Resolve a problem in font generation when y_offset is negative. #1268
Conversation
|
Found some char is wrong mapping in other languages, will try to fix as well. |
…r all languages including cjk.
|
Can anybody help to test different languages? This should fix all languages wrong char problem. |
|
I can build a set of all of them if needed, but would need some native/bilingual people to actually test/read it... |
|
I will try to test a few one anyway. It is interesting that this problem is here for some time but seems nobody reported this. |
|
And well, it should works anyway, now the CJK and other languages shared the same code in special character handling. The differences are just the code mapping, CJK used 2 byte encoding while others use the remaining space besides standard and extended, and I checked the language that has the most special character (CZ) and it seems ok in the simulator. |
|
I have tested 5 languages in my TX16S (en, cn, tw, es, cz) , all looks good. I think now need somebody to test the lowest config of a B/W radio with languages (en, es, and cz), and if these are ok, other languages should be good to go. |
|
@pfeerick As this will not affect B/W radios, I think it is good to merge now. Should merge to 2.6 branch as well. |
|
Sounds good to me... I'll give it a try later today and merge/cherrypick across. |
|
The old code tried to translate these Upper case to lower case: Not sure about the purpose, but it seems no harm for not doing translations, characters can still properly displayed anyway. |
|
To me... these pictures say it all (before and after) 😁
The lower case conversion may be used by the B&Ws... i.e.
Whilst the after for X9D+2019 is an improvement... I believe it should be |
|
Let me check the B/W font and see if Uppercase chars missing. Umm......... B/W used fixed fonts, encoding may have a problem then. Let me check how B/W fonts are managed first. Fixed fonts requires fixed encoding, and dynamic encoding will break anyway. |
|
Just another sample for you
|
|
I think I will try to isolate B/W and color encoding method, based on the fonts used. |
|
I used a fixed encoding for B/W LCD that matches the font files, hope this solves the problems in B/W LCD. @pfeerick Please help to do some tests. |
|
A few before/after images for cz and es, all looks good now.
|
…uages, and some wrong characters used (#1268) * Resolve a problem in font generation when y_offset is negative. * Change to use full dynamic detection of special chars, generalized for all languages including cjk. * Isolated colorlcd and bwlcd translation encoding method. * Fixed the cz char mapping. * Only cz requires subset lowercase handling.
…uages, and some wrong characters used (EdgeTX#1268) * Resolve a problem in font generation when y_offset is negative. * Change to use full dynamic detection of special chars, generalized for all languages including cjk. * Isolated colorlcd and bwlcd translation encoding method. * Fixed the cz char mapping. * Only cz requires subset lowercase handling.
Fixes #1199
Summary of changes: