Improvements to the manual thanks to Huan Wu.
• Thanks to Rob Beezer we now have instructions on how to install Python bindings in a virtual environment.

Deprecation notice

• None

Backwards incompatible changes

• If liblouis has been compiled for 16-bit Unicode it will now error if a table contains characters outside of that range. Previously it would issue a warning and ignore the problem. Thanks to Christian Egli.

Invisible changes

• None

New, renamed or removed tables

New

• critical-apparatus.uti
• he-common-consonants.uti
• he-common-vowels-ihbc.uti

Renamed

• None

Removed

• None

Noteworthy changes in this release

This release adds support for new braille tables, such as Thai contracted and Greek International US, and updates existing braille tables such as the Finnish table which has been updated to the latest braille standard. Benehalo has uncovered a number of security problems in the liblouis code and has thankfully also provided fixes for them by hardening the code.

I'd like to thank all the very dedicated contributors for helping to push liblouis forward.

For a detailed list of all the changes refer to the list of closed issues.

New features

• None

Bug fixes

• Benehalo discovered quite a number of security problems and kindly enough also just fixed them by adding some check guards and making the code more robust.

Braille table improvements

• Hungarian improvements (added more exceptions) thanks to Attila Hammer.
• The usage of match opcodes with empty pre- and post patterns, e.g. match - disk - 145-24-16 in some tables has been replaced with plain old always opcodes, as their implementation is much faster.
• A change to the Spanish back-translation that went into the last release was reverted as it caused more problems than it solved.
• The Dutch display table has been improved thanks to Bert Frees.
• The Finnish table has been updated to reflect the latest Finnish braille standard from 2021 (Pistekirjoituksen perusteet, 2021) thanks to Sami Maatta from Celia.
• Added a new Thai contracted braille table according to the guidelines of "Thai Braille Use Manual" of the National Braille Promotion and Development Fund. Thanks to Chatupol Nootatong.
• Add a table for Greek International US braille. This table is based on the composed Greek characters, unlike the existing English table which is based on the decomposed system. Thanks to Sarah Blake LaRose and Leonard de Ruijter.

Other changes

• None

Deprecation notice

• None

Backwards incompatible changes

• None

Invisible changes

• None

New, renamed or removed tables

New

• th-g1.utb
• grc-international-en-composed.utb

Renamed

None

Removed

None

Noteworthy changes in this release

We are very excited to bring this release as a lot of enthusiasm and energy has been gone into it. A lot of of new contributors have improved liblouis and added support for new languages to it. We now have support for many ancient languages like Akkadian, Ugaritic, Biblical Hebrew or Syriac. There is also support for Yiddish now. Also the Dutch tables got a very innovative upgrade.

I'd like to thank all the very dedicated contributors for helping to bring liblouis forward.

For a detailed list of all the changes refer to the list of closed issues.

New features

• There is a new uncontracted braille table for Yiddish, created by Abby Howell and Matthew Shifrin, a teacher and student of Yiddish at the Boston Worker's Circle.
• A group of incredibly dedicated people has been working on support for ancient languages, namely Matt Yeater and Paul Geoghegan with the help of Sarah Blake LaRose, Eric Harvey, Leonard de Ruijter and Abby Howell. The languages that are supported so far are:
  • Biblical Hebrew
  • Akkadian according to Borger system
  • Akkadian braille according to U.S. system
  • Syriac
  • Ugaritic
  • A table to provide support for the languages represented in the Cuneiform / Akkadian Writing systems

Bug fixes

• Fix a bug in the Python bindings that occurs on bigendian architectures, thanks to Daniel Garcia.
• Fix a memory problem when handling parameters in lou_translate, thanks to benehalo.

Braille table improvements

• Minor improvements to the Danish tables thanks to Bue Vester-Andersen:
  • Fixed a missing contraction for "gjorde".
• Improvements to Hungarian (added more exceptions) thanks to Attila Hammer.
• Corrections to the Malay Table thanks to lbk2907.
• Major improvements to Serbian backtranslation thanks to Aleksandar Sredojevic:
  • Backtranslation now works to both Cyrillic and Latin script.
• Fix the metadata for the Cantonese (Hong Kong) table thanks to André Polykanine.
• Improvements to backtranslation of Spanish grade 2 thanks to Iván Argote Pérez with the help of Abby Howell.
• Major upgrade to the Dutch tables thanks to Bert Frees. These tables use virtual dots in an innovative way to help sighted readers better understand the generated braille. They also use a new technique using m4 macros to simplify writing tables that use many repetitive opcodes.

Other changes

• The metadata now has support for unicode-range, i.e. whether the characters used in a table all fall inside the 0-U+FFFF Unicode range (ucs2), or whether some characters are in the > U+10000-U+10FFFF range (ucs4), thanks to Bert Frees.

Deprecation notice

• None

Backwards incompatible changes

• None

Invisible changes

• None

New, renamed or removed tables

New

• sr-Cyrl.ctb
• yi.utb
• hbo.utb
• akk.utb
• akk-borger.yaml
• syc.utb
• uga.utb
• cuneiform-transliterated.utb

Renamed

None

Removed

None

Noteworthy changes in this release

A lot of hard work by Andrey, Anthony, Attila, Bert, Bue and James went into this release. We have many improvements for UEB, better support for Belarusian and Ukrainian literary braille, a major upgrade for Turkish and miscellaneous enhancements for Kannada, Danish, Dutch, Esperanto and Malay. There is also a brand new table for classical Greek for Spanish braille readers.

I'd like to thank everybody for helping to bring liblouis forward.

For a detailed list of all the changes refer to the list of closed issues.

New features

• There is a new uncontracted braille table to read classical Greek characters according to the technical document B12-1 by the Spanish Braille Commission, thanks to Iván Novegil.

Bug fixes

• None

Braille table improvements

• Fixed issues with Kannada table kannada.cti thanks to Jake Kyle.
• Improvements to the Belarusian and Ukrainian literary braille tables thanks to Andrey Yakuboy:
  • Added support for back translation to the Belarusian and Ukrainian literary Braille tables (bel.utb and uk.utb respectively).
  • Changed the dot patterns for parentheses in the Ukrainian braille table from 126-345 to 2356.
  • Added new tables for Belarusian and Ukrainian braille with indication of capital letters (bel-detailed.utb and uk-detailed.utb respectively).
• Minor improvements to the Danish tables thanks to Bue Vester-Andersen:
  • Fixed definition of dollar sign ($)
  • Fixed rules for applying section sign (§)
  • Fixed contraction of some words
• Updates to UEB Contracted thanks to Anthony Tibbs:
  • Final letter groupsigns will not be used immediately following capital indicators (e.g. when the first letter of the groupsign is capitalized, or the groupsign follows a capitalized passage) per UEB Rule 10.8.2
• Updates to UEB thanks to James Bowden:
  • Back translating braille dot 3 now results in the ordinary apostrophe U+0027 rather than U+2019
  • Dot 5, used for "spaced digits" is only applied if a no-break space U+00a0 is used in the input. It fixes an issue with British postcodes which should be treated as two separate items. It also means that if someone wants spaced digits, numbers have to be joined together with no-break spaces deliberately.
• Do not collapse Unicode braille in Dutch, thanks to Davy Kager.
• Improvements to Hungarian (added more exceptions) thanks to Attila Hammer.
• Added support for back-translation to the Esperanto table, thanks to Bue Vester-Andersen.
• Major upgrade to the Turkish tables thanks to Oğuz UĞUR with lots of help from Attila Hammer.
  • add support for accented characters
  • add punctuation symbols from the Turkish Braille Writing Guideline
  • fix some contractions
• Corrections to the Malay Table thanks to lbk2907.

Other changes

• Duplicate emphasis classes are no longer warnings thanks to Christian Egli. There are cases where emphasis redefinition is legitimate. They are simply logged now.

Deprecation notice

• None

Backwards incompatible changes

• None

Invisible changes

• None

New, renamed or removed tables

New

• bel-detailed.utb
• uk-detailed.utb

Renamed

None

Removed

None

Noteworthy changes in this release

This release manages to bring as many new features, table improvements and bug fixes as we haven't seen in a long time. There are a number of major bug fixes that Bert worked on that fix bugs and inconsistencies. On the table side there is a new table for Filipino contracted braille, James has tirelessly been working on UEB improvements and (as usual) Attila worked on the Hungarian tables.

I'd like to thank everybody for helping to bring liblouis forward.

For a detailed list of all the changes refer to the list of closed issues.

New features

• lou_translate has a new option to specify the table to use with a query. See the updated documentation on how to make use of this new feature.

Bug fixes

• Fix a bug when matching context rules that begin with uppercase letters. Thanks to Anthony Tibbs for the very detailed bug report and thanks to Bert for the actual fix.
• Fix a memory corruption error in the code for the swap opcodes. Thanks to the OSS-fuzz project for reporting and to Christian Egli for fixing it.
• Change how emphasis phrases are determined. Also allow non-whole words at the start and end of a phrase. This fixes a problem in the UEB handling. Thanks to Ken Perry from APH for reporting and to Bert Frees for fixing it.
• Character definition and base rules are now always considered in the order in which they are defined in the table. In other words the topmost rule always wins. This fixes a number of problems and inconsistencies. Thanks to Bert Frees for the fix.
  • Fix a problem with compbrlAtCursor and compbrlLeftCursor that was reported by Daniel Mayr
• When a table does not define any capital marks, rules are now case sensitive thanks to Bert Frees

Braille table improvements

• Improvements to UEB
  • fixes to many words starting with letters "BE-".
  • Fix back translation of dots 26 - when standing alone; should be the word enough.
  • Fix assignment for the right double arrow and added left, up, and down double arrows.
  • Fix typos in a couple of rules.
  • Fix words beginning "CONS-" and "DISHD-" when written uppercase, all above thanks to James Bowden.
  • Ensure letter indicators are added to letters A-J following numbers in a capitalized passage, thanks to Anthony Tibbs.
• Improvements to Hungarian (added more exceptions) thanks to Attila Hammer.
• New table for Filipino contracted braille, thanks to Allan Mesoga and Dang Hoai Phúc.

Other changes

• The manywarnigs module from gnulib is no longer used. It caused massive slowdowns when compiling with gcc.
• Add the name of the concerned table when logging warnings thanks to Bert Frees for the patch and to Attila Hammer for the PR.

Deprecation notice

• None

Backwards incompatible changes

• The order of how character definition and base rules are considered has changed. All included tables were adapted, but if you maintain your own tables you might want to look at the change to see what you need to do.

Invisible changes

New, renamed or removed tables

New

• fil-g2.ctb

Renamed

None

Removed

None

Noteworthy changes in this release

This release contains support for two more languages, namely Thai and Romanian uncontracted braille. There is also the usual assortment of incremental improvements to Bengali, Danish, Hungarian, Swedish, Urdu and US English. To top it off, some great new contributors did some code cleanups and bug fixes.

I'd like to thank everybody for helping to bring liblouis forward.

For a detailed list of all the changes refer to the list of closed issues.

New features

• Added support for Thai braille thanks to Chatupol Nootatong, Dr. Wiraman Niyomphol and Itsawat Banlawanich from Thai braille fund and Foundation for the Blind in Thailand.
• Added a new Romanian 6 dot table, thanks to BAUM Engineering.

Braille table improvements

• Update to Urdu, removing old rule where dots 256 was used for Urdu letters TEH ALEF. Thanks to Jake Kyle.
• Added the dari character to Bengali thanks to Attila Hammer.
• Improvements to Hungarian (added more exceptions) thanks to Attila Hammer.
• Improvements to Hungarian to support users if typing physical braille display, thanks to Attila Hammer.
• Fixed the translation of Unicode ellipses in US English grade 1 and 2 tables thanks to Anthony Tibbs.
• Added some more symbols to the Nemeth character definitions thanks to Norbert Markus and Attila Hammer.
• Fixed a bug in Danish when back-translating a word contraction followed by "!", thanks to Bue Vester-Andersen.
• Improved handling of numbers and fixed a couple of whole-word contraction errors in Swedish braille thanks to Insyn Scandinavia AB.
• Added the possibility to retain the virtual dots on Dutch braille, so that it should be possible to distinguish between identical dot patterns that originate from different input. Thanks to Bert Frees.

Other changes

• Removed some code smell thanks to Aaron Cannon.
• Duplicate emphasis class definitions now cause a warning, thanks to Attila Hammer.

Deprecation notice

• None

Backwards incompatible changes

• The line continuation feature was removed, as it was no longer needed and caused problems (e.g. issue 239). The tables that depended on it were updated. Thanks to Aaron Cannon.
• Invalid attribute names now cause a table compilation error instead of just a warning. Thanks to Attila Hammer.

New, renamed or removed tables

New

None

Renamed

None

Removed

None

Noteworthy changes in this release

This release contains one new feature, namely the possibility to mark a test as an expected failure just for back- or forward translation. On top of that we have updates for Danish, Hungarian, Norwegian and Nemeth Braille plus the usual assortment of code cleanups and bug fixes.

I'd like to thank everybody for helping to bring liblouis forward.

For a detailed list of all the changes refer to the list of closed issues.

New features

• The YAML syntax to define tests has been improved so that expected failures can now contain a direction attribute. In other words you can now specify whether you expect a test to fail for both directions or just for forward or backward translation. For more details consult the updated documentation on how to write YAML tests. Thanks to Christian Egli.

Bug fixes

• Fix an out-of-bound write in brl_checks.c thanks to Bert Frees for the fix and to Han Zheng for the report.
• Clear the display table cache on lou_free thanks to Bert Frees for the fix and to David Korczynski for the report.

Braille table improvements

• Improvements to Hungarian (added more exceptions) thanks to Attila Hammer.
• Changed the Danish grade 2 tables in accordance with updates from the Danish Braille committee, thanks to thanks to Bue Vester-Andersen.
• Added support for more symbols in the Nemeth braille tables thanks to Norbert Markus.
• Improvements to Norwegian braille thanks to Lars Bjørndal
  • Put noback before acute sign to enable input at sign (email addresses).
  • Added more noback prefixes.
  • Corrected leftwards arrow and added definition for left right arrow.
  • Changed bullet point from 2356 to 36-36. According to the Norwegian braille standard, this is the preferred unordered list item prefix.

Other changes

• Clarify the README for Python thanks to Rob Beezer.
• Properly remove support for Python2 in the Python bindings again thanks to Rob Beezer.

Deprecation notice

• None

Backwards incompatible changes

• None

Invisible changes

• None

New, renamed or removed tables

New

• None

Renamed

• None

Removed

• None