Node:Simple Data Types, Next:Connecting to the Server, Previous:Notation, Up:Fundamentals
INT32, INT16, INT8 and BOOL are non-negative
integers which must fit in 32, 16, 8 and 1 bits, respectively. They are
transmitted to the server in ASCII-encoded decimal notation.
HOLLERITH denotes character strings of arbitrary length. They are
transmitted as nHtext where text is the string and
n is the number of characters in text in decimal
notation. All byte values are allowed in the string itself, including
nulls.
Long live FORTRAN!
The character set used in the strings is not yet specified by Protocol A. In the future, some Unicode encoding will probably be used, but it is not yet decided which one or how the transition will be handled. Bug 99 is about the need for a Unicode roadmap; check that bug for the current state of the plans.
For now, which character set to use is a local policy of each server
installation. There is not yet any way in the protocol to specify the
character set that a certain server uses. Most clients currently
assume that ISO 8859-1 (Latin-1) is used, and the default collate
table of lyskomd also assumes ISO 8859-1. Conference names must
currently use an 8-bit character set encoding where whitespace is
defined as in ASCII, or conference matching won't work.
get-collate-table contains some more information about
character set issues.
BITSTRING is a string of bits, commonly used for a set of
boolean-valued flags. Bit strings are denoted as
BITSTRING ( name-1; name-2; name-3; ... )
in this specification. They are transmitted as a sequence of ASCII ones and zeroes in the order the fields are listed.
For instance, given the specification
shape-of-world ::= BITSTRING (
is-flat;
is-round;
is-2d;
is-3d;
)
most peoples idea of shape-of-world would be sent as 0101
(round and three-dimensional.)
ENUMERATION is an integer constant. It is transmitted as an INT32,
but only fixed values are permitted. Clients should be prepared to
receive numbers outside the enumeration and either handle this
gracefully as an error or use a reasonable default value in place of an
invalid enumeration value.
An enumeration is specified as
ENUMERATION (
name-1=value-1;
name-2=value-2;
name-3=value-3;
...
)
This specification states that name-1 is represented by the integer value-1, name-2 is represented by value-2 and name-3 is represented by value-3.
An enumeration can also be inherited from a SELECTION datatype:
Info-type ::= ENUMERATION-OF(Misc-Info)
This means that Info-type is an enumeration, that contains the same keys and values as the SELECTION Misc-Info.
For example, in the following specification, the constant guwal will be
transmitted as the integer 2, ciokwe as the integer 3, and hopi as the
integer 5.
language ::= ENUMERATION ( hakka = 1;
guwal = 2;
ciokwe = 3;
yoruba = 4;
hopi = 5;
)
ARRAY is a list of a fixed number of elements of a single type.
The specification for an array is ARRAY type where
type is the type of the array elements.
Arrays are transmitted as an n { element
element ... } where n is the number of elements and
each element is an element of the array. A special case is when
the array is empty, in which case the server transmits it as 0 *.
Note that the client must always transmit empty arrays as 0 {
}.
In some calls the client can ask the server not to send the array
contents, only its length. In these cases the array is transmitted as
n * where n is the number of elements in the array.
SELECTION is tagged data. It consists of an INT32 selector
followed by a tail of an arbitrary type and is specified as
SELECTION (
n=name tail : type;
n=name tail : type;
...
)
where each n is the selector value, name the selector
name and tail the name of the selector tail and type
its type. name and tail are often very similar, such
as sent-by and sender.
When transmitted, the selector is transmitted as an INT32 followed by
the tail belonging to that selector. For instance, given the
specification
description ::= SELECTION (
1=name the_name : HOLLERITH;
2=age years : INT32;
)
two legal messages of the type description are 1
4HJohn and 2 18.
RPC is a notation used to specify calls to the server. An RPC
specification has the following form:
RPC (
call [n] ( request ) -> ( reply ) ;
call [n] ( request ) -> ( reply ) ;
)
where each call is the name of a call, n is the call number, request is a single data element sent as a request and reply is a single data element sent in reply from the server.
RPC calls are transmitted as n request where n and request have the same meaning as above. Note that in the client-server dialog a reference number must also be supplied with each request. This reference number is not part of the RPC itself, but is required for communications; see Client-Server Dialog.
Structures are collections of several data items. In the specification
they are written as
( name : type ;
name : type ;
...
)
where each name is the name of a data field and the corresponding type is its type.
Structures are transmitted as a sequence of their fields.