6 use vars qw(@EXPORT @EXPORT_OK $VERSION @ISA);
15 $VERSION = eval $VERSION;
17 @EXPORT = @EXPORT_OK = qw(try catch);
20 my ( $try, $catch ) = @_;
22 # we need to save this here, the eval block will be in scalar context due
24 my $wantarray = wantarray;
26 # save the value of $@ so we can set $@ back to it in the beginning of the eval
29 my ( @ret, $error, $failed );
31 # FIXME consider using local $SIG{__DIE__} to accumulate all errors. It's
32 # not perfect, but we could provide a list of additional errors for
36 # localize $@ to prevent clobbering of previous value by a successful
40 # failed will be true if the eval dies, because 1 will not be returned
45 # evaluate the try block in the correct context
48 } elsif ( defined $wantarray ) {
54 return 1; # properly set $fail to false
57 # copy $@ to $error; when we leave this scope, local $@ will revert $@
58 # back to its previous value
62 # at this point $failed contains a true value if the eval died, even if some
63 # destructor overwrote $@ as the eval was unwinding.
65 # if we got an error, invoke the catch block.
67 # This works like given($error), but is backwards compatible and
68 # sets $_ in the dynamic scope for the body of C<$catch>
70 return $catch->($error);
73 # in case when() was used without an explicit return, the C<for>
74 # loop will be aborted and there's no useful return value
79 # no failure, $@ is back to what it was, everything is fine
80 return $wantarray ? @ret : $ret[0];
97 Try::Tiny - minimal try/catch with proper localization of $@
101 # handle errors with a catch handler
105 warn "caught error: $_";
108 # just silence errors
115 This module provides bare bones C<try>/C<catch> statements that are designed to
116 minimize common mistakes with eval blocks, and NOTHING else.
118 This is unlike L<TryCatch> which provides a nice syntax and avoids adding
119 another call stack layer, and supports calling C<return> from the try block to
120 return from the parent subroutine. These extra features come at a cost of a few
121 dependencies, namely L<Devel::Declare> and L<Scope::Upper> which are
122 occasionally problematic, and the additional catch filtering uses L<Moose>
123 type constraints which may not be desirable either.
125 The main focus of this module is to provide simple and reliable error handling
126 for those having a hard time installing L<TryCatch>, but who still want to
127 write correct C<eval> blocks without 5 lines of boilerplate each time.
129 It's designed to work as correctly as possible in light of the various
130 pathological edge cases (see L<BACKGROUND>) and to be compatible with any style
131 of error values (simple strings, references, objects, overloaded objects, etc).
135 All functions are exported by default using L<Exporter>.
137 If you need to rename the C<try> or C<catch> keyword consider using
138 L<Sub::Import> to get L<Sub::Exporter>'s flexibility.
144 Takes one mandatory try subroutine and one optional catch subroutine.
146 The mandatory subroutine is evaluated in the context of an C<eval> block.
148 If no error occurred the value from the first block is returned, preserving
151 If there was an error and the second subroutine was given it will be invoked
152 with the error in C<$_> (localized) and as that block's first and only
155 Note that the error may be false, but if that happens the C<catch> block will
160 Intended to be used in the second argument position of C<try>.
162 Just returns the subroutine it was given.
170 Inside the catch block the previous value of C<$@> is still available for use.
171 This value may or may not be meaningful depending on what happened before the
172 C<try>, but it might be a good idea to preserve it in an error stack.
178 There are a number of issues with C<eval>.
182 When you run an eval block and it succeeds, C<$@> will be cleared, potentially
183 clobbering an error that is currently being caught.
185 This causes action at a distance, clearing previous errors your caller may have
188 C<$@> must be properly localized before invoking C<eval> in order to avoid this
191 More specifically, C<$@> is clobbered at the begining of the C<eval>, which
192 also makes it impossible to capture the previous error before you die (for
193 instance when making exception objects with error stacks).
195 For this reason C<try> will actually set C<$@> to its previous value (before
196 the localization) in the beginning of the C<eval> block.
198 =head2 Localizing $@ silently masks errors
200 Inside an eval block C<die> behaves sort of like:
204 return_undef_from_eval();
207 This means that if you were polite and localized C<$@> you can't die in that
208 scope, or your error will be discarded (printing "Something's wrong" instead).
210 The workaround is very ugly:
221 =head2 $@ might not be a true value
229 because due to the previous caveats it may have been unset.
231 C<$@> could also be an overloaded error object that evaluates to false, but
232 that's asking for trouble anyway.
234 The classic failure mode is:
236 sub Object::DESTROY {
241 my $obj = Object->new;
250 In this case since C<Object::DESTROY> is not localizing C<$@> but still uses
251 C<eval>, it will set C<$@> to C<"">.
253 The destructor is called when the stack is unwound, after C<die> sets C<$@> to
254 C<"foo at Foo.pm line 42\n">, so by the time C<if ( $@ )> is evaluated it has
255 been cleared by C<eval> in the destructor.
257 The workaround for this is even uglier than the previous ones. Even though we
258 can't save the value of C<$@> from code that doesn't localize, we can at least
259 be sure the eval was aborted due to an error:
261 my $failed = not eval {
267 This is because an C<eval> that caught a C<die> will always return a false
272 Using Perl 5.10 you can use L<perlsyn/"Switch statements">.
274 The C<catch> block is invoked in a topicalizer context (like a C<given> block),
275 but note that you can't return a useful value from C<catch> using the C<when>
276 blocks without an explicit C<return>.
278 This is somewhat similar to Perl 6's C<CATCH> blocks. You can use it to
279 concisely match errors:
284 when (/^Can't locate .*?\.pm in \@INC/) { } # ignore
294 C<@_> is not available, you need to name your args:
297 my ( $self, @args ) = @_;
298 try { $self->bar(@args) }
303 C<return> returns from the C<try> block, not from the parent sub (note that
304 this is also how C<eval> works, but not how L<TryCatch> works):
307 try { return "foo" };
315 C<try> introduces another caller stack frame. L<Sub::Uplevel> is not used. L<Carp>
316 will report this when using full stack traces. This lack of magic is considered
321 The value of C<$_> in the C<catch> block is not guaranteed to be preserved,
322 there is no safe way to ensure this if C<eval> is used unhygenically in
323 destructors. It's only guaranteed that the C<catch> will be called.
333 Much more feature complete, more convenient semantics, but at the cost of
334 implementation complexity.
338 Automatic error throwing for builtin functions and more. Also designed to
339 work well with C<given>/C<when>.
343 A lightweight role for rolling your own exception classes.
347 Exception object implementation with a C<try> statement. Does not localize
350 =item L<Exception::Class::TryCatch>
352 Provides a C<catch> statement, but properly calling C<eval> is your
355 The C<try> keyword pushes C<$@> onto an error stack, avoiding some of the
356 issues with C<$@>, but you still need to localize to prevent clobbering.
360 =head1 LIGHTNING TALK
362 I gave a lightning talk about this module, you can see the slides (Firefox
365 L<http://nothingmuch.woobling.org/talks/takahashi.xul?data=try_tiny.txt>
369 L<http://nothingmuch.woobling.org/talks/yapc_asia_2009/try_tiny.yml>
371 =head1 VERSION CONTROL
373 L<http://github.com/nothingmuch/try-tiny/>
377 Yuval Kogman E<lt>nothingmuch@woobling.orgE<gt>
381 Copyright (c) 2009 Yuval Kogman. All rights reserved.
382 This program is free software; you can redistribute
383 it and/or modify it under the terms of the MIT license.