-
-
Notifications
You must be signed in to change notification settings - Fork 631
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
LibLouis 3, UEB table name/file names, and provisions for other enhancements. #6110
Conversation
One of the enhancements in LibLouis 3 is centralized pattern matching and translation routines. Among other things, it allows emphasais to be handled better. Thus add pattern.c to list of files to be compiled as part of compiling NVDA from source, as without this, MS Linker (link.exe) presents unresolved symbol error.
…ccess#4194. As the old uebc-g* tables are no longer maintained, it was decided to use the new en-ueb-g* set. Not only this makes the file names more consistent, it allows exciting changes to be included as part of LibLouis 3.0.0. For string compatibility with old NVDA releases, table name presented to users will still be named Unified English Braille Code grade 1/2, with braille table name entries now mapped under en-ueb-g*.
The Dutch table name also changed, and possibly the Norwegian one. |
@dkager: Do you know the file names of these tables? I found Norwegian (no-no-comp8) but would like to get a list of file name changes so I can commit them in a single commit. Thanks. |
…name for Telugu. re nvaccess#6109. Davy Kager reports that Norwegian table file name has changed (confirmed), thus added the new file name. Telugu: it is actually called 'Telugu', not 'Telegu' accoridng to Wikipedia, table header itself and other sources. No impact for translators as it is just English name that's different.
In case of te-in: it is actualy called Telugu, not Telegu. No impact for translators (just English label has been corrected). Thanks. |
Hey Joseph, I wouldn't be worried about commit cleanlyness too much, because @jcsteh On 6/24/2016 3:03 PM, Joseph Lee wrote:
Websites: email me at [email protected] mailto:[email protected] |
The Dutch tables are nl-BE-g0.utb (Belgium) and nl-NL-g0.utb (Netherlands). |
Hi, how about g1 files? Thanks. From: Davy Kager [mailto:[email protected]] The Dutch tables are nl-BE-g0.utb (Belgium) and nl-NL-g0.utb (Netherlands). — |
The table has been renamed. From: Joseph Lee [mailto:[email protected]] Hi, how about g1 files? Thanks. From: Davy Kager [ mailto:[email protected] mailto:[email protected]] The Dutch tables are nl-BE-g0.utb (Belgium) and nl-NL-g0.utb (Netherlands). — — |
Ug. As is far too often the case in liblouis land, this is ridiculously confusing. The old tables still exist... but now we have these new ones.
|
I thought they had been replaced. Something to look at.
Dutch only has one grade, so dropping the "grade 1" in the name sounds fine. This is uncontracted literary braille. |
Hi, @dkager: Thanks for the answer. Will apply it. For other table maintainers: I think now is the time to tell us which tables are which, especially those that were modified (not due to opcode changes). Thanks. |
According to Davy Kager, there is really one braille code for Dutch (no separate grade 0 nor grade 1). Thus: * Dutch (Netherlands) Grade 1 -> Dutch (Netherlands) * Duethc (Belgium) grade 1 -> Duethc (Belgium) Also, the old g0 tables are incompatible iwth new optocdes introduced as part of LibLouis 3 (grade 1 tables use the new ones).
|
…n routines that loads liblouis.dll. nvaccess#6109
Hi, In the end, due to file inclusion issues and build issues on Windows (vanilla Liblouis def), it was decided that NVDA will use its own liblouis.def. Thanks. |
Merge note: This should not be merged to master until liblouis 3.0.0 final is released. Once that happens, do the following:
What's New entry: in Changes:
|
Instead of the NMake file using liblouis.def we should probably have MSVC generate one. This requires some code changes I'll have to discuss with Christian. |
No, the 2005 braille standard only includes an uncontracted 6-dot code. In the past we had multiple levels of contractions (3 I believe), but now we don't. So this is grade 0, but since there are no other grades we can leave that out IMO. On a final note, the Netherlands and Belgium tables differ only in some character definitions. The rules are the same for the whole Dutch language region. At some point we wanted to consolidate the two tables, but it hasn't happened yet. |
Merge note: On further reflection and discussion, this should not be merged to master until #6140 (backwards compat for renamed tables) is done. |
Hi, first commit on #6140 (for UEB series and Norwegian) will be done tonight, the rest will be done once beta comes out. Thanks. From: James Teh [mailto:[email protected]] Merge note: On further reflection and discussion, this should not be merged to master until #6140 #6140 (backwards compat for renamed tables) is done. — |
…w filenames. re nvaccess#6140. Given that people would be surprised if some tables would revert back to defaults, it would be handy to provide a short migration routine to map old table filenames to new ones. Thus, a tempoary routine and a map is introduced to handle it (to be removed at a future date).
More table renames for Danish (6-dot series) and Dutch, along with entries for them in migration map.
I want to get this into 2016.3, so in the interest of speed, I've made a few tiny changes myself (rather than going back and forth with a tiny review). |
@michaelDCurran, can you please review these commits? I reviewed most of this, but then made some minor tweaks myself in the interest of speed when I was trying to get this into 2016.3 (a decision we obviously reversed). So, my small tweaks should probably be reviewed. |
I am happy with these 4 commits. |
What's New entry: in Changes:
Fixes #6109. Fixes #4194. Fixes #6220. Fixes #6140.