If you don't need any additional detail in the .keyboard_info (fonts, examples, etc), then the build process will generate the BCP 47 codes in the .keyboard_info automatically from the .kps. That means, if users don't want to specify fonts or examples in the .keyboard_info, they can leave the languages tag out entirely of the keyboard_info (a minimal keyboard_info is still required which references license for example). The build process will in that situation generate the BCP 47 codes in the final .keyboard_info file based on what is found in the .kps.
Martin H should be scraping the .keyboard_info files on downloads.keyman.com/keyboards, not those in the source repository. The ones on downloads.keyman.com (e.g. https://downloads.keyman.com/keyboards/acoli/1.0/acoli.keyboard_info) include all the metadata scraped and are much more useful.
Even better is if he uses api.keyman.com to do queries -- then he's working with a published API that won't change under him.
From Marc Durdin Feb2020
If you don't need any additional detail in the .keyboard_info (fonts, examples, etc), then the build process will generate the BCP 47 codes in the .keyboard_info automatically from the .kps. That means, if users don't want to specify fonts or examples in the .keyboard_info, they can leave the languages tag out entirely of the keyboard_info (a minimal keyboard_info is still required which references license for example). The build process will in that situation generate the BCP 47 codes in the final .keyboard_info file based on what is found in the .kps.
Martin H should be scraping the .keyboard_info files on downloads.keyman.com/keyboards, not those in the source repository. The ones on downloads.keyman.com (e.g. https://downloads.keyman.com/keyboards/acoli/1.0/acoli.keyboard_info) include all the metadata scraped and are much more useful.
Even better is if he uses api.keyman.com to do queries -- then he's working with a published API that won't change under him.