api.fabaccess-api/utils.capnp
Nadja Reitzenstein 21574d8185 update
2022-10-31 15:20:12 +01:00

72 lines
3.4 KiB
Cap'n Proto

@0xed0c02f41fea6b5a;
interface L10NString {
# Any string type that is intended to be displayed to an user that is more
# than an identifier to be used as-is must be able to be localized into the
# users preferred language. This includes description, help messages, etc.
# but of course does not extend to usernames.
# TODO: Potentially make generic over the localized content (e.g. dates)?
# Can be done after the fact without braking protocol, so no big issue.
get @0 ( lang :Text ) -> ( lang :Text, content :Text );
# Retrieve the string in the given locale. The input parameter MUST be a
# RFC5646-formatted locale identifier (e.g: "en-US", "de-DE", "az-Arab-IR").
#
# If a server can't find a localized version matching exactly it MUST try to
# substitute it. Substitution MUST always return more specific matches for
# general queries. e.g. if "it" is requested and the server has "it-CH"
# available it returns this string.
#
# Substitution SHOULD NOT cross language barriers, e.g. returning "en-GB"
# for a string requested in "cy-GB". Substitution MUST NOT return a
# localization in a different language unless server has a priori knowledge
# that the user can read and understand said language.
#
# Substitution SHOULD prefer unspecified subtags over wrong subtags. If
# "es-AR" is requested and a server has "es", and "es-VE" available, "es"
# should be selected.
#
# A server MUST set the output `lang` field to the exact tag that the
# content it sends was written in and `content` to the localized string.
# e.g. If a string is requested for "sr" and the server has found a string
# that was configured as "sr-Cyrl-BA" the server sets lang to "sr-Cyrl-BA".
#
# If a server can't find a suitable substitute it MUST set the output
# `content` to a NULL pointer and set the output `lang` to the input `lang`
# it was passed.
# If a server can't parse a given `lang` tag it MUST set the output `lang`
# to NULL.
available @1 () -> ( langs :List(Text) );
# Returns the list of locales this content is available in.
}
struct UUID {
# UUID type used to identify machines.
#
# Consider using this algorithm for assembling the 128-bit integer:
# (assuming ISO9899:2018 shifting & casting rules)
# uint128_t num = (upper << 64) + lower;
# And then respectively this code for deconstructing it:
# uint64_t lower = (uint64_t) num;
# uint64_t upper = (uint64_t) (num >> 64);
lower @0 :UInt64;
# lower 8 bytes of the uuid, containing the LSB.
upper @1 :UInt64;
# upper 8 bytes of the uuid, containing the MSB.
}
using OID = Data;
# An OID is encoded as a sequence of varints. In this encoding the lower 7 bits
# of each octet contain data bits while the MSB indicates if the *following*
# octet is still part of this edge. It is the same encoding UTF-8 uses. To
# decode you simply collect octets until you find an octet <128 and then concat
# the data bits of all the octets you've accumulated, including the current one.
# This gives you the value of one node. Continue until you've exhausted the
# available data. This is a rather efficient encoding since almost all edges of
# the OID tree are smaller than 128 and thus encode into one byte.
# X.208 does *not* limit the size of nodes! However, a reasonable size limit is
# 128 bit per node, which is the size of the UUID nodes in the `2.25` subtree.