README

print() is a fully type-safe, templated version of printf(). It supports printing to any stream, or even any object that supports insertion (<<).

Why?

Traditionally, printf() does not do any kind of type checking, and it uses C-style variadic functions. Modern compilers support printf() as an intrinsic, so they can do safety checks on the given arguments. However, this only works for static format strings, and cannot work on format strings that are, for example, read from a file during runtime. Also, printf() can only support basic datatypes, and cannot be overloaded for user-defined types.

How?

Using print() is very simple. The basic version is defined as follows:

template <typename... Ts>
void print (char const*, Ts&&...);

The first argument is the format string, which currently must be a char const*. The remaining arguments are either items to be printed or format parameters.

The format string's specification is practically identical to printf(). For example:

print ("Integer: %$; Float: %$;\n", 7, 7.5);

The %$ is the field specifier. It is analogous to %d or %f when using printf(), except it automatically detects the proper datatype.

print() even supports most of the format flags used by printf().

print ("Width   10: %10$;\n", 7);
print ("Width   ??: %*$;\n", 10, 7);
print ("Left-align: %-10$;\n", 7);

The complete field format specification is as follows:

%[flags][width][.[pflags]precision]$

flags can be any combination of the following:

Flag	Effect
-	Left-align
+	Show + for positive numbers.
0	Pad the field with zeroes instead of spaces.

width can either be a number to set as the field width, or it can be * to use the current parameter as the width.

pflags affect floating-point numbers, and can be the following:

Flag	Effect
e	Use scientific notation
f	Use fixed decimal point

precision works just like width, but instead sets the precision for floating-point numbers. If the f flag is used, precision will be the number of digits after the decimal point.

Using Streams

By default, print() will print to std::cout. However, there is a second form of print():

template <typename S, typename... Ts>
void print (S&&, char const*, Ts&&...);

It's the same as the first version, except the first parameter is the stream to use. The given stream does not have to be derived from std::ostream, but print() uses stream sytax (<< and manipulators), so be aware of possible incompatabilities.

License

Print is licensed under the unmodified zlib/libpng liscense. Full text can be found here.

For those unfamiliar, this basically means that you can do whatever you want with the source code as long as you don't claim ownership or remove the license.

Contact

Jeramy Harrison

Email: dbralir@gmail.com