Quick debugging methods :
@print($a);
is equivalent to
if isset($a) echo $a ;
@a++;
is equivalent to
if isset($a) $a++ ;
else $a = 1;
PHP 支持一个错误控制运算符：@。当将其放置在一个 PHP 表达式之前，该表达式可能产生的任何错误信息都被忽略掉。
如果用 set_error_handler() 设定了自定义的错误处理函数，仍然会被调用，但是此错误处理函数可以（并且也应该）调用 error_reporting()，而该函数在出错语句前有 @ 时将返回 0。
如果激活了 track_errors 特性，表达式所产生的任何错误信息都被存放在变量 $php_errormsg 中。此变量在每次出错时都会被覆盖，所以如果想用它的话就要尽早检查。
<?php
/* Intentional file error */
$my_file = @file ('non_existent_file') or
die ("Failed opening file: error was '$php_errormsg'");
// this works for any expression, not just functions:
$value = @$cache[$key];
// will not issue a notice if the index $key doesn't exist.
?>
Note: @ 运算符只对表达式有效。对新手来说一个简单的规则就是：如果能从某处得到值，就能在它前面加上 @ 运算符。例如，可以把它放在变量，函数和 include 调用，常量，等等之前。不能把它放在函数或类的定义之前，也不能用于条件结构例如 if 和 foreach 等。
参见 error_reporting() 及手册中错误处理及日志函数的有关章节。
目前的"@"错误控制运算符前缀甚至使导致脚本终止的严重错误的错误报告也失效。这意味着如果在某个不存在或者敲错了字母的函数调用前用了"@"来抑制错误信息，那脚本会没有任何迹象显示原因而死在那里。
Quick debugging methods :
@print($a);
is equivalent to
if isset($a) echo $a ;
@a++;
is equivalent to
if isset($a) $a++ ;
else $a = 1;
In PHP it is extremely beneficial to turn all notices into exceptions. This helps in creating bug free code through finding errors sooner rather than later. It also helps reduce the impact of bugs as code when entering an erroneous state ends sooner rather than later and at all. In the worst case without that you can have much more scenarios where code fails yet appears as though it succeeded.
However there are rare cases in which notices and warnings are produced where the above behavour might be unproductive. Worse yet your error handling will kick out that exception before the function gets to return.
They are rare cases such as socket handling where certain states are expressed through errors causing ambiguity.
What is PHP's behavior for a variable that is assigned the return value of an expression protected by the Error Control Operator when the expression encounteres an error?
Based on the following code, the result is NULL (but it would be nice if this were confirmed to be true in all cases).
<?php
$var = 3;
$arr = array();
$var = @$arr['x']; // what is the value of $var after this assignment?
// is it its previous value (3) as if the assignment never took place?
// is it FALSE or NULL?
// is it some kind of exception or error message or error number?
var_dump($var); // prints "NULL"
?>
There is no reason to NOT use something just because "it can be misused". You could as well say "unlink is evil, you can delete files with it so don't ever use unlink".
It's a valid point that the @ operator hides all errors - so my rule of thumb is: use it only if you're aware of all possible errors your expression can throw AND you consider all of them irrelevant.
A simple example is
<?php
$x = @$a["name"];
?>
There are only 2 possible problems here: a missing variable or a missing index. If you're sure you're fine with both cases, you're good to go. And again: suppressing errors is not a crime. Not knowing when it's safe to suppress them is definitely worse.
While you should definitely not be too liberal with the @ operator, I also disagree with people who claim it's the ultimate sin.
For example, a very reasonable use is to suppress the notice-level error generated by parse_ini_file() if you know the .ini file may be missing.
In my case getting the FALSE return value was enough to handle that situation, but I didn't want notice errors being output by my API.
TL;DR: Use it, but only if you know what you're suppressing and why.
Prepending @ before statement like you are doing a crime with yourself.
I was wondering if anyone (else) might find a directive to disable/enable to error operator would be a useful addition. That is, instead of something like (which I have seen for a few places in some code):
<?php
if (defined(PRODUCTION)) {
@function();
}
else {
function();
}
?>
There could be something like this:
<?php
if (defined(PRODUCTION)) {
ini_set('error.silent',TRUE);
}
else {
ini_set('error.silent',FALSE);
}
?>
This operator is affectionately known by veteran phpers as the stfu operator.
If you use the ErrorException exception to have a unified error management, I'll advise you to test against error_reporting in the error handler, not in the exception handler as you might encounter some headaches like blank pages as error_reporting might not be transmitted to exception handler.
So instead of :
<?php
function exception_error_handler($errno, $errstr, $errfile, $errline )
{
throw new ErrorException($errstr, 0, $errno, $errfile, $errline);
}
set_error_handler("exception_error_handler");
function catchException($e)
{
if (error_reporting() === 0)
{
return;
}
// Do some stuff
}
set_exception_handler('catchException');
?>
It would be better to do :
<?php
function exception_error_handler($errno, $errstr, $errfile, $errline )
{
if (error_reporting() === 0)
{
return;
}
throw new ErrorException($errstr, 0, $errno, $errfile, $errline);
}
set_error_handler("exception_error_handler");
function catchException($e)
{
// Do some stuff
}
set_exception_handler('catchException');
?>
If you're wondering what the performance impact of using the @ operator is, consider this example. Here, the second script (using the @ operator) takes 1.75x as long to execute...almost double the time of the first script.
So while yes, there is some overhead, per iteration, we see that the @ operator added only .005 ms per call. Not reason enough, imho, to avoid using the @ operator.
<?php
function x() { }
for ($i = 0; $i < 1000000; $i++) { x(); }
?>
real 0m7.617s
user 0m6.788s
sys 0m0.792s
vs
<?php
function x() { }
for ($i = 0; $i < 1000000; $i++) { @x(); }
?>
real 0m13.333s
user 0m12.437s
sys 0m0.836s
Be aware that using @ is dog-slow, as PHP incurs overhead to suppressing errors in this way. It's a trade-off between speed and convenience.
After some time investigating as to why I was still getting errors that were supposed to be suppressed with @ I found the following.
1. If you have set your own default error handler then the error still gets sent to the error handler regardless of the @ sign.
2. As mentioned below the @ suppression only changes the error level for that call. This is not to say that in your error handler you can check the given $errno for a value of 0 as the $errno will still refer to the TYPE(not the error level) of error e.g. E_WARNING or E_ERROR etc
3. The @ only changes the rumtime error reporting level just for that one call to 0. This means inside your custom error handler you can check the current runtime error_reporting level using error_reporting() (note that one must NOT pass any parameter to this function if you want to get the current value) and if its zero then you know that it has been suppressed.
<?php
// Custom error handler
function myErrorHandler($errno, $errstr, $errfile, $errline)
{
if ( 0 == error_reporting () ) {
// Error reporting is currently turned off or suppressed with @
return;
}
// Do your normal custom error reporting here
}
?>
For more info on setting a custom error handler see: http://php.net/manual/en/function.set-error-handler.php
For more info on error_reporting see: http://www.php.net/manual/en/function.error-reporting.php
Be aware of using error control operator in statements before include() like this:
<?PHP
(@include("file.php"))
OR die("Could not find file.php!");
?>
This cause, that error reporting level is set to zero also for the included file. So if there are some errors in the included file, they will be not displayed.
Error suppression should be avoided if possible as it doesn't just suppress the error that you are trying to stop, but will also suppress errors that you didn't predict would ever occur. This will make debugging a nightmare.
It is far better to test for the condition that you know will cause an error before preceding to run the code. This way only the error that you know about will be suppressed and not all future errors associated with that piece of code.
There may be a good reason for using outright error suppression in favor of the method I have suggested, however in the many years I've spent programming web apps I've yet to come across a situation where it was a good solution. The examples given on this manual page are certainly not situations where the error control operator should be used.
I was confused as to what the @ symbol actually does, and after a few experiments have concluded the following:
* the error handler that is set gets called regardless of what level the error reporting is set on, or whether the statement is preceeded with @
* it is up to the error handler to impart some meaning on the different error levels. You could make your custom error handler echo all errors, even if error reporting is set to NONE.
* so what does the @ operator do? It temporarily sets the error reporting level to 0 for that line. If that line triggers an error, the error handler will still be called, but it will be called with an error level of 0
Hope this helps someone
If you want to log all the error messages for a php script from a session you can use something like this:
<?php
session_start();
function error($error, $return=FALSE) {
global $php_errormsg;
if(isset($_SESSION['php_errors'])) {
$_SESSION['php_errors'] = array();
}
$_SESSION['php_errors'][] = $error; // Maybe use $php_errormsg
if($return == TRUE) {
$message = "";
foreach($_SESSION['php_errors'] as $php_error) {
$messages .= $php_error."\n";
}
return $messages; // Or you can use use $_SESSION['php_errors']
}
}
?>
Hope this helps someone...
error_reporting()==0 for detecting the @ error suppression assumes that you did not set the error level to 0 in the first place.
However, typically if you want to set your own error handler, you would set the error_reporting to 0. Therefore, an alternative to detect the @ error suppression is required.
To suppress errors for a new class/object:
<?php
// Tested: PHP 5.1.2 ~ 2006-10-13
// Typical Example
$var = @some_function();
// Class/Object Example
$var = @new some_class();
// Does NOT Work!
//$var = new @some_class(); // syntax error
?>
I found this most useful when connecting to a
database, where i wanted to control the errors
and warnings displayed to the client, while still
using the class style of access.
If you wish to display some text when an error occurs, echo doesn't work. Use print instead. This is explained on the following link 'What is the difference between echo and print?':
http://www.faqts.com/knowledge_base/view.phtml/aid/1/fid/40
It says "print can be used as part of a more complex expression where echo cannot".
Also, you can add multiple code to the result when an error occurs by separating each line with "and". Here is an example:
<?php
$my_file = @file ('non_existent_file') or print 'File not found.' and $string = ' Honest!' and print $string and $fp = fopen ('error_log.txt', 'wb+') and fwrite($fp, $string) and fclose($fp);
?>
A shame you can't use curly brackets above to enclose multiple lines of code, like you can with an if statement or a loop. It could make for a single long line of code. You could always call a function instead.
Better use the function trigger_error() (http://de.php.net/manual/en/function.trigger-error.php)
to display defined notices, warnings and errors than check the error level your self. this lets you write messages to logfiles if defined in the php.ini, output
messages in dependency to the error_reporting() level and suppress output using the @-sign.