7 $VERSION = eval $VERSION if $VERSION =~ /_/;
10 our @EXPORT = our @EXPORT_OK = qw(try catch finally);
13 $Carp::Internal{+__PACKAGE__}++;
15 # Need to prototype as @ not $$ because of the way Perl evaluates the prototype.
16 # Keeping it at $$ means you only ever get 1 sub because we need to eval in a list
17 # context & not a scalar one
20 my ( $try, @code_refs ) = @_;
22 # we need to save this here, the eval block will be in scalar context due
24 my $wantarray = wantarray;
26 my ( $catch, @finally );
28 # find labeled blocks in the argument list.
29 # catch and finally tag the blocks by blessing a scalar reference to them.
30 foreach my $code_ref (@code_refs) {
32 if ( ref($code_ref) eq 'Try::Tiny::Catch' ) {
33 croak 'A try() may not be followed by multiple catch() blocks'
35 $catch = ${$code_ref};
36 } elsif ( ref($code_ref) eq 'Try::Tiny::Finally' ) {
37 push @finally, ${$code_ref};
40 'try() encountered an unexpected argument ('
41 . ( defined $code_ref ? $code_ref : 'undef' )
42 . ') - perhaps a missing semi-colon before or'
47 # FIXME consider using local $SIG{__DIE__} to accumulate all errors. It's
48 # not perfect, but we could provide a list of additional errors for
51 # save the value of $@ so we can set $@ back to it in the beginning of the eval
52 # and restore $@ after the eval finishes
57 # failed will be true if the eval dies, because 1 will not be returned
59 my $failed = not eval {
62 # evaluate the try block in the correct context
65 } elsif ( defined $wantarray ) {
71 return 1; # properly set $fail to false
74 # preserve the current error and reset the original value of $@
78 # set up a scope guard to invoke the finally block at the end
80 map { Try::Tiny::ScopeGuard->_new($_, $failed ? $error : ()) }
83 # at this point $failed contains a true value if the eval died, even if some
84 # destructor overwrote $@ as the eval was unwinding.
86 # if we got an error, invoke the catch block.
88 # This works like given($error), but is backwards compatible and
89 # sets $_ in the dynamic scope for the body of C<$catch>
91 return $catch->($error);
94 # in case when() was used without an explicit return, the C<for>
95 # loop will be aborted and there's no useful return value
100 # no failure, $@ is back to what it was, everything is fine
101 return $wantarray ? @ret : $ret[0];
106 my ( $block, @rest ) = @_;
108 croak 'Useless bare catch()' unless defined wantarray;
111 bless(\$block, 'Try::Tiny::Catch'),
117 my ( $block, @rest ) = @_;
119 croak 'Useless bare finally()' unless defined wantarray;
122 bless(\$block, 'Try::Tiny::Finally'),
128 package # hide from PAUSE
129 Try::Tiny::ScopeGuard;
131 use constant UNSTABLE_DOLLARAT => ($] < '5.013002') ? 1 : 0;
139 my ($code, @args) = @{ $_[0] };
141 local $@ if UNSTABLE_DOLLARAT;
147 "Execution of finally() block $code resulted in an exception, which "
148 . '*CAN NOT BE PROPAGATED* due to fundamental limitations of Perl. '
149 . 'Your program will continue as if this event never took place. '
150 . "Original exception text follows:\n\n"
151 . (defined $@ ? $@ : '$@ left undefined...')
166 Try::Tiny - minimal try/catch with proper preservation of $@
170 You can use Try::Tiny's C<try> and C<catch> to expect and handle exceptional
171 conditions, avoiding quirks in Perl and common mistakes:
173 # handle errors with a catch handler
177 warn "caught error: $_"; # not $@
180 You can also use it like a standalone C<eval> to catch and ignore any error
181 conditions. Obviously, this is an extreme measure not to be undertaken
184 # just silence errors
191 This module provides bare bones C<try>/C<catch>/C<finally> statements that are designed to
192 minimize common mistakes with eval blocks, and NOTHING else.
194 This is unlike L<TryCatch> which provides a nice syntax and avoids adding
195 another call stack layer, and supports calling C<return> from the C<try> block to
196 return from the parent subroutine. These extra features come at a cost of a few
197 dependencies, namely L<Devel::Declare> and L<Scope::Upper> which are
198 occasionally problematic, and the additional catch filtering uses L<Moose>
199 type constraints which may not be desirable either.
201 The main focus of this module is to provide simple and reliable error handling
202 for those having a hard time installing L<TryCatch>, but who still want to
203 write correct C<eval> blocks without 5 lines of boilerplate each time.
205 It's designed to work as correctly as possible in light of the various
206 pathological edge cases (see L</BACKGROUND>) and to be compatible with any style
207 of error values (simple strings, references, objects, overloaded objects, etc).
209 If the C<try> block dies, it returns the value of the last statement executed in
210 the C<catch> block, if there is one. Otherwise, it returns C<undef> in scalar
211 context or the empty list in list context. The following examples all
212 assign C<"bar"> to C<$x>:
214 my $x = try { die "foo" } catch { "bar" };
215 my $x = try { die "foo" } || { "bar" };
216 my $x = (try { die "foo" }) // { "bar" };
218 my $x = eval { die "foo" } || "bar";
220 You can add C<finally> blocks, yielding the following:
223 try { die 'foo' } finally { $x = 'bar' };
224 try { die 'foo' } catch { warn "Got a die: $_" } finally { $x = 'bar' };
226 C<finally> blocks are always executed making them suitable for cleanup code
227 which cannot be handled using local. You can add as many C<finally> blocks to a
228 given C<try> block as you like.
232 All functions are exported by default using L<Exporter>.
234 If you need to rename the C<try>, C<catch> or C<finally> keyword consider using
235 L<Sub::Import> to get L<Sub::Exporter>'s flexibility.
241 Takes one mandatory C<try> subroutine, an optional C<catch> subroutine and C<finally>
244 The mandatory subroutine is evaluated in the context of an C<eval> block.
246 If no error occurred the value from the first block is returned, preserving
249 If there was an error and the second subroutine was given it will be invoked
250 with the error in C<$_> (localized) and as that block's first and only
253 C<$@> does B<not> contain the error. Inside the C<catch> block it has the same
254 value it had before the C<try> block was executed.
256 Note that the error may be false, but if that happens the C<catch> block will
259 Once all execution is finished then the C<finally> block, if given, will execute.
263 Intended to be used in the second argument position of C<try>.
265 Returns a reference to the subroutine it was given but blessed as
266 C<Try::Tiny::Catch> which allows try to decode correctly what to do
267 with this code reference.
271 Inside the C<catch> block the caught error is stored in C<$_>, while previous
272 value of C<$@> is still available for use. This value may or may not be
273 meaningful depending on what happened before the C<try>, but it might be a good
274 idea to preserve it in an error stack.
276 For code that captures C<$@> when throwing new errors (i.e.
277 L<Class::Throwable>), you'll need to do:
298 Intended to be the second or third element of C<try>. C<finally> blocks are always
299 executed in the event of a successful C<try> or if C<catch> is run. This allows
300 you to locate cleanup code which cannot be done via C<local()> e.g. closing a file
303 When invoked, the C<finally> block is passed the error that was caught. If no
304 error was caught, it is passed nothing. (Note that the C<finally> block does not
305 localize C<$_> with the error, since unlike in a C<catch> block, there is no way
306 to know if C<$_ == undef> implies that there were no errors.) In other words,
307 the following code does just what you would expect:
312 # ...code run in case of error
315 print "The try block died with: @_\n";
317 print "The try block ran without error.\n";
321 B<You must always do your own error handling in the C<finally> block>. C<Try::Tiny> will
322 not do anything about handling possible errors coming from code located in these
325 Furthermore B<exceptions in C<finally> blocks are not trappable and are unable
326 to influence the execution of your program>. This is due to limitation of
327 C<DESTROY>-based scope guards, which C<finally> is implemented on top of. This
328 may change in a future version of Try::Tiny.
330 In the same way C<catch()> blesses the code reference this subroutine does the same
331 except it bless them as C<Try::Tiny::Finally>.
337 There are a number of issues with C<eval>.
341 When you run an C<eval> block and it succeeds, C<$@> will be cleared, potentially
342 clobbering an error that is currently being caught.
344 This causes action at a distance, clearing previous errors your caller may have
347 C<$@> must be properly localized before invoking C<eval> in order to avoid this
350 More specifically, C<$@> is clobbered at the beginning of the C<eval>, which
351 also makes it impossible to capture the previous error before you die (for
352 instance when making exception objects with error stacks).
354 For this reason C<try> will actually set C<$@> to its previous value (the one
355 available before entering the C<try> block) in the beginning of the C<eval>
358 =head2 Localizing $@ silently masks errors
360 Inside an C<eval> block, C<die> behaves sort of like:
364 return_undef_from_eval();
367 This means that if you were polite and localized C<$@> you can't die in that
368 scope, or your error will be discarded (printing "Something's wrong" instead).
370 The workaround is very ugly:
381 =head2 $@ might not be a true value
389 because due to the previous caveats it may have been unset.
391 C<$@> could also be an overloaded error object that evaluates to false, but
392 that's asking for trouble anyway.
394 The classic failure mode is:
396 sub Object::DESTROY {
401 my $obj = Object->new;
410 In this case since C<Object::DESTROY> is not localizing C<$@> but still uses
411 C<eval>, it will set C<$@> to C<"">.
413 The destructor is called when the stack is unwound, after C<die> sets C<$@> to
414 C<"foo at Foo.pm line 42\n">, so by the time C<if ( $@ )> is evaluated it has
415 been cleared by C<eval> in the destructor.
417 The workaround for this is even uglier than the previous ones. Even though we
418 can't save the value of C<$@> from code that doesn't localize, we can at least
419 be sure the C<eval> was aborted due to an error:
421 my $failed = not eval {
427 This is because an C<eval> that caught a C<die> will always return a false
432 Using Perl 5.10 you can use L<perlsyn/"Switch statements">.
434 The C<catch> block is invoked in a topicalizer context (like a C<given> block),
435 but note that you can't return a useful value from C<catch> using the C<when>
436 blocks without an explicit C<return>.
438 This is somewhat similar to Perl 6's C<CATCH> blocks. You can use it to
439 concisely match errors:
444 when (/^Can't locate .*?\.pm in \@INC/) { } # ignore
454 C<@_> is not available within the C<try> block, so you need to copy your
455 arglist. In case you want to work with argument values directly via C<@_>
456 aliasing (i.e. allow C<$_[1] = "foo">), you need to pass C<@_> by reference:
459 my ( $self, @args ) = @_;
460 try { $self->bar(@args) }
468 try { $_ = $self->bar($_) for @$args }
473 C<return> returns from the C<try> block, not from the parent sub (note that
474 this is also how C<eval> works, but not how L<TryCatch> works):
484 say "this text WILL be displayed, even though an exception is thrown";
487 Instead, you should capture the return value:
494 return unless $success;
496 say "This text WILL NEVER appear!";
499 Note that if you have a C<catch> block, it must return C<undef> for this to work,
500 since if a C<catch> block exists, its return value is returned in place of C<undef>
501 when an exception is thrown.
505 C<try> introduces another caller stack frame. L<Sub::Uplevel> is not used. L<Carp>
506 will not report this when using full stack traces, though, because
507 C<%Carp::Internal> is used. This lack of magic is considered a feature.
511 The value of C<$_> in the C<catch> block is not guaranteed to be the value of
512 the exception thrown (C<$@>) in the C<try> block. There is no safe way to
513 ensure this, since C<eval> may be used unhygenically in destructors. The only
514 guarantee is that the C<catch> will be called if an exception is thrown.
518 The return value of the C<catch> block is not ignored, so if testing the result
519 of the expression for truth on success, be sure to return a false value from
527 return; # avoid returning a true value;
534 C<$SIG{__DIE__}> is still in effect.
536 Though it can be argued that C<$SIG{__DIE__}> should be disabled inside of
537 C<eval> blocks, since it isn't people have grown to rely on it. Therefore in
538 the interests of compatibility, C<try> does not disable C<$SIG{__DIE__}> for
539 the scope of the error throwing code.
543 Lexical C<$_> may override the one set by C<catch>.
545 For example Perl 5.10's C<given> form uses a lexical C<$_>, creating some
553 warn $_; # will print $foo, not the error
554 warn $_[0]; # instead, get the error like this
559 Note that this behavior was changed once again in L<Perl5 version 18
560 |https://metacpan.org/module/perldelta#given-now-aliases-the-global-_>.
561 However, since the entirety of lexical C<$_> is now L<considired experimental
562 |https://metacpan.org/module/perldelta#Lexical-_-is-now-experimental>, it
563 is unclear whether the new version 18 behavior is final.
573 Much more feature complete, more convenient semantics, but at the cost of
574 implementation complexity.
578 Automatic error throwing for builtin functions and more. Also designed to
579 work well with C<given>/C<when>.
583 A lightweight role for rolling your own exception classes.
587 Exception object implementation with a C<try> statement. Does not localize
590 =item L<Exception::Class::TryCatch>
592 Provides a C<catch> statement, but properly calling C<eval> is your
595 The C<try> keyword pushes C<$@> onto an error stack, avoiding some of the
596 issues with C<$@>, but you still need to localize to prevent clobbering.
600 =head1 LIGHTNING TALK
602 I gave a lightning talk about this module, you can see the slides (Firefox
605 L<http://web.archive.org/web/20100628040134/http://nothingmuch.woobling.org/talks/takahashi.xul>
609 L<http://web.archive.org/web/20100305133605/http://nothingmuch.woobling.org/talks/yapc_asia_2009/try_tiny.yml>
611 =head1 VERSION CONTROL
613 L<http://github.com/nothingmuch/try-tiny/>
617 Yuval Kogman E<lt>nothingmuch@woobling.orgE<gt>
621 Copyright (c) 2009 Yuval Kogman. All rights reserved.
622 This program is free software; you can redistribute
623 it and/or modify it under the terms of the MIT license.