|
NAMEBadger::Debug - base class mixin module implement debugging methodsSYNOPSISpackage Your::Module; use Badger::Debug default => 0; # default value for $DEBUG and DEBUG sub some_method { my $self = shift; # DEBUG is a compile-time constant, so very efficient $self->debug("First Message") if DEBUG; # $DEBUG is a runtime variable, so more flexible $self->debug("Second Message") if $DEBUG; } package main; use Your::Module; Your::Module->some_method; # no output, debugging off by default Your::Module->debugging(1); # turns runtime debugging on Your::Module->some_method; # [Your::Module line 13] Second Message DESCRIPTIONThis mixin module implements a number of methods for debugging. Read "The Whole Caboodle" if you just want to get started quickly. Read "Picky Picky Picky" if you want to get all picky about what you want to use or want more information on the individual features.Note that all of the debugging methods described below work equally well as both object and class methods even if we don't explicitly show them being used both ways. # class method Your::Module->debug('called as a class method'); # object method my $object = Your::Module->new; $object->debug('called as an object method'); The Whole CaboodleThe default import option is the all-in-one option that enables all debugging features. The value you specify with it will be used as the default debugging status. Use 0 if you want debugging off by default, or any true value if you want it on.package Your::Module; use Badger::Debug default => 0; The default option imports the debug() and debugging() methods, the $DEBUG package variable (set to the default value you specified unless it's already defined to be something else), and the DEBUG constant subroutine (defined to have the same value as the $DEBUG variable). In your module's methods you can call the debug() method to generate debugging messages. You can use the DEBUG constant or the $DEBUG variable as a condition so that messages only get displayed when debugging is enbled. sub some_method { my $self = shift; # DEBUG is a compile-time constant, so very efficient $self->debug("First Message") if DEBUG; # $DEBUG is a runtime variable, so more flexible $self->debug("Second Message") if $DEBUG; } The DEBUG constant is resolved at compile time so it results in more efficient code. When debugging is off, Perl will completely eliminate the first call to the debug() method in the above example. The end result is that there's no performance overhead incurred by including debugging statements like these. The $DEBUG package variable is a little more flexible because you can change the value at any point during the execution of your program. You might want to do this from inside the module (say to enable debugging in one particular method that's causing problems), or outside the module from a calling program or another module. The debugging() method is provided as a convenient way to change the $DEBUG package variable for a module. Your::Module->debugging(0); # turn runtime debugging off Your::Module->debugging(1); # turn runtime debugging on The downside is that checking the $DEBUG variable at runtime is less efficient than using the DEBUG compile time constant. Unless you're working on performance critical code, it's probably not something that you should worry about. However, if you are the worrying type then you can use "Badger::Debug" to get some of the best bits of both worlds. When your module is loaded, both DEBUG and $DEBUG will be set to the default value you specified unless $DEBUG is already defined. If it is defined then the DEBUG constant will be set to whatever value it has. So if you define the $DEBUG package variable before loading the module then you'll be able to enable both run time and compile time debugging messages without having to go and edit the source code of your module. $Your::Module::DEBUG = 1; require Your::Module; Alternately, you can let "Badger::Debug" do it for you. The modules import option allows you to specify one or more modules that you want debugging enabled for. use Badger::Debug modules => 'My::Module::One My::Module::Two'; use My::Module::One; # both runtime and compile time use My::Module::Two; # debugging enabled in both modules The benefit of this approach is that it happens at compile time. If you do it before you "use" your modules, then you'll get both compile time and run time debugging enabled. If you do it after then you'll get just runtime debugging enabled. Best of all - you don't need to change any of your existing code to load modules via "require" instead of "use" Picky Picky PickyThe "Badger::Debug" module allow you to be more selective about what you want to use. This section described the individual debugging methods and the DEBUG and $DEBUG flags that can be used to control debugging.In the simplest case, you can import the debug() method into your own module for generating debugging messages. package Your::Module; use Badger::Debug 'debug'; sub some_method { my $self = shift; $self->debug("Hello from some_method()"); } In most cases you'll want to be able to turn debugging messages on and off. You could do something like this: # initialise $DEBUG if it's not already set our $DEBUG = 0 unless defined $DEBUG; sub some_method { my $self = shift; $self->debug("Hello from some_method()") if $DEBUG; } If you use the "unless defined $DEBUG" idiom shown in the example shown above then it will also allow you to set the $DEBUG flag before your module is loaded. This is particularly useful if the module is auto-loaded on demand by another module or your own code. # set $DEBUG flag for your module $Your::Module::DEBUG = 1; # later... require Your::Module; # debugging is enabled You can also achieve the same effect at compile time using the "Badger::Debug" modules export option. use Badger::Debug modules => 'Your::Module'; # sets $Your::Module::DEBUG = 1 use Your::Module; # debugging is enabled The advantage of using the $DEBUG package variable is that you can change the value at any point to turn debugging on or off. For example, if you've got a section of code that requires debugging enabled to track down a particular bug then you can write something like this: sub gnarly_method { my $self = shift; local $DEBUG = 1; $self->debug("Trying to track down the cause bug 666"); # the rest of your code... $self->some_method; } Making the change to $DEBUG "local" means that it'll only stay set to 1 until the end of the "gnarly_method()". It's a good idea to add a debugging message any time you make temporary changes like this. The message generated will contain the file and line number so that you can easily find it later when the bug has been squashed and either comment it out (for next time) or remove it. The "Badger::Debug" module has a $DEBUG export hook which will define the the $DEBUG variable for you. The value you provide will be used as the default for $DEBUG if it isn't already defined. package Your::Module; use Badger::Debug 'debug', '$DEBUG' => 0; sub some_method { my $self = shift; $self->debug("Hello from some_method()") if $DEBUG; } The debugging() method can also be imported from "Badger::Debug". This provides a simple way to set the $DEBUG variable. Your::Module->debugging(1); # debugging on Your::Module->debugging(0); # debugging off The downside to using a package variable is that it slows your code down every time you check the $DEBUG flag. In all but the most extreme cases, this should be of no concern to you whatsoever. Write your code in the way that is most convenient for you, not the machine. WARNING: Do not even begin to consider entertaining the merest thought of optimising your code to make it run faster until your company is on the verge of financial ruin due to your poorly performing application and your boss has told you (with confirmation in writing, countersigned by at least 3 members of the board of directors) that you will be fired first thing tomorrow morning unless you make the code run faster RIGHT NOW. Another approach is to define a constant DEBUG value. package Your::Module; use Badger::Debug 'debug'; use constant DEBUG => 0; sub some_method { my $self = shift; $self->debug("Hello from some_method()") if DEBUG; } This is an all-or-nothing approach. Debugging is on or off and there's nothing you can do about it except for changing the constant definition in the source code and running the program again. The benefit of this approach is that DEBUG is defined as a compile time constant. When DEBUG is set to 0, Perl will effectively remove the entire debugging line at compile time because it's based on a premise ("if DEBUG") that is known to be false. The end result is that there's no runtime performance penalty whatsoever. "Badger::Debug" also provides the DEBUG hook if this is the kind of thing you want. package Your::Module; use Badger::Debug 'debug', 'DEBUG' => 0; sub some_method { my $self = shift; $self->debug("Hello from some_method()") if DEBUG; } What makes this extra-special is that you're only specifying the default value for the "DEBUG" constant. If the $DEBUG package variable is defined when the module is loaded then that value will be used instead. So although it's not possible to enable or disable debugging for different parts of a module, you can still enable debugging for the whole module by setting the $DEBUG package variable before loading it. # set $DEBUG flag for your module $Your::Module::DEBUG = 1; # later... require Your::Module; # debugging is enabled Here's a reminder of the other way to achieve the same thing at compile time using the "Badger::Debug" modules export option. use Badger::Debug modules => 'Your::Module'; # sets $Your::Module::DEBUG = 1 use Your::Module; # debugging is enabled You can combine the use of both $DEBUG and DEBUG in your code, for a two-level approach to debugging. The DEBUG tests will always be resolved at compile time so they're suitable for low-level debugging that either has a performance impact or is rarely required. The $DEBUG tests will be resolved at run time, so they can be enabled or disabled at any time or place. sub some_method { my $self = shift; $self->debug("Hello from some_method()") if DEBUG; $self->debug("Goodbye from some_method()") if $DEBUG; } IMPORT OPTIONSAll of the debugging methods can be imported selectively into your module. For example:use Badger::Debug 'debug debugging debug_caller'; The following import options are also provided. defaultUsed to set the default debugging value and import various debugging methods and flags.use Badger::Debug default => 0; # debugging off by default It imports the debug() and debugging() methods along with the $DEBUG package variable and DEBUG constant. See "The Whole Caboodle" for further discussion on using it. $DEBUGUsed to define a $DEBUG variable in your module. A default value should be specified which will be used to set the $DEBUG value if it isn't already defined.use Badger::Debug '$DEBUG' => 0; # debugging off by default print $DEBUG; # 0 DEBUGUsed to define a "DEBUG" constant in your module. If the $DEBUG package variable is defined then the "DEBUG" constant will be set to whatever value it contains. Otherwise it will be set to the default value you provide.use Badger::Debug 'DEBUG' => 0; # debugging off by default print DEBUG; # 0 modulesThis option can be used to set the $DEBUG value true in one or more packages. This ensures that any debugging will be enabled in those modules.use Badger::Debug modules => 'My::Module::One My::Module::Two'; use My::Module::One; # debugging enabled in both modules use My::Module::Two; Modules that haven't yet been loaded will have both compile time (DEBUG) and run time ($DEBUG) debugging enabled. Modules that have already been loaded will only have run time debugging enabled. dumpsThis option can be used to construct a specialised dump() method for your module. The method is used to display nested data in serialised text form for debugging purposes. The default dump() method for an object will display all items stored within the object. The "dumps" import option can be used to limit the dump to only display the fields specified.package Your::Module; use Badger::Debug dumps => 'foo bar baz'; # ...more code... package main; my $object = Your::Module->new; print $object->dump; # dumps foo, bar and baz colour / colorEither of these (depending on your spelling preference) can be used to enable colourful (or colorful) debugging.use Badger::Debug 'colour'; Debugging messages will then appear in colour (on a terminal supporting ANSI escape sequences). See the Badger::Test module for an example of this in use. :debugImports all of the debug(), debugging(), debug_up(), debug_caller(), debug_callers and debug_args() methods.:dumpImports all of the dump(), dump_ref(), dump_hash(), dump_list(), dump_text(), dump_data() and dump_data_inline() methods.DEBUGGING METHODSdebug($msg1, $msg2, ...)This method can be used to generate debugging messages.$object->debug("Hello ", "World\n"); It prints all argument to STDERR with a prefix indicating the class name, file name and line number from where the "debug()" method was called. [Badger::Example line 42] Hello World At some point in the future this will be extended to allow you to tie in debug hooks, e.g. to forward to a logging module. debugf($format, $arg1, $arg2, ...)This method provides a "printf()"-like wrapper around debug().$object->debugf('%s is %s', e => 2.718); # e is 2.718 debug_up($n, $msg1, $msg2, ...)The debug() method generates a message showing the file and line number from where the method was called. The "debug_up()" method can be used to report the error from somewhere higher up the call stack. This is typically used when you create your own debugging methods, as shown in the following example.sub parse { my $self = shift; while (my ($foo, $bar) = $self->get_foo_bar) { $self->trace($foo, $bar); # report line here # do something } } sub trace { my ($self, $foo, $bar) = @_; $self->debug_up(2, "foo: $foo bar: $bar"); # not here } The "trace()" method calls the debug_up() method telling it to look two levels up in the caller stack instead of the usual one (thus "debug_up(1,...)" has the same effect as "debug(...)"). So instead of reporting the line number in the "trace()" subroutine (which would be the case if we called "debug(...)" or "debug_up(1,...)"), it will correctly reporting the line number of the call to "trace()" in the "parse()" method. debug_at($info, $message)This method is a wrapper around debug() that allows you to specify a different location to be added to the message generated.$at->debug_at( { where => 'At the edge of time', line => 420 }, 'Flying sideways' ); This generates the following debug message: [At the edge of time line 420] Flying sideways Far out, man! You can change the $FORMAT package variable to define a different message structure. As well as the pre-defined placeholders (see the $FORMAT documentation) you can also define your own custom placeholders like "<server>" in the following example. $Badger::Debug::FORMAT = '<server>: <msg> at line <line> of <file>'; You must then provide values for the additional placeholder in the $info hash array when you call the debug_at() method. $at->debug_at( { server => 'Alpha' }, 'Normality is resumed' ); You can also specify a custom format in the $info hash array. $at->debug_at( { format => '<msg> at line <line> of <file>' }, 'Normality is resumed' ); debug_caller()Prints debugging information about the current caller.sub wibble { my $self = shift; $self->debug_caller; } debug_callers()Prints debugging information about the complete call stack.sub wibble { my $self = shift; $self->debug_callers; } debug_args()Prints debugging information about the arguments passed.sub wibble { my $self = shift; $self->debug_args(@_); } debugging($flag)This method of convenience can be used to set the $DEBUG variable for a module. It can be called as a class or object method.Your::Module->debugging(1); # turn debugging on Your::Module->debugging(0); # turn debugging off debug_modules(@modules)This method can be used to set the $DEBUG true in one or more modules. Modules can be specified as a list of package names, a reference to a list, or a whitespace delimited string.Badger::Debug->debug_modules('Your::Module::One Your::Module::Two'); The method is also accessible via the modules import option. DATA INSPECTION METHODSThese methods of convenience can be used to inspect data structures. The emphasis is on brevity for the sake of debugging rather than full blown inspection. Use Data::Dumper or on of the other fine modules available from CPAN if you want something more thorough.The methods below are recursive, so dump_list(), on finding a hash reference in the list will call dump_hash() and so on. However, this recursion is deliberately limited to no more than $MAX_DEPTH levels deep (3 by default). Remember, the emphasis here is on being able to see enough of the data you're dealing with, neatly formatted for debugging purposes, rather than being overwhelmed with the big picture. If any of the methods encounter an object then they will call its dump() method if it has one. Otherwise they fall back on dump_ref() to expose the internals of the underlying data type. You can create your own custom dump() method for you objects or use the dumps import option to have a custom dump() method defined for you. dump()Debugging method which returns a text representation of the object internals.print STDERR $object->dump(); You can define your own "dump()" for an object and this will be called whenever your object is dumped. The dumps import option can be used to generate a custom "dump()" method. dump_ref($ref)Does The Right Thing to call the appropriate dump method for a reference of some kind.dump_hash(\%hash)Debugging method which returns a text representation of the hash array passed by reference as the first argument.print STDERR $object->dump_hash(\%hash); dump_list(\@list)Debugging method which returns a text representation of the array passed by reference as the first argument.print STDERR $object->dump_list(\@list); dump_text($text)Debugging method which returns a truncated and sanitised representation of the text string passed (directly or by reference) as the first argument.print STDERR $object->dump_text($text); The string will be truncated to $MAX_TEXT characters and any newlines will be converted to "\n" representations. dump_data($item)Debugging method which calls the appropriate dump method for the item passed as the first argument. If it is an object with a dump() method then that will be called, otherwise it will fall back on dump_ref(), as it will for any other non-object references. Non-references are passed to the dump_text() method.print STDERR $object->dump_data($item); dump_data_inline($item)Wrapper around dump_data() which strips any newlines from the generated output, suitable for a more compact debugging output.print STDERR $object->dump_data_inline($item); MISCELLANEOUS METHODSenable_colour()Enables colourful debugging and error messages.Badger::Debug->enable_colour; PACKAGE VARIABLES$FORMATThe debug() method uses the message format in the $FORMAT package variable to generate debugging messages. The default value is:[<where> line <line>] <msg> The "<where<gt">, "<line>" and "<msg>" markers denote the positions where the class name, line number and debugging message are inserted. You can embed any of the following placeholders into the message format: msg The debugging message file The name of the file where the debug() method was called from line The line number that it was called from pkg The package that it was called from class The class name of the object that the method was called against where A summary of the package and class date The current date time The current time If the "class" is the same as the "pkg" then "where" will contain the same value. If they are different then "where" will be set equivalent to "<pkg> (<class>)". This is the case when the debug() method is called from a base class method ("pkg" will be the base class name from where the call was made) against a subclass object ("class" will be the subclass name). See also the debug_at() method which allows you to specify a custom format and/or additional placeholder values. $MAX_DEPTHThe maximum depth that the data inspection methods will recurse to.$MAX_TEXTThe maximum length of text that will be returned by dump_text().AUTHORAndy Wardley <http://wardley.org/>COPYRIGHTCopyright (C) 1996-2009 Andy Wardley. All Rights Reserved.This module is free software; you can redistribute it and/or modify it under the same terms as Perl itself.
Visit the GSP FreeBSD Man Page Interface. |