← Index
NYTProf Performance Profile   « block view • line view • sub view »
For 05.Domain_and_Item.t
  Run on Tue May 4 17:21:41 2010
Reported on Tue May 4 17:23:13 2010

File /usr/local/lib/perl5/site_perl/5.10.1/Try/Tiny.pm
Statements Executed 11487
Statement Execution Time 7.55ms
Subroutines — ordered by exclusive time
Calls P F Exclusive
Time
Inclusive
Time
Subroutine
522556.29ms655msTry::Tiny::::try Try::Tiny::try
521441.68ms1.68msTry::Tiny::::catch Try::Tiny::catch
11119µs27µsTry::Tiny::::BEGIN@3 Try::Tiny::BEGIN@3
11114µs14µsTry::Tiny::::BEGIN@8 Try::Tiny::BEGIN@8
1119µs80µsTry::Tiny::::BEGIN@6 Try::Tiny::BEGIN@6
1119µs51µsTry::Tiny::::BEGIN@46 Try::Tiny::BEGIN@46
0000s0sTry::Tiny::ScopeGuard::::DESTROYTry::Tiny::ScopeGuard::DESTROY
0000s0sTry::Tiny::::finally Try::Tiny::finally
Call graph for these subroutines as a Graphviz dot language file.
Line State
ments
Time
on line
Calls Time
in subs
Code
1package Try::Tiny;
2
3332µs235µs
# spent 27µs (19+8) within Try::Tiny::BEGIN@3 which was called # once (19µs+8µs) by Class::MOP::BEGIN@13 at line 3
use strict;
# spent 27µs making 1 call to Try::Tiny::BEGIN@3 # spent 8µs making 1 call to strict::import
4#use warnings;
5
6344µs2152µs
# spent 80µs (9+72) within Try::Tiny::BEGIN@6 which was called # once (9µs+72µs) by Class::MOP::BEGIN@13 at line 6
use vars qw(@EXPORT @EXPORT_OK $VERSION @ISA);
# spent 80µs making 1 call to Try::Tiny::BEGIN@6 # spent 72µs making 1 call to vars::import
7
8
# spent 14µs within Try::Tiny::BEGIN@8 which was called # once (14µs+0s) by Class::MOP::BEGIN@13 at line 11
BEGIN {
91600ns require Exporter;
10113µs @ISA = qw(Exporter);
111109µs114µs}
# spent 14µs making 1 call to Try::Tiny::BEGIN@8
12
131800ns$VERSION = "0.04";
14
15122µs$VERSION = eval $VERSION;
16
1712µs@EXPORT = @EXPORT_OK = qw(try catch finally);
18
1911µs$Carp::Internal{+__PACKAGE__}++;
20
21# Need to prototype as @ not $$ because of the way Perl evaluates the prototype.
22# Keeping it at $$ means you only ever get 1 sub because we need to eval in a list
23# context & not a scalar one
24
25
# spent 655ms (6.29+649) within Try::Tiny::try which was called 522 times, avg 1.25ms/call: # 276 times (3.13ms+15.5ms) by Class::MOP::Attribute::_process_accessors at line 343 of Class/MOP/Attribute.pm, avg 67µs/call # 201 times (2.31ms+35.1ms) by Class::MOP::Class::_post_add_attribute at line 519 of Class/MOP/Class.pm, avg 186µs/call # 36 times (676µs+748µs) by Moose::Object::DESTROY at line 100 of Moose/Object.pm, avg 40µs/call # 8 times (143µs+597ms) by Class::MOP::load_first_existing_class at line 110 of Class/MOP.pm, avg 74.7ms/call # once (26µs+-26µs) by Moose::Meta::Attribute::does at line 35 of Moose/Meta/Attribute.pm
sub try (&;@) {
26522285µs my ( $try, @code_refs ) = @_;
27
28 # we need to save this here, the eval block will be in scalar context due
29 # to $failed
3052297µs my $wantarray = wantarray;
31
32522107µs my ( $catch, $finally );
33
34 # find labeled blocks in the argument list.
35 # catch and finally tag the blocks by blessing a scalar reference to them.
36522333µs foreach my $code_ref (@code_refs) {
3752154µs next unless $code_ref;
38
39521176µs my $ref = ref($code_ref);
40
41521518µs if ( $ref eq 'Try::Tiny::Catch' ) {
42 $catch = ${$code_ref};
43 } elsif ( $ref eq 'Try::Tiny::Finally' ) {
44 $finally = ${$code_ref};
45 } else {
463297µs293µs
# spent 51µs (9+42) within Try::Tiny::BEGIN@46 which was called # once (9µs+42µs) by Class::MOP::BEGIN@13 at line 46
use Carp;
# spent 51µs making 1 call to Try::Tiny::BEGIN@46 # spent 42µs making 1 call to Exporter::import
47 confess("Unknown code ref type given '${ref}'. Check your usage & try again");
48 }
49 }
50
51 # save the value of $@ so we can set $@ back to it in the beginning of the eval
52522104µs my $prev_error = $@;
53
5452291µs my ( @ret, $error, $failed );
55
56 # FIXME consider using local $SIG{__DIE__} to accumulate all errors. It's
57 # not perfect, but we could provide a list of additional errors for
58 # $catch->();
59
60 {
61 # localize $@ to prevent clobbering of previous value by a successful
62 # eval.
631044147µs local $@;
64
65 # failed will be true if the eval dies, because 1 will not be returned
66 # from the eval body
67522318µs $failed = not eval {
6852259µs $@ = $prev_error;
69
70 # evaluate the try block in the correct context
71522244µs9664ms if ( $wantarray ) {
# spent 664ms making 8 calls to Class::MOP::__ANON__[Class/MOP.pm:103], avg 83.0ms/call # spent 11µs making 1 call to Moose::Meta::Attribute::__ANON__[Moose/Meta/Attribute.pm:35]
72 @ret = $try->();
73 } elsif ( defined $wantarray ) {
74 $ret[0] = $try->();
75 } else {
76513563µs513186ms $try->();
# spent 105ms making 201 calls to Class::MOP::Class::__ANON__[Class/MOP/Class.pm:515], avg 520µs/call # spent 80.3ms making 276 calls to Class::MOP::Attribute::__ANON__[Class/MOP/Attribute.pm:340], avg 291µs/call # spent 748µs making 36 calls to Moose::Object::__ANON__[Moose/Object.pm:94], avg 21µs/call
77 };
78
79521174µs return 1; # properly set $fail to false
80 };
81
82 # copy $@ to $error; when we leave this scope, local $@ will revert $@
83 # back to its previous value
84522263µs $error = $@;
85 }
86
87 # set up a scope guard to invoke the finally block at the end
88522116µs my $guard = $finally && bless \$finally, "Try::Tiny::ScopeGuard";
89
90 # at this point $failed contains a true value if the eval died, even if some
91 # destructor overwrote $@ as the eval was unwinding.
92522107µs if ( $failed ) {
93 # if we got an error, invoke the catch block.
9411µs if ( $catch ) {
95 # This works like given($error), but is backwards compatible and
96 # sets $_ in the dynamic scope for the body of C<$catch>
97 for ($error) {
9815µs129µs return $catch->($error);
# spent 29µs making 1 call to Class::MOP::__ANON__[Class/MOP.pm:110]
99 }
100
101 # in case when() was used without an explicit return, the C<for>
102 # loop will be aborted and there's no useful return value
103 }
104
105 return;
106 } else {
107 # no failure, $@ is back to what it was, everything is fine
1085211.21ms return $wantarray ? @ret : $ret[0];
109 }
110}
111
112
# spent 1.68ms within Try::Tiny::catch which was called 521 times, avg 3µs/call: # 276 times (733µs+0s) by Class::MOP::Attribute::_process_accessors at line 343 of Class/MOP/Attribute.pm, avg 3µs/call # 201 times (666µs+0s) by Class::MOP::Class::_post_add_attribute at line 519 of Class/MOP/Class.pm, avg 3µs/call # 36 times (242µs+0s) by Moose::Object::DESTROY at line 100 of Moose/Object.pm, avg 7µs/call # 8 times (37µs+0s) by Class::MOP::load_first_existing_class at line 110 of Class/MOP.pm, avg 5µs/call
sub catch (&;@) {
113521302µs my ( $block, @rest ) = @_;
114
115 return (
1165211.74ms bless(\$block, 'Try::Tiny::Catch'),
117 @rest,
118 );
119}
120
121sub finally (&;@) {
122 my ( $block, @rest ) = @_;
123
124 return (
125 bless(\$block, 'Try::Tiny::Finally'),
126 @rest,
127 );
128}
129
130sub Try::Tiny::ScopeGuard::DESTROY {
131 my $self = shift;
132 $$self->();
133}
134
135113µs__PACKAGE__
136
137__END__
138
139=pod
140
141=head1 NAME
142
143Try::Tiny - minimal try/catch with proper localization of $@
144
145=head1 SYNOPSIS
146
147 # handle errors with a catch handler
148 try {
149 die "foo";
150 } catch {
151 warn "caught error: $_";
152 };
153
154 # just silence errors
155 try {
156 die "foo";
157 };
158
159=head1 DESCRIPTION
160
161This module provides bare bones C<try>/C<catch>/C<finally> statements that are designed to
162minimize common mistakes with eval blocks, and NOTHING else.
163
164This is unlike L<TryCatch> which provides a nice syntax and avoids adding
165another call stack layer, and supports calling C<return> from the try block to
166return from the parent subroutine. These extra features come at a cost of a few
167dependencies, namely L<Devel::Declare> and L<Scope::Upper> which are
168occasionally problematic, and the additional catch filtering uses L<Moose>
169type constraints which may not be desirable either.
170
171The main focus of this module is to provide simple and reliable error handling
172for those having a hard time installing L<TryCatch>, but who still want to
173write correct C<eval> blocks without 5 lines of boilerplate each time.
174
175It's designed to work as correctly as possible in light of the various
176pathological edge cases (see L<BACKGROUND>) and to be compatible with any style
177of error values (simple strings, references, objects, overloaded objects, etc).
178
179If the try block dies, it returns the value of the last statement executed in
180the catch block, if there is one. Otherwise, it returns C<undef> in scalar
181context or the empty list in list context. The following two examples both
182assign C<"bar"> to C<$x>.
183
184 my $x = try { die "foo" } catch { "bar" };
185
186 my $x = eval { die "foo" } || "bar";
187
188You can add finally blocks making the following true.
189
190 my $x;
191 try { die 'foo' } finally { $x = 'bar' };
192 try { die 'foo' } catch { warn "Got a die: $_" } finally { $x = 'bar' };
193
194Finally blocks are always executed making them suitable for cleanup code
195which cannot be handled using local.
196
197=head1 EXPORTS
198
199All functions are exported by default using L<Exporter>.
200
201If you need to rename the C<try>, C<catch> or C<finally> keyword consider using
202L<Sub::Import> to get L<Sub::Exporter>'s flexibility.
203
204=over 4
205
206=item try (&;@)
207
208Takes one mandatory try subroutine, an optional catch subroutine & finally
209subroutine.
210
211The mandatory subroutine is evaluated in the context of an C<eval> block.
212
213If no error occurred the value from the first block is returned, preserving
214list/scalar context.
215
216If there was an error and the second subroutine was given it will be invoked
217with the error in C<$_> (localized) and as that block's first and only
218argument.
219
220Note that the error may be false, but if that happens the C<catch> block will
221still be invoked.
222
223Once all execution is finished then the finally block if given will execute.
224
225=item catch (&;$)
226
227Intended to be used in the second argument position of C<try>.
228
229Returns a reference to the subroutine it was given but blessed as
230C<Try::Tiny::Catch> which allows try to decode correctly what to do
231with this code reference.
232
233 catch { ... }
234
235Inside the catch block the previous value of C<$@> is still available for use.
236This value may or may not be meaningful depending on what happened before the
237C<try>, but it might be a good idea to preserve it in an error stack.
238
239=item finally (&;$)
240
241 try { ... }
242 catch { ... }
243 finally { ... };
244
245Or
246
247 try { ... }
248 finally { ... };
249
250Or even
251
252 try { ... }
253 finally { ... }
254 catch { ... };
255
256Intended to be the second or third element of C<try>. Finally blocks are always
257executed in the event of a successful C<try> or if C<catch> is run. This allows
258you to locate cleanup code which cannot be done via C<local()> e.g. closing a file
259handle.
260
261B<You must always do your own error handling in the finally block>. C<Try::Tiny> will
262not do anything about handling possible errors coming from code located in these
263blocks.
264
265In the same way C<catch()> blesses the code reference this subroutine does the same
266except it bless them as C<Try::Tiny::Finally>.
267
268=back
269
270=head1 BACKGROUND
271
272There are a number of issues with C<eval>.
273
274=head2 Clobbering $@
275
276When you run an eval block and it succeeds, C<$@> will be cleared, potentially
277clobbering an error that is currently being caught.
278
279This causes action at a distance, clearing previous errors your caller may have
280not yet handled.
281
282C<$@> must be properly localized before invoking C<eval> in order to avoid this
283issue.
284
285More specifically, C<$@> is clobbered at the begining of the C<eval>, which
286also makes it impossible to capture the previous error before you die (for
287instance when making exception objects with error stacks).
288
289For this reason C<try> will actually set C<$@> to its previous value (before
290the localization) in the beginning of the C<eval> block.
291
292=head2 Localizing $@ silently masks errors
293
294Inside an eval block C<die> behaves sort of like:
295
296 sub die {
297 $@ = $_[0];
298 return_undef_from_eval();
299 }
300
301This means that if you were polite and localized C<$@> you can't die in that
302scope, or your error will be discarded (printing "Something's wrong" instead).
303
304The workaround is very ugly:
305
306 my $error = do {
307 local $@;
308 eval { ... };
309 $@;
310 };
311
312 ...
313 die $error;
314
315=head2 $@ might not be a true value
316
317This code is wrong:
318
319 if ( $@ ) {
320 ...
321 }
322
323because due to the previous caveats it may have been unset.
324
325C<$@> could also be an overloaded error object that evaluates to false, but
326that's asking for trouble anyway.
327
328The classic failure mode is:
329
330 sub Object::DESTROY {
331 eval { ... }
332 }
333
334 eval {
335 my $obj = Object->new;
336
337 die "foo";
338 };
339
340 if ( $@ ) {
341
342 }
343
344In this case since C<Object::DESTROY> is not localizing C<$@> but still uses
345C<eval>, it will set C<$@> to C<"">.
346
347The destructor is called when the stack is unwound, after C<die> sets C<$@> to
348C<"foo at Foo.pm line 42\n">, so by the time C<if ( $@ )> is evaluated it has
349been cleared by C<eval> in the destructor.
350
351The workaround for this is even uglier than the previous ones. Even though we
352can't save the value of C<$@> from code that doesn't localize, we can at least
353be sure the eval was aborted due to an error:
354
355 my $failed = not eval {
356 ...
357
358 return 1;
359 };
360
361This is because an C<eval> that caught a C<die> will always return a false
362value.
363
364=head1 SHINY SYNTAX
365
366Using Perl 5.10 you can use L<perlsyn/"Switch statements">.
367
368The C<catch> block is invoked in a topicalizer context (like a C<given> block),
369but note that you can't return a useful value from C<catch> using the C<when>
370blocks without an explicit C<return>.
371
372This is somewhat similar to Perl 6's C<CATCH> blocks. You can use it to
373concisely match errors:
374
375 try {
376 require Foo;
377 } catch {
378 when (/^Can't locate .*?\.pm in \@INC/) { } # ignore
379 default { die $_ }
380 };
381
382=head1 CAVEATS
383
384=over 4
385
386=item *
387
388C<@_> is not available, you need to name your args:
389
390 sub foo {
391 my ( $self, @args ) = @_;
392 try { $self->bar(@args) }
393 }
394
395=item *
396
397C<return> returns from the C<try> block, not from the parent sub (note that
398this is also how C<eval> works, but not how L<TryCatch> works):
399
400 sub bar {
401 try { return "foo" };
402 return "baz";
403 }
404
405 say bar(); # "baz"
406
407=item *
408
409C<try> introduces another caller stack frame. L<Sub::Uplevel> is not used. L<Carp>
410will report this when using full stack traces. This lack of magic is considered
411a feature.
412
413=item *
414
415The value of C<$_> in the C<catch> block is not guaranteed to be the value of
416the exception thrown (C<$@>) in the C<try> block. There is no safe way to
417ensure this, since C<eval> may be used unhygenically in destructors. The only
418guarantee is that the C<catch> will be called if an exception is thrown.
419
420=item *
421
422The return value of the C<catch> block is not ignored, so if testing the result
423of the expression for truth on success, be sure to return a false value from
424the C<catch> block:
425
426 my $obj = try {
427 MightFail->new;
428 } catch {
429 ...
430
431 return; # avoid returning a true value;
432 };
433
434 return unless $obj;
435
436=back
437
438=head1 SEE ALSO
439
440=over 4
441
442=item L<TryCatch>
443
444Much more feature complete, more convenient semantics, but at the cost of
445implementation complexity.
446
447=item L<autodie>
448
449Automatic error throwing for builtin functions and more. Also designed to
450work well with C<given>/C<when>.
451
452=item L<Throwable>
453
454A lightweight role for rolling your own exception classes.
455
456=item L<Error>
457
458Exception object implementation with a C<try> statement. Does not localize
459C<$@>.
460
461=item L<Exception::Class::TryCatch>
462
463Provides a C<catch> statement, but properly calling C<eval> is your
464responsibility.
465
466The C<try> keyword pushes C<$@> onto an error stack, avoiding some of the
467issues with C<$@>, but you still need to localize to prevent clobbering.
468
469=back
470
471=head1 LIGHTNING TALK
472
473I gave a lightning talk about this module, you can see the slides (Firefox
474only):
475
476L<http://nothingmuch.woobling.org/talks/takahashi.xul?data=try_tiny.txt>
477
478Or read the source:
479
480L<http://nothingmuch.woobling.org/talks/yapc_asia_2009/try_tiny.yml>
481
482=head1 VERSION CONTROL
483
484L<http://github.com/nothingmuch/try-tiny/>
485
486=head1 AUTHOR
487
488Yuval Kogman E<lt>nothingmuch@woobling.orgE<gt>
489
490=head1 COPYRIGHT
491
492 Copyright (c) 2009 Yuval Kogman. All rights reserved.
493 This program is free software; you can redistribute
494 it and/or modify it under the terms of the MIT license.
495
496=cut
497