BRF (Braille Format), aka ASCII Braille, is a usual way of encoding formated braille texts, ready to submit to a braille embosser. One can find a lot of .brf file on internet, so it would be very helpful to have this encoding in iconv. Attached is a simple table-based implementation (not tested, but generated from brltty tools) with a test-case (written by hand and tested with brltty tools).
Created attachment 1037 [details] libc/localedata/charmaps/BRF
Created attachment 1038 [details] libc/iconvdata/brf.c
Created attachment 1039 [details] libc/iconvdata/testdata/BRF Testcase for usual BRF range (0x20-0x5f)
Created attachment 1040 [details] libc/iconvdata/testdata/BRF..UTF8 Result of conversion of the previously-attached testcase into UTF-8
Note: there are duplicates. Strict BRF only defines 0x00-0x5f: 0x00-0x1f for usual ascii commands, and 0x20-05f is bijective with 6-dots patterns. In addition to that, a-z{}|~` are usually defined as I described in the table because it often happens that .brf files (erroneously) include some of them.
Mmm, thanks to a debian snapshot, I could realize that duplicates pose problem. Here are files without duplicates, which actually got tested.
Created attachment 1043 [details] libc/localedata/charmaps/BRF
Created attachment 1044 [details] libc/iconvdata/testdata/BRF
Created attachment 1045 [details] libc/iconvdata/testdata/BRF..UTF8
Any remaining problem on this? This charset just got registered by IANA.
I've added the files and the required infrastructure changes upstream.