Skip to content

Advanced Usage

Greg Sherwood edited this page Aug 20, 2014 · 32 revisions

Specifying Valid File Extensions

By default, PHP_CodeSniffer will check any file it finds with a .inc, .php, .js or .css extension, although not all standards will actually check all these file types. Sometimes, this means that PHP_CodeSniffer is not checking enough of your files. Sometimes, the opposite is true. PHP_CodeSniffer allows you to specify a list of valid file extensions using the --extensions command line argument. Extensions are separated by commas.

To only check .php files:

$ phpcs --extensions=php /path/to/code

To check .php, .inc and .lib files:

$ phpcs --extensions=php,inc,lib /path/to/code

If you have asked PHP_CodeSniffer to check a specific file rather than an entire directory, the extension of the specified file will be ignored. The file will be checked even if it has an invalid extension or no extension at all. In the following example, the main.inc file will be checked by PHP_CodeSniffer even though the --extensions command line argument specifies that only .php files should be checked.

$ phpcs --extensions=php /path/to/code/main.inc

The ignoring of file extensions for specific files is a feature of PHP_CodeSniffer and is the only way to check files without an extension. If you check an entire directory of files, all files without extensions will be ignored, so you must check each of these file separately.

Ignoring Files and Folders

Sometimes you want PHP_CodeSniffer to run over a very large number of files, but you want some files and folders to be skipped. The --ignore command line argument can be used to tell PHP_CodeSniffer to skip files and folders that match one or more patterns.

In the following example, PHP_CodeSniffer will skip all files inside the package's tests and data directories. This is useful if you are checking a PEAR package but don't want your test or data files to conform to your coding standard.

$ phpcs --ignore=*/tests/*,*/data/* /path/to/code

The ignore patterns can also be complete regular expressions. If you do specify a regular expression, be aware that * is converted to .* for the convenience in simple patterns, like those used in the example above. So use * anywhere you would normally use .*.

You can also tell PHP_CodeSniffer to ignore a file using a special comment inserted at the top of the file. This will stop the file being checked even if it does not match the ignore pattern.

<?php
// @codingStandardsIgnoreFile
$xmlPackage = new XMLPackage;
$xmlPackage['error_code'] = get_default_error_code_value();
$xmlPackage->send();
?>

Ignoring Parts of a File

Some parts of your code may be unable to conform to your coding standard. For example, you might have to break your standard to integrate with an external library or web service. To stop PHP_CodeSniffer generating errors for this code, you can wrap it in special comments. PHP_CodeSniffer will then hide all errors and warnings that are generated for these lines of code.

$xmlPackage = new XMLPackage;
// @codingStandardsIgnoreStart
$xmlPackage['error_code'] = get_default_error_code_value();
// @codingStandardsIgnoreEnd
$xmlPackage->send();

Limiting Results to Specific Sniffs

By default, PHP_CodeSniffer will check your code using all sniffs in the specified standard. Sometimes you may want to find all occurrences of an error to eliminate it more quickly or understand the scope of the problem. PHP_CodeSniffer allows you to specify a list of sniffs to limit results to using the --sniffs command line argument. Sniffs are separated by commas.

$ phpcs --standard=PEAR --sniffs=Generic.PHP.LowerCaseConstant,PEAR.WhiteSpace.ScopeIndent /path/to/code

This feature is a message filter and not a quick way to define a custom coding standard. All sniffs specified on the command line in this way must be used in the coding standard you are using to check your files.

To view source codes for error messages, use the -s command line argument. This will print source codes in the full, summary and source reports.

Filtering Errors and Warnings Based on Severity

By default, PHP_CodeSniffer assigns a severity of 5 to all errors and warnings. Standards may change the severity of some messages so they are hidden by default or even so that they are raised to indicate greater importance. PHP_CodeSniffer allows you to decide what the minimum severity level must be to show a message in its report using the --severity command line argument.

To hide errors and warnings with a severity less than 3:

$ phpcs --severity=3 /path/to/code

You can specify different values for errors and warnings using the --error-severity and --warning-severity command line arguments.

To show all errors, but only warnings with a severity of 8 or more:

$ phpcs --error-severity=1 --warning-severity=8 /path/to/code

Setting the severity of warnings to 0 is the same as using the -n command line argument. If you set the severity of errors to 0 PHP_CodeSniffer will not show any errors, which may be useful if you just want to show warnings.

This feature is particularly useful during manual code reviews. During normal development, or an automated build, you may want to only check code formatting issues. But while during a code review, you may wish to show less severe errors and warnings that may need manual peer review.

Replacing Tabs with Spaces

Most of the sniffs written for PHP_CodeSniffer do not support the usage of tabs for indentation and alignment. You can write your own sniffs that check for tabs instead of spaces, but you can also get PHP_CodeSniffer to convert your tabs into spaces before a file is checked. This allows you to use the existing space-based sniffs on your tab-based files.

In the following example, PHP_CodeSniffer will replace all tabs in the files being checked with between 1 and 4 spaces, depending on the column the tab indents to.

$ phpcs --tab-width=4 /path/to/code

Specifying an Encoding

Some PHP_CodeSniffer reports output UTF-8 encoded XML, which can cause problems if your files are already UTF-8 encoded. In this case, some content from your files (generally comments) are used within error messages and may be double-encoded. To help PHP_CodeSniffer encode reports correctly, you can specify the encoding of your source files using the --encoding command line argument.

$ phpcs --encoding=utf-8 /path/to/code

The default encoding used by PHP_CodeSniffer is ISO-8859-1.

Specifying php.ini Settings

PHP_CodeSniffer allows you to set temporary php.ini settings during a run using the -d command line argument. The name of the php.ini setting must be specified on the command line, but the value is optional. If no value is set, the php.ini setting will be given a value of TRUE.

$ phpcs -d memory_limit=32M /path/to/code

You can also specific multiple values:

$ phpcs -d memory_limit=32M -d include_path=.:/php/includes /path/to/code

Setting Configuration Options

PHP_CodeSniffer has some configuration options that can be set. Individual coding standards may also require configuration options to be set before functionality can be used. View a full list of configuration options.

To set a configuration option, use the --config-set command line argument.

$ phpcs --config-set <option> <value>

Configuration options are written to a global configuration file. If you want to set them for a single run only, use the --runtime-set command line argument.

$ phpcs --runtime-set <option> <value> /path/to/code

Deleting Configuration Options

PHP_CodeSniffer allows you to delete any configuration option, reverting it to its default value. View a full list of configuration options.

To delete a configuration option, use the --config-delete command line argument.

$ phpcs --config-delete <option>

Viewing Configuration Options

To view the currently set configuration options, use the 1--config-show1 command line argument.

$ phpcs --config-show
Array
(
    [default_standard] => PEAR
    [zend_ca_path] => /path/to/ZendCodeAnalyzer
)

Printing Verbose Tokeniser Output

This feature is provided for debugging purposes only. Using this feature will dramatically increase screen output and script running time.

PHP_CodeSniffer contains multiple verbosity levels. Level 2 (indicated by the command line argument -vv) will print all verbosity information for level 1 (file specific token and line counts with running times) as well as verbose tokeniser output.

The output of the PHP_CodeSniffer tokeniser shows the step-by-step creation of the scope map and the level map.

The Scope Map

The scope map is best explained with an example. For the following file:

<?php
if ($condition) {
    echo 'Condition was true';
}
?>

The scope map output is:

*** START SCOPE MAP ***
Start scope map at 1: T_IF => if
Process token 2 []: T_WHITESPACE =>  
Process token 3 []: T_OPEN_PARENTHESIS => (
* skipping parenthesis *
Process token 6 []: T_WHITESPACE =>  
Process token 7 []: T_OPEN_CURLY_BRACKET => {
=> Found scope opener for 1 (T_IF)
Process token 8 [opener:7;]: T_WHITESPACE => \n
Process token 9 [opener:7;]: T_WHITESPACE =>     
Process token 10 [opener:7;]: T_ECHO => echo
Process token 11 [opener:7;]: T_WHITESPACE =>  
Process token 12 [opener:7;]: T_CONSTANT_ENCAPSED_STRING => 'Condition was true'
Process token 13 [opener:7;]: T_SEMICOLON => ;
Process token 14 [opener:7;]: T_WHITESPACE => \n
Process token 15 [opener:7;]: T_CLOSE_CURLY_BRACKET => }
=> Found scope closer for 1 (T_IF)
*** END SCOPE MAP ***

The scope map output above shows the following pieces of information about the file:

  • A scope token if was found at token 1 (note that token 0 is the open PHP tag).
  • The opener for the if statement, the open curly brace, was found at token 7.
  • The closer for the if statement, the close curly brace, was found at token 15.
  • Tokens 8 - 15 are all included in the scope set by the scope opener at token 7, the open curly brace. This indicates that these tokens are all within the if statement.

The scope map output is most useful when debugging PHP_CodeSniffer's scope map, which is critically important to the successful checking of a file, but is also useful for checking the type of a particular token. For example, if you are unsure of the token type for an opening curly brace, the scope map output shows you that the type is T_OPEN_CURLY_BRACKET and not, for example, T_OPEN_CURLY_BRACE.

The Level Map

The level map is best explained with an example. For the following file:

<?php
if ($condition) {
    echo 'Condition was true';
}
?>

The level map output is:

*** START LEVEL MAP ***
Process token 0 on line 1 [lvl:0;]: T_OPEN_TAG => <?php\n
Process token 1 on line 2 [lvl:0;]: T_IF => if
Process token 2 on line 2 [lvl:0;]: T_WHITESPACE =>  
Process token 3 on line 2 [lvl:0;]: T_OPEN_PARENTHESIS => (
Process token 4 on line 2 [lvl:0;]: T_VARIABLE => $condition
Process token 5 on line 2 [lvl:0;]: T_CLOSE_PARENTHESIS => )
Process token 6 on line 2 [lvl:0;]: T_WHITESPACE =>  
Process token 7 on line 2 [lvl:0;]: T_OPEN_CURLY_BRACKET => {
=> Found scope opener for 1 (T_IF)
    * level increased *
    * token 1 (T_IF) added to conditions array *
    Process token 8 on line 2 [lvl:1;conds;T_IF;]: T_WHITESPACE => \n
    Process token 9 on line 3 [lvl:1;conds;T_IF;]: T_WHITESPACE =>     
    Process token 10 on line 3 [lvl:1;conds;T_IF;]: T_ECHO => echo
    Process token 11 on line 3 [lvl:1;conds;T_IF;]: T_WHITESPACE =>  
    Process token 12 on line 3 [lvl:1;conds;T_IF;]: T_CONSTANT_ENCAPSED_STRING => 'Condition was true'
    Process token 13 on line 3 [lvl:1;conds;T_IF;]: T_SEMICOLON => ;
    Process token 14 on line 3 [lvl:1;conds;T_IF;]: T_WHITESPACE => \n
    Process token 15 on line 4 [lvl:1;conds;T_IF;]: T_CLOSE_CURLY_BRACKET => }
    => Found scope closer for 7 (T_OPEN_CURLY_BRACKET)
    * token T_IF removed from conditions array *
    * level decreased *
Process token 16 on line 4 [lvl:0;]: T_WHITESPACE => \n
Process token 17 on line 5 [lvl:0;]: T_CLOSE_TAG => ?>\n
*** END LEVEL MAP ***

The level map output above shows the following pieces of information about the file:

  • A scope opener, an open curly brace, was found at token 7 and opened the scope for an if statement, defined at token 1.
  • Tokens 8 - 15 are all included in the scope set by the scope opener at token 7, the open curly brace. All these tokens are at level 1, indicating that they are enclosed in 1 scope condition, and all these tokens are enclosed in a single condition; an if statement.

The level map is most commonly used to determine indentation rules (e.g., a token 4 levels deep requires 16 spaces of indentation) or to determine if a particular token is within a particular scope (eg. a function keyword is within a class scope, making it a method).

Printing Verbose Token Processing Output

This feature is provided for debugging purposes only. Using this feature will dramatically increase screen output and script running time.

PHP_CodeSniffer contains multiple verbosity levels. Level 3 (indicated by the command line argument -vvv) will print all verbosity information for level 1 (file specific token and line counts with running times), level 2 (tokeniser output) as well as token processing output with sniff running times.

The token processing output is best explained with an example. For the following file:

<?php
if ($condition) {
    echo 'Condition was true';
}
?>

The token processing output is:

*** START TOKEN PROCESSING ***
Process token 0: T_OPEN_TAG => <?php\n
    Processing PEAR_Sniffs_Commenting_FileCommentSniff... DONE in 0 seconds
    Processing Generic_Sniffs_PHP_DisallowShortOpenTagSniff... DONE in 0 seconds
    Processing Generic_Sniffs_Files_LineLengthSniff... DONE in 0.0001 seconds
    Processing Generic_Sniffs_Files_LineEndingsSniff... DONE in 0 seconds
Process token 1: T_IF => if
    Processing PEAR_Sniffs_ControlStructures_ControlSignatureSniff... DONE in 0.0001 seconds
    Processing PEAR_Sniffs_ControlStructures_MultiLineConditionSniff... DONE in 0 seconds
    Processing PEAR_Sniffs_WhiteSpace_ScopeClosingBraceSniff... DONE in 0 seconds
    Processing PEAR_Sniffs_WhiteSpace_ScopeIndentSniff... DONE in 0 seconds
    Processing Generic_Sniffs_ControlStructures_InlineControlStructureSniff... DONE in 0 seconds
Process token 2: T_WHITESPACE =>  
    Processing Generic_Sniffs_WhiteSpace_DisallowTabIndentSniff... DONE in 0 seconds
Process token 3: T_OPEN_PARENTHESIS => (
Process token 4: T_VARIABLE => $condition
    Processing PEAR_Sniffs_NamingConventions_ValidVariableNameSniff... DONE in 0 seconds
Process token 5: T_CLOSE_PARENTHESIS => )
Process token 6: T_WHITESPACE =>  
    Processing Generic_Sniffs_WhiteSpace_DisallowTabIndentSniff... DONE in 0 seconds
Process token 7: T_OPEN_CURLY_BRACKET => {
Process token 8: T_WHITESPACE => \n
    Processing Generic_Sniffs_WhiteSpace_DisallowTabIndentSniff... DONE in 0 seconds
Process token 9: T_WHITESPACE =>     
    Processing Generic_Sniffs_WhiteSpace_DisallowTabIndentSniff... DONE in 0 seconds
Process token 10: T_ECHO => echo
Process token 11: T_WHITESPACE =>  
    Processing Generic_Sniffs_WhiteSpace_DisallowTabIndentSniff... DONE in 0 seconds
Process token 12: T_CONSTANT_ENCAPSED_STRING => 'Condition was true'
Process token 13: T_SEMICOLON => ;
Process token 14: T_WHITESPACE => \n
    Processing Generic_Sniffs_WhiteSpace_DisallowTabIndentSniff... DONE in 0 seconds
Process token 15: T_CLOSE_CURLY_BRACKET => }
Process token 16: T_WHITESPACE => \n
    Processing Generic_Sniffs_WhiteSpace_DisallowTabIndentSniff... DONE in 0 seconds
Process token 17: T_CLOSE_TAG => ?>\n
*** END TOKEN PROCESSING ***

Every token processed is shown, along with its ID, type and contents. For each token, all sniffs that were executed on the token are displayed, along with the running time.

For example, the output above shows us that token 1, an if keyword, had 5 sniffs executed on it; the ControlSignature sniff, the MultiLineCondition sniff, the ScopeClosingBrace sniff, the ScopeIndent sniff and the InlineControlStructure sniff. Each was executed fairly quickly, but the slowest was the ControlSignature sniff, taking 0.0001 seconds to process that token.

The other interesting piece of information we get from the output above is that some tokens didn't have any sniffs executed on them. This is normal behaviour for PHP_CodeSniffer as most sniffs listen for specific or rarely used tokens and then execute on it and a number of tokens following it.

For example, the ScopeIndentSniff executes on the if statement's token only, but actually checks the indentation of every line within the if statement. The sniff uses the scope map to find all tokens within the if statement.