I'm learning OO PHP and am trying to get some of the coding practices straight. Here is a pared down version of some code I'm using for error (and exception) handling:
final class MyErrorExceptionHandler {
private $level = array(); // error levels to be handled as standard errors
private $path = array(); // full path to file
private $path_short; // filename plus working dir
public function myErrorHandler($severity, $message, $file, $line) {
if (error_reporting() & $severity) { // error code is included in error_reporting
$this->level = array(E_WARNING => 'warning',
E_NOTICE => 'notice',
E_USER_WARNING => 'user warning',
E_USER_NOTICE => 'user notice');
if (array_key_exists($severity, $this->level)) { // handle as standard error
/*$this->severity = $severity;
$this->message = $message;
$this->file = $file;
$this->line = $line;*/
$this->printMessage($severity, $message, $file, $line);
} else { // fatal: E_USER_ERROR or E_RECOVERABLE_ERROR use php's ErrorException converter
throw new ErrorException($message, 0, $severity, $file, $line);
}
}
} // fn myErrorHandler
private function printMessage($severity, $message, $file, $line) {
echo ucfirst($this->level[$severity]) . ': ' . $message;
$this->shortenPath($file);
echo ' in ' . $this->path_short . ' on line ' . $line;
} // fn printMessage
private function shortenPath($file) {
$this->path_short = $file;
$this->path = explode(DIRECTORY_SEPARATOR, $file);
if (count($this->path) > 2) { // shorten path to one dir, if more than one dir
$this->path_short = array_pop($this->path); // filename
$this->path_short = end($this->path) . DIRECTORY_SEPARATOR . $this->path_short; // dir+file
}
} // fn shortenPath
} // cl MyErrorExceptionHandler
The title of this question is probably a little bit off because I'm not 100% on the terminology. Basically I'm trying to figure out a few things.
Is it right to explicitly declare $level and $path as arrays?
Should $level be declared as it is (and made $this->level)? If so, have I assigned its value (E_WARNING etc.) in a wise place? Would the constructor (not shown here) be a smarter choice?
Note the commented block in myErrorHandler(). Originally I had declared all of these properties at the top of the class, and then called $this->printMessage() without any parameters. Which is the more correct way? If I keep the code as is, would I want to then use $this->severity = $severity etc. inside printMessage()?
So, would it be better to:
replace
$this->shortenPath($file);
echo ' in ' . $this->path_short . ' on line ' . $line;
with
$path_short = $this->shortenPath($file);
echo ' in ' . $path_short . ' on line ' . $line;
ultimately, and give a return value in shortenPath()?
I realize this is a mishmash of several different questions, but what I'm trying to get at is a common inquiry about the proper style of declaring/using variables/properties, specifically when dealing with methods.
To summarize: When should I use $this->foo = $foo?
EDIT: sorry, I have assumed below that you would create a new instance of the 'object' with each error which obviously you are not doing. Just edited my answer to reflect this.
"When should I use $this->foo = $foo?"
There can be several cases in which you would do this, but it's usually if you create $foo within a method and wish to have that then accessed by the entire object.
For example, if you wanted to call on an object and use that within this particular object (if it doesn't make sense to extend). You would do something like:
$foo = new DataModel();
$this->foo = $foo;
OR
$this->foo = new DataModel();
That object may be a decorator or something else related to error handling and the above code would usually feature in your constructor. You could then access the methods of that object any time by using:
$this->foo->objectMethod();
..and to express something noted in the comments to this answer:
"would you assign $file to the object as that is used in several methods?"
I wouldn't assign $file to the object,
here's why. The semantics of the word
"property" means "belongs to". In your
case, your class is a error handler.
$file doesn't belong to the error
handler, it belongs to an error
instance. If your class was
MyErrorHandler_Error (created for each
instance of a triggered error), then
$file would be a property of that
class, along with $line and $level.
To answer what I can from your other questions:
It's neither. I would consider it preference.
Yes - any variables or values which should be available to your entire object and required for the object to run properly, should probably be set within your constructor, if not within your variable declarations (not sure of terminology there) at the top of the class.
read the comments below. Because this particular class deals with multiple instances of errors - assigning the properties of those errors to the object wouldn't be best practice as you will be overwriting them with each new error. However, it does make sense to store all of your errors and error properties within an array assigned to the object if you require to access historical data. For example, at the moment, if you create a new error - that is all you are doing. You have no way of accessing any old errors this object has created.
see above
You should also think about conflicts when assigning properties to objects. Are you likely to reassign, because if so, the old property will be gone. Fairly simple but still something you have to consider.
Related
We have PHP code in production that sometimes fails with "Call to member function on null", although the same code path executes fine several times before that in one invocation. We have a test that reproduces the error consistently at the same run of the loop.
I already proved that the object gets created correctly in the factory even if it gets returned as null. The factory method must not return null in any case, as indicated in the DocBlock. This question is not related to nullable return types or something like that.
The process does not exceed memory or runtime limitations and I already tried turning off the garbage collector, but no luck. The error happens both in PHP 7.0 and 7.3 on Debian, did not try on other versions or operating systems.
I am not allowed to paste the real code here, but I wrote a simple mockup to explain in more detail. Please keep in mind that this demo code will not result in the error, it is just meant to show the general structure of the program that runs into this fault.
// Three placeholder classes with common methods
class Bender
{
public function common()
{
echo "Bend, bend!" . PHP_EOL;
}
}
class Clamper
{
public function common()
{
echo "Clamp, clamp!" . PHP_EOL;
}
}
class Worker
{
public function common()
{
echo "Work, work!" . PHP_EOL;
}
}
// abstract class with static factory to produce objects
abstract class MomCorp
{
/**
* Factory to create one of several objects
*
* #param string $name
* #return Bender|Clamper|Worker
*/
public static function factory($name)
{
$type = self::managementDecision($name);
switch ($type)
{
case "bender":
$robot = new Bender();
break;
case "clamper":
$robot = new Clamper();
break;
default:
$robot = new Worker();
}
// optional QA works flawlessly here, object is fine all the time!
// $robot->common();
return $robot;
}
public static function managementDecision($name)
{
// irrelevant magic happens on $name here
return "bender";
}
}
foreach (['Rodriguez', 'Angle-ine', 'Flexo', 'Billie'] as $newName)
{
echo "$newName: ";
// The next two lines break after some loops - why?
// The perfectly functional object in factory gets returned as null
$bot = MomCorp::factory($newName);
$bot->common();
}
// SAMPLE OUTPUT
// Rodriguez: Bend, bend!
// Angle-ine: Bend, bend!
// Flexo: Bend, bend!
// Billie: Call to a member function common() on null
Has anyone experienced the same and has any hints on what might cause such an error and how to fix it?
Suppose I have the following PHP code:
class Foo {
function getBar() {
return 1;
}
}
function check( Foo $foo ) {
if ( $foo->getBar() == 1 ) {
// here could be more code ...
return 'Oh no, there was an error in class' .
get_class( $foo ) . ', method ' .
'getBar';
}
}
The last string in check bothers me because if Foo::bar gets renamed by a refactoring tool, the error message will be wrong. Is there any way to get around this without using a string somewhere?
You can use __METHOD__ to get the name of the current method.
But to get reference to other method that would allow you some kind of automatic refactoring - no, it's not possible in php.
Can be done by using method_exists()
class Foo {
function getBar() {
return 1;
}
}
function check( Foo $foo , $method = 'getBar') {
if (!method_exists($foo, $method) ) {
// here could be more code ...
return 'Oh no, there was an error in class' .
get_class( $foo ) . ', method ' .
$method;
}
}
It is not possible in PHP per se, but you can implement such a feature. One possible implementation would work as follows: somewhere the file path, class name, method name and some kind of a description of where and what should match what. Your new feature whenever triggered would check the given files, check whether some values changed, fix whatever needs to be fixed and log a report about the task. It would not be simple to implement something like this, but, important to note is that there is a solution.
I've got this autoloader method which is used to include class files as needed.
public static function autoloader($className) {
$fileExists = false;
foreach(array(LIBS_PATH, CONTROLLERS_PATH, MODELS_PATH) as $path) {
// Check if the class file exists and is readable
if(is_readable($classPath = $path . '/' . $className . '.php')) {
// Include the file
require($classPath);
$fileExists = true;
break;
}
}
if($fileExists === false) {
exit('The application cannot continue as it is not able to locate a required class.');
}
}
That works fine but I was thinking is this way better:
public static function autoloader($className) {
foreach(array(LIBS_PATH, CONTROLLERS_PATH, MODELS_PATH) as $path) {
// Check if the class file exists and is readable
if(is_readable($classPath = $path . '/' . $className . '.php')) {
// Include the file
require($classPath);
return;
}
}
exit('The application cannot continue as it is not able to locate a required class.');
}
As you can see I'm using the return statement in the middle of the loop to terminate the rest of the function because the class file has been included and the method has done its job.
Which way is the best way to break out of the loop if a match has been found? I'm just unsure about using the return statement because I always associated that with returning some value from a method.
Return can be used to simply break out of a method/function, and this use is perfectly legal. The real question to ask is what impact do you feel it has on readability?
There are different schools of thought on early returns.
One school maintains a single entry/exit point, stating that it makes the code easier to read since one can be assured that logic at the bottom of the code will be reached.
Another school states that the first school is outdated, and this is not as pressing a concern, especially if one maintains shorter method/function lengths.
A medium exists between the two where a trade-off between an early return and convoluted logic is present.
It comes down to your judgment.
No, it is not wrong to use return this way.
At least look at manual:
http://php.net/manual/en/function.autoload.php
void __autoload ( string $class )
void means that it should not return anything.
But it is not an error.
And also better use require_once when including class definitions.
Is it possible to define private variables in a PHP script so these variables are only visible in this single PHP script and nowhere else? I want to have an include file which does something without polluting the global namespace. It must work with PHP 5.2 so PHP namespaces are not an option. And no OOP is used here so I'm not searching for private class members. I'm searching for "somewhat-global" variables which are global in the current script but nowhere else.
In C I could do it with the static keyword but is there something similar in PHP?
Here is a short example of a "common.php" script:
$dir = dirname(__FILE__);
set_include_path($dir . PATH_SEPARATOR . get_include_path());
// Do more stuff with the $dir variable
When I include this file in some script then the $dir variable is visible in all other scripts as well and I don't want that. So how can I prevent this?
There are a few things you could do to keep $dir out of subsequent files
Example 1
set_include_path(dirname(__FILE__) . PATH_SEPARATOR . get_include_path());
This is the most obvious.
Example 2
$dir = dirname(__FILE__);
set_include_path($dir . PATH_SEPARATOR . get_include_path());
// work with $dir
unset($dir);
Just unset the variable after defining it and using it. Note this will unset any variable named $dir used prior to including this script.
Example 3
define('DIR_THIS', dirname(__FILE__));
set_include_path(DIR_THIS . PATH_SEPARATOR . get_include_path());
It is less likely I suppose to redefine a global constant like this.
Example 4
function my_set_include_path {
$dir = dirname(__FILE__);
set_include_path($dir . PATH_SEPARATOR . get_include_path());
// Do more stuff with the $dir variable
$my_other_var = 'is trapped within this function';
}
my_set_include_path();
You can define as many variables within that function and not affect the global namespace.
Conclusion
The first method is the easiest way to solve this problem, however because you want to use $dir again, it may not be ideal. The last example will at least keep that $dir (and any others defined in that function) out of the global namespace.
The only way you're going to accomplish anything close to what you want is to wrap everything in that included file in a function, and call it. If the file needs to execute itself you could still do
<?php
run_myfile()
function run_myfile() {
...
}
?>
There is no generic way to make a variable scoped to only a file outside of namespaces, classes, or functions.
Well, I'm probably getting flailed for this, but you if you are totally desperate you could use a Registry for that. I've whipped up a small one that does without classes (since I assume from And no OOP is used here so I'm not searching for private class members. means you don't want to do it with OOP at all)
function ®istry_get_instance()
{
static $data = array();
return $data;
}
The static $data variable inside is persisted inside the function scope, so you can call the function wherever you like and always get the same contents. The crucial point is returning by reference, e.g.
$registry = ®istry_get_instance(); // get $data array by reference
$registry['foo'] = 'bar'; // set something to $data
unset($registry); // delete global reference to $data
print_r(®istry_get_instance()); // show $data
Obviously you'd still have $registry as a variable in the global scope when calling this method from the global scope. So, you could add some more functions to make the Registry more convenient to use, e.g. for setting data to the Registry:
function registry_set($key, $value)
{
$registry = ®istry_get_instance();
$registry[$key] = $value;
}
and for getting it out again:
function registry_get($key)
{
$registry = ®istry_get_instance();
if(array_key_exists($key, $registry)) {
return $registry[$key];
} else {
trigger_error(sprintf(
'Undefined Index: %s', htmlentities($key)
), E_USER_NOTICE);
}
}
and for checking if a key exists:
function registry_isset($key)
{
$registry = ®istry_get_instance();
return array_key_exists($key, $registry);
}
which you could then use like:
registry_set('foo', 'bar'); // setting something to the registry
var_dump( registry_isset('foo') ); // check foo is in the registry now
echo registry_get('foo'); // prints 'bar'
echo registry_get('punt'); // raises Notice
You could populate the Registry from an include file with an additional method like this:
function registry_load_file($file)
{
if(!is_readable(realpath($file))) {
return trigger_error(sprintf(
'File is not readable: %s', htmlentities($file)
), E_USER_WARNING);
}
$config = include $file;
if(!is_array($config)) {
return trigger_error(sprintf(
'Expected file %s to return an array', htmlentities($file))
, E_USER_WARNING);
}
$registry = ®istry_get_instance();
$registry += $config;
}
with the include file having to return an array:
// config.php
return array(
'setting1' => 'something'
);
and then you can do
registry_load_from_file('config.php'); // add the contents of config to registry
print_r(registry_get_instance()); // show content of registry
Of course, this is now six functions in the global scope just for not having a global variable. Don't know if it's worth it, especially since I consider static in functions and all that reference stuff doubtful practice.
Take it as a proof of concept :)
Why not just put everything in a static class? Then you only have a single "variable" that could possibly conflict with the global namespace.
class MyClass {
public static $myvar = 1;
public static $myvar2 = "xyz";
public static function myfunction() {
self::$myvar++;
self::$myvar2 = "abc";
}
}
// References to class items, if needed
MyClass::myfunction();
MyClass::$myvar += 3;
If the problem you are trying to is just:
$dir = dirname(__FILE__);
set_include_path($dir . PATH_SEPARATOR . get_include_path());
// Do more stuff with the $dir variable
Then the solution would be to change the include path relative to '.' in your ini settings. E.g. change:
include_path=includes:/usr/local/php
to
include_path=./includes:/usr/local/php
Note that a script does not come into scope except where you explicitly include/require it (both the _once check applies globally) however I would recommend strongly against calling include/require from within a function - its much more transparent having the includes/requires at the top of the script.
I think that the problem you are trying to solve is based on a false premise and you should look for another way of fixing it. If you want the code in an include file to behave differently depending on what includes it, then really you should seperate it out into 2 seperate files - or maybe even 3 - 2 for the different behaviours and 1 for the common.
C.
I'm getting a "Call to a member function process_data() on a non-object in page.class.php on line 35" even though the object has been called.
Here is the index.php extraction showing the object being instantised
// require our common files
require("modules/module.php");
require("registry/objects/datetime.class.php");
require("registry/objects/page.class.php");
// load in all the objects
$datetime = new dateandtime;
$page = new page;
$module = new module;
It then passes to the Process class
require("template.class.php");
$template = new template($php_path . "controllers/themes/adm/" . $page . ".html");
// Place in both commonly used language and page specific language
$template->language($php_path . "controllers/language/en/adm/common.php");
$template->language($php_path . "controllers/language/en/adm/" . $page . ".php");
// Tell the page's module to load in data it needs
$module->process_data("module_" . $page);
// Output the final result
$template->output();
It's at this point PHP is throwing the error. The contents of the module.php file is as follows
class module {
public function process_data ($child) {
require($child . ".php");
read_data();
return true;
}
}
I've tried moving the instance declaration to within the second pasted code, but that generates more errors, because the class that "module" calls in then uses some of the "template" classes as well - so the same issue occurs just further down the line.
What am I getting wrong her, or completely missing, I'm sure it's the latter but I really need help here. Thanks
It looks to me as if variable $module was not in the scope when you try to call object method. Could you try var_dump($module) before $module->process_data("module_" . $page). What is the result of this function? Quick solution may be declaring $module global, but globals are not a very good idea anyway (but you may check if it works).