The way ECLiPSe terms are printed can be customised in a number of ways. The most flexible predicates to print terms are write_term/3 and printf/3. They both allow all variants of term output, but the format is specified in a different way. The following figure gives an overview.
| write_term/2,3 | ||
| output option | printf/2,3 %..w | Meaning |
| as(term) | do not assume any particular meaning of the printed term | |
| as(clause) | C | print the term as a clause (apply clause transformations) |
| as(goal) | G | print the term as a goal (apply goal transformations) |
| attributes(none) | do not print any variable attributes | |
| attributes(pretty) | m | print attributes using the corresponding print handlers |
| attributes(full) | M | print the full contents of all variable attributes |
| compact(false) | print blank space around operators, after commas, etc. | |
| compact(true) | K | don’t print blank space unless necessary |
| depth(Max) | <Max> | print the term only up to a maximum nesting depth of Max (a positive integer) |
| depth(0) | observe the stream-specific or global flag print_depth | |
| depth(full) | D | print the whole term (may loop when the term is cyclic!) |
| dotlists(false) | write lists in square bracket notation, e.g., [a,b] | |
| dotlists(true) | . | write lists as terms with functor ./2 |
| newlines(false) | print newlines inside quotes as the escape
sequence \n | |
| newlines(true) | N | print newlines as line breaks even inside quotes |
| nl(false) | do not add a newline | |
| nl(true) | L | print a newline sequence (as with nl/1) after the term. |
| fullstop(false) | do not add a fullstop | |
| fullstop(true) | F | terminate the term with a fullstop, so it can be read back. |
| flush(false) | do not force a flush after printing the term | |
| flush(true) | b | flush after printing the term |
| numbervars(false) | do not treat ’$VAR’/1 terms specially | |
| numbervars(true) | I | print terms of the form ’$VAR’(N) as named variables: ’$VAR’(0) is printed as A, ’$VAR’(25) as Z, ’$VAR’(26) as A1 and so on. When the argument is an atom or a string, just this argument is printed. |
| operators(true) | obey operator declarations and print prefix/infix/postfix | |
| operators(false) | O | ignore operator declarations and print functor notation |
| portrayed(false) | do not use portray/1,2 | |
| portrayed(true) | P | call the user-defined predicate portray/1,2 for printing |
| precedence(Prec) | print assuming given context precedence | |
| quoted(false) | do not print quotes around strings or atoms | |
| quoted(true) | Q | quote strings and atoms if necessary |
| transform(true) | apply portray transformations (write macros) | |
| transform(false) | T | do not apply portray transformations (write macros). |
| variable_names(VarNames) | use the given variable names when printing the corresponding variables. VarsNames is a list of terms of the form Name=Var. | |
| variables(default) | print variables using their source name (if available) | |
| variables(raw) | v | print variables using a system-generated name, e.g., _123 |
| variables(full) | V | print variables as source name plus number, e.g., Alpha_132 |
| variables(anonymous) | _ | print every variable as a simple underscore |
Overview of term output options (see write_term/3 for more details)
The write_term/2 and write_term/3 predicates print a single ECLiPSe term and accept a list of output options (first column in the table).
The printf/2 and printf/3 predicates are similar to C’s printf(3) function, but provide additional format characters for printing ECLiPSe terms. The basic format string for printing arbitrary terms is %w. Additional format characters can go between % and w, according to the second column in the table.
For example, the following pairs of printing goals are equivalent:
printf("%mw", [X]) <-> write_term(X, [attributes(pretty)])
printf("%O.w", [X]) <-> write_term(X, [operators(false),dotlist(true)])
printf("%5_w", [X]) <-> write_term(X, [depth(5),variables(anonymous)])
The other term output predicates write/1, writeln/1, writeq/1, write_canonical/1, display/1, print/1 can all be defined in terms of write_term/2 (or, similarly in terms of printf/2) as follows:
write(X) :- write_term(X, [numbervars(true)]).
writeln(X) :- write_term(X, [numbervars(true)]), nl.
writeq(X) :- write_term(X, [variables(raw), attributes(full),
transform(false), quoted(true), depth(full),
numbervars(true)]).
write_canonical(X) :-
write_term(X, [variables(raw), attributes(full),
transform(false), quoted(true), depth(full),
dotlist(true), operators(false)]).
display(X) :- write_term(X, [dotlist(true), operators(false)]).
print(X) :- write_term(X, [portrayed(true), numbervars(true)]).
It is possible to set default output options for an output stream in order to globally affect all output to this particular stream. The set_stream_property/3 predicate can be used to assign default options (in the same form as accepted by write_term/3) to a stream. These options will then be observed by all output predicates which do not override the particular option.