Closed okapia closed 8 years ago
This is a minor nit but the normal Unix conventions would be to have all the grammars in a fixed location. So instead of /usr/share/rsync/rsync.grammar, it would be something like /usr/share/cli-grammar/rsync
Fixed.
That'd also allow an environment variable to list where to find them.
Right, some sort of $CLI_GRAMMAR_PATH envvar with the same syntax as $PATH.
Okay, it might need a catchy name and there might be a file extension. Not sure if you were planning yaml, json, xml or something custom?
How about "grammy" for the name? I went ahead and grabbed https://github.com/grammy-org/, hopefully the music people won't try to seize it.
Re file format: I intentionally haven't given that any thought yet. I was thinking the first task would be to write up what features the file format should have. Only later will deciding on a particular encoding come up. That is: data model before data encoding. I now filed this as #2.
Thanks for your input!
This is a minor nit but the normal Unix conventions would be to have all the grammars in a fixed location. So instead of /usr/share/rsync/rsync.grammar, it would be something like /usr/share/cli-grammar/rsync
That'd also allow an environment variable to list where to find them.
Okay, it might need a catchy name and there might be a file extension. Not sure if you were planning yaml, json, xml or something custom?