NAME Perinci::Sub::Wrapper - A multi-purpose subroutine wrapping framework VERSION version 0.41 SYNOPSIS use Perinci::Sub::Wrapper qw(wrap_sub); my $res = wrap_sub(sub => sub {die "test\n"}, meta=>{...}); my ($wrapped_sub, $meta) = ($res->[2]{sub}, $res->[2]{meta}); $wrapped_sub->(); # call the wrapped function DESCRIPTION Perinci::Sub::Wrapper is an extensible subroutine wrapping framework. It works by creating a single "large" wrapper function from a composite bits of code, instead of using multiple small wrappers (a la Python's decorator). The single-wrapper approach has the benefit of smaller function call overhead. You can still wrap multiple times if needed. This module is used to enforce Rinci properties, e.g. "args" (by performing schema validation before calling the function), "timeout" (by doing function call inside an "eval()" and using "alarm()" to limit the execution), or "retry" (by wrapping function call inside a simple retry loop). It can also be used to convert argument passing style, e.g. from "args_as" "array" to "args_as" "hash", so you can call function using named arguments even though the function accepts positional arguments, or vice versa. There are many other possible uses. This module uses Log::Any for logging. USAGE Suppose you have a subroutine like this: sub gen_random_array { my %args = @_; my $len = $args{len} // 10; die "Length too big" if $len > 1000; die "Please specify at least length=1" if $len < 1; [map {rand} 1..$len]; } Wrapping can, among others, validate arguments and give default values. First you add a Rinci metadata to your subroutine: our %SPEC; $SPEC{gen_random_array} = { v => 1.1, summary=> 'Generate an array of specified length containing random values', args => { len => {req=>1, schema => ["int*" => between => [1, 1000]]}, }, result_naked=>1, }; You can then remove code that validates arguments and gives default values. You might also want to make sure that your subroutine is run wrapped. sub gen_random_array { my %args = @_; die "This subroutine needs wrapping" unless $args{-wrapped}; # optional [map {rand} 1..$args{len}]; } Most wrapping options can also be put in "_perinci.sub.wrapper.*" attributes. For example: $SPEC{gen_random_array} = { v => 1.1, args => { len => {req=>1, schema => ["int*" => between => [1, 1000]]}, }, result_naked=>1, # skip validating arguments because sub already implements it "_perinci.sub.wrapper.validate_args" => 0, }; sub gen_random_array { my %args = @_; my $len = $args{len} // 10; die "Length too big" if $len > 1000; die "Please specify at least length=1" if $len < 1; [map {rand} 1..$len]; } See also Dist::Zilla::Plugin::Rinci::Validate which can insert validation code into your Perl source code files so you can skip doing it again in validation. EXTENDING The framework is simple and extensible. Please delve directly into the source code for now. Some notes: The internal uses OO. The main wrapper building mechanism is in the "wrap()" method. For each Rinci property, it will call "handle_NAME()" wrapper handler method. The "handlemeta_NAME()" methods are called first, to determine order of processing. You can supply these methods either by subclassing the class or, more simply, monkeypatching the method in the "Perinci::Sub::Wrapper" package. The wrapper handler method will be called with a hash argument, containing these keys: value (property value), new (this key will exist if "convert" argument of "wrap()" exists, to convert a property to a new value). For properties that have name in the form of "NAME1.NAME2.NAME3" (i.e., dotted) only the first part of the name will be used (i.e., "handle_NAME1()"). VARIABLES $Log_Wrapper_Code (BOOL) Whether to log wrapper result. Default is from environment variable LOG_PERINCI_WRAPPER_CODE, or false. Logging is done with Log::Any at trace level. METHODS The OO interface is only used internally or when you want to extend the wrapper. ENVIRONMENT LOG_PERINCI_WRAPPER_CODE (bool) If set to 1, will log the generated wrapper code. This value is used to set $Log_Wrapper_Code if it is not already set. PERINCI_WRAPPER_VALIDATE_ARGS (bool, default 1) Can be set to 0 to skip adding validation code. This provides a default for "validate_args" wrap_sub() argument. PERFORMANCE NOTES The following numbers are produced on an Asus Zenbook UX31 laptop (Intel Core i5 1.7GHz) using Perinci::Sub::Wrapper v0.33 and Perl v5.14.2. Operating system is Ubuntu 11.10 (64bit). For perspective, empty subroutine ("sub {}") as well as "sub { [200, "OK"] }" can be called around 4.3 mil/sec. Wrapping this subroutine "sub { [200, "OK"] }" and this simple metadata "{v=>1.1, args=>{a=>{schema=>"int"}}}" using default options yields call performance for "$sub->()" of about 0.28 mil/sec. For "$sub->(a=>1)" it is about 0.12 mil/sec. So if your sub needs to be called a million times a second, the wrapping adds too big of an overhead. By default, wrapper provides these functionality: checking invalid and unknown arguments, argument value validation, exception trapping ("eval {}"), and result checking. If we turn off all these features except argument validation (by adding options "allow_invalid_args=>1, trap=>0, validate_result=>0") call performance increases to around 0.47 mil/sec (for "$sub->()" and 0.24 mil/sec (for "$sub->(a=>1)"). As more arguments are introduced in the schema and passed, and as argument schemas become more complex, overhead will increase. For example, for 5 int arguments being declared and passed, call performance is around 0.11 mil/sec. Without passing any argument when calling, call performance is still around 0.43 mil/sec, indicating that the significant portion of the overhead is in argument validation. FAQ How to display the wrapper code being generated? If environment variable LOG_PERINCI_WRAPPER_CODE or package variable $Log_Perinci_Wrapper_Code is set to true, generated wrapper source code is logged at trace level using Log::Any. It can be displayed, for example, using Log::Any::App: % LOG_PERINCI_WRAPPER_CODE=1 TRACE=1 \ perl -MLog::Any::App -MPerinci::Sub::Wrapper=wrap_sub \ -e 'wrap_sub(sub=>sub{}, meta=>{v=>1.1, args=>{a=>{schema=>"int"}}});' How do I tell if I am being wrapped? Wrapper code passes "-wrapped" special argument with a true value. So you can do something like this: sub my_sub { my %args = @_; return [412, "I need to be wrapped"] unless $args{-wrapped}; ... } Your subroutine needs accept arguments as hash/hashref. caller() doesn't work from inside my wrapped code! Wrapping adds at least one or two levels of calls: one for the wrapper subroutine itself, the other is for the eval trap loop which can be disabled but is enabled by default. The 'goto &NAME' special form, which can replace subroutine and avoid adding another call level, cannot be used because wrapping also needs to postprocess function result. This poses a problem if you need to call caller() from within your wrapped code; it will also be off by at least one or two. The solution is for your function to use the caller() replacement, provided by Perinci::Sub::Util. But that is not transparent! True. The wrapped module needs to load and use that utility module explicitly. An alternative is for Perinci::Sub::Wrapper to use Sub::Uplevel. Currently though, this module does not use Sub::Uplevel because, as explained in its manpage, it is rather slow. If you don't use caller(), your subroutine actually doesn't need to care if it is wrapped nor it needs "uplevel-ing". SEE ALSO Perinci AUTHOR Steven Haryanto COPYRIGHT AND LICENSE This software is copyright (c) 2013 by Steven Haryanto. This is free software; you can redistribute it and/or modify it under the same terms as the Perl 5 programming language system itself. FUNCTIONS None are exported by default, but they are exportable. wrap_all_subs(%args) -> [status, msg, result, meta] Wrap all subroutines in a package and replace them with the wrapped version. This function will search all subroutines in a package which have metadata, wrap them, then replace the original subroutines and metadata with the wrapped version. One common use case is to put something like this at the bottom of your module: Perinci::Sub::Wrapper::wrap_all_subs(); to wrap ("protect") all your module's subroutines and discard the original unwrapped version. Arguments ('*' denotes required arguments): * package => *str* Package to search subroutines in. Default is caller package. * wrap_args => *hash* Arguments to pass to wrap_sub(). Each subroutine will be wrapped by wrap*sub(). This argument specifies what arguments to pass to wrap*sub(). Note: If you need different arguments for different subroutine, perhaps this function is not for you. You can perform your own loop and wrap_sub(). Return value: Returns an enveloped result (an array). First element (status) is an integer containing HTTP status code (200 means OK, 4xx caller error, 5xx function error). Second element (msg) is a string containing error message, or 'OK' if status is 200. Third element (result) is optional, the actual result. Fourth element (meta) is called result metadata and is optional, a hash that contains extra information. wrap_sub(%args) -> [status, msg, result, meta] Wrap subroutine to do various things, like enforcing Rinci properties. Will wrap subroutine and bless the generated wrapped subroutine (by default into "Perinci::Sub::Wrapped") as a way of marking that the subroutine is a wrapped one. Arguments ('*' denotes required arguments): * allow_invalid_args => *bool* (default: 0) Whether to allow invalid arguments. By default, wrapper will require that all argument names are valid ("/\A-?\w+\z/"), except when this option is turned on. * allow_unknown_args => *bool* (default: 0) Whether to allow unknown arguments. By default, this setting is set to false, which means that wrapper will require that all arguments are specified in "args" property, except for special arguments (those started with underscore), which will be allowed nevertheless. Will only be done if "allow_invalid_args" is set to false. * compile => *bool* (default: 1) Whether to compile the generated wrapper. Can be set to 0 to not actually wrap but just return the generated wrapper source code. * convert => *hash* Properties to convert to new value. Not all properties can be converted, but these are a partial list of those that can: v (usually do not need to be specified when converting from 1.0 to 1.1, will be done automatically), args*as, result*naked, default_lang. * debug => *bool* (default: 0) Generate code with debugging. If turned on, will produce various debugging in the generated code. Currently what this does: * add more comments (e.g. for each property handler) * forbid_tags => *array* Forbid properties which have certain wrapping tags. Some property wrapper, like dies*on*error (see Perinci::Sub::Property::dies*on*error) has tags 'die', to signify that it can cause wrapping code to die. Sometimes such properties are not desirable, e.g. in daemon environment. The use of such properties can be forbidden using this setting. * meta* => *hash* The function metadata. * normalize_schemas => *bool* (default: 1) Whether to normalize schemas in metadata. By default, wrapper normalize Sah schemas in metadata, like in 'args' or 'result' property, for convenience so that it does not need to be normalized again prior to use. If you want to turn off this behaviour, set to false. * remove_internal_properties => *bool* (default: 1) Whether to remove properties prefixed with _. By default, wrapper removes internal properties (properties which start with underscore) in the new metadata. Set this to false to keep them. * skip => *array* Properties to skip (treat as if they do not exist in metadata). * sub* => *code* The code to wrap. * sub_name => *str* The name of the code, e.g. Foo::func. It is a good idea to supply this so that wrapper code can display this information when they need to (e.g. see "Perinci::Sub::Property::dies_on_error"). * trap => *bool* (default: 1) Whether to trap exception using an eval block. If set to true, will wrap call using an eval {} block and return 500 /undef if function dies. Note that if some other properties requires an eval block (like 'timeout') an eval block will be added regardless of this parameter. * validate_args => *bool* (default: 1) Whether wrapper should validate arguments. If set to true, will validate arguments. Validation error will cause status 400 to be returned. This will only be done for arguments which has "schema" arg spec key. Will not be done if "args" property is skipped. * validate_result => *bool* (default: 1) Whether wrapper should validate arguments. If set to true, will validate sub's result. Validation error will cause wrapper to return status 500 instead of sub's result. This will only be done if "schema" or "statuses" keys are set in the "result" property. Will not be done if "result" property is skipped. Return value: Returns an enveloped result (an array). First element (status) is an integer containing HTTP status code (200 means OK, 4xx caller error, 5xx function error). Second element (msg) is a string containing error message, or 'OK' if status is 200. Third element (result) is optional, the actual result. Fourth element (meta) is called result metadata and is optional, a hash that contains extra information.