Don't rely on parenthesis used for `print` construct:
print 1 . print(2) + 3;
print PHP_EOL;
print 1 . (print(2)) + 3;
(PHP 4, PHP 5, PHP 7, PHP 8)
print — 输出字符串
$arg
) : int
输出 arg。
print 实际上不是函数（而是语言结构），所以可以不用圆括号包围参数列表。
和 echo 最主要的区别：
print 仅支持一个参数，并总是返回 1。
arg输入数据。
总是返回 1。
Example #1 print 范例
<?php
print("Hello World");
print "print() also works without parentheses.";
print "This spans
multiple lines. The newlines will be
output as well";
print "This spans\nmultiple lines. The newlines will be\noutput as well.";
print "escaping characters is done \"Like this\".";
// 可以在打印语句中使用变量
$foo = "foobar";
$bar = "barbaz";
print "foo is $foo"; // foo is foobar
// 也可以使用数组
$bar = array("value" => "foo");
print "this is {$bar['value']} !"; // this is foo !
// 使用单引号将打印变量名，而不是变量的值
print 'foo is $foo'; // foo is $foo
// 如果没有使用任何其他字符，可以仅打印变量
print $foo; // foobar
print <<<END
This uses the "here document" syntax to output
multiple lines with $variable interpolation. Note
that the here document terminator must appear on a
line with just a semicolon no extra whitespace!
END;
?>
Note: 因为是一个语言构造器而不是一个函数，不能被 可变函数 调用。
Don't rely on parenthesis used for `print` construct:
print 1 . print(2) + 3;
print PHP_EOL;
print 1 . (print(2)) + 3;
mvpetrovich of 2007 could just use single quotes as his string delimiters (see the example in the current documentation).
It's not ALWAYS appropriate, but generally it is best (the Zend Framework coding standards have a good take on this). It yields a number of interesting benefits:
1: Nobody will be tempted to write functions to replace backticks or other characters with double quotes. Such functions may cause a (negligible) loss of efficiency, and maybe other undesired effects.
2: You will be able to use double quotes without escaping. This is recommended (although not required) for HTML and XML attributes, as well as quoted text.
3: The script will hit the browser very slightly slightly faster since PHP doesn't have to scan through the string looking for variables, escaped characters, curly braces or other things.
4: Your code gets ten times easier to read. (as mvpetrovich pointed out)
If, in spite of these four excellent benefits, you really MUST still use double quotes to delimit boring old string constants (and seriously, why would you?), you could use the slightly less favourable single quotes as delimiters for most markup languages.
HTML served as HTML will even let you lay out unquoted attributes (yuck).
It should also be noted though that if you are just printing bare strings, you may as well shut off the php parser. The quickest way to send a string is to write it as plain text, OUTSIDE of the php tags. This will also make your code look excellent in a lot of syntax highlighters.
There are few disadvantages to doing this, if any. Output buffering still works. All your classes and objects and includes remain in place. Your script runs faster. World peace is obtained.
An update to the println function I wrote below, this is a more efficient, correct and returns a value (1, always; (print)).
<?php
function println($string_message = '') {
return isset($_SERVER['SERVER_PROTOCOL']) ? print "$string_message<br />" . PHP_EOL:
print $string_message . PHP_EOL;
}
?>
Be careful when using print. Since print is a language construct and not a function, the parentheses around the argument is not required.
In fact, using parentheses can cause confusion with the syntax of a function and SHOULD be omited.
Most would expect the following behavior:
<?php
if (print("foo") && print("bar")) {
// "foo" and "bar" had been printed
}
?>
But since the parenthesis around the argument are not required, they are interpretet as part of the argument.
This means that the argument of the first print is
("foo") && print("bar")
and the argument of the second print is just
("bar")
For the expected behavior of the first example, you need to write:
<?php
if ((print "foo") && (print "bar")) {
// "foo" and "bar" had been printed
}
?>
I wrote a println function that determines whether a \n or a <br /> should be appended to the line depending on whether it's being executed in a shell or a browser window. People have probably thought of this before but I thought I'd post it anyway - it may help a couple of people.
<?php
function println ($string_message) {
$_SERVER['SERVER_PROTOCOL'] ? print "$string_message<br />" : print "$string_message\n";
}
?>
Examples:
Running in a browser:
<?php println ("Hello, world!"); ?>
Output: Hello, world!<br />
Running in a shell:
<?php println ("Hello, world!"); ?>
Output: Hello, world!\n
the FAQTs article can be found archived at http://web.archive.org/web/20060601063513/http
://www.faqts.com/knowledge_base/view.phtml/aid/1/fid/40
(url split to get past the line-length limitation)
I have written a script to benchmark the several methods of outputting data in PHP: via single quotes, double quotes, heredoc, and printf. The script constructs a paragraph of text with each method. It performs this construction 10,000 times, then records how long it took. In total, it prints 160,000 times and records 16 timings. Here are the raw results.
Outputted straight to browser--
Single quotes: 2,813 ms
...with concatenation: 1,179 ms
Double quotes: 5,180 ms
...with concatenation: 3,937 ms
heredoc: 7,300 ms
...with concatenation: 6,288 ms
printf: 9,527 ms
...with concatenation: 8,564 ms
Outputted to the output buffer--
Single quotes: 8 ms
...with concatenation: 38 ms
Double quotes: 8 ms
...with concatenation: 47 ms
heredoc: 17 ms
...with concatenation: 49 ms
printf: 54 ms
...with concatenation: 52 ms
A nice graph of the script's output can be found here:
http://i3x171um.com/output_benchmarks/ob.gif
So what should you choose to print your text? I found several things out writing this.
First, it should be noted that the print and echo keywords are interchangeable, performance-wise. The timings show that one is probably an alias for the other. So use whichever you feel most comfortable with.
Second, if you've ever wondered which was better, the definitive answer is single quotes. Single quotes are at least four times faster in any situation. Double quotes, while more convenient, do pose a debatably significant performance issue when outputting massive amounts of data.
Third, stay away from heredoc, and absolutely stay away from [s]printf. They're slow, and the alternatives are there.
The source of my script can be found here:
http://i3x171um.com/output_benchmarks/ob.txt
DO NOT RUN THE SCRIPT ON THE INTERNET! Run it instead from localhost. The script outputs ~45 megabytes of text in an html comment at the top of the page by default. Expect the benchmark to take ~45 seconds. If this is too long, you can change the amount of iterations to a lower number (the results scale accurately down to about 1,000 iterations).
I have a small utility run from the command line that processes a potentially huge list of files. As it can take hours to complete, I stuck a
print '.';
statement in the body of the main loop to prove that something was happening.
For reasons unknown to me, the utiliity suddenly started buffering the output such that it printed nothing until completion, defeating the purpose of the running monitor. Adding flush() statements did nothing. The problem was solved by using
fputs(STDOUT, '.');
but I have no idea why.
This is a simple function for printing debug comments that I didn't think of for a long time. Maybe it'll serve you good too.
<?php
function printd($str) {
if ($debug) { echo $str; }
}
// ...
if ($valueCalculatedEarlierInTheScript == 3) {
doSomethingWithNoOutput();
printd("doSomethingWithNoOutput() has executed.");
}
?>
It's mostly just to make sure everything is running without having to go through everything and put in echo "Step #whatever has executed" whenever something mysterious isn't working.