3 # ABSTRACT: minimal try/catch with proper preservation of $@
10 use Exporter 5.57 'import';
11 our @EXPORT = our @EXPORT_OK = qw(try catch finally);
14 $Carp::Internal{+__PACKAGE__}++;
17 my $su = $INC{'Sub/Util.pm'} && defined &Sub::Util::set_subname;
18 my $sn = $INC{'Sub/Name.pm'} && eval { Sub::Name->VERSION(0.08) };
20 $su = eval { require Sub::Util; } && defined &Sub::Util::set_subname;
22 $sn = eval { require Sub::Name; Sub::Name->VERSION(0.08) };
26 *_subname = $su ? \&Sub::Util::set_subname
27 : $sn ? \&Sub::Name::subname
29 *_HAS_SUBNAME = ($su || $sn) ? sub(){1} : sub(){0};
34 # Need to prototype as @ not $$ because of the way Perl evaluates the prototype.
35 # Keeping it at $$ means you only ever get 1 sub because we need to eval in a list
36 # context & not a scalar one
39 my ( $try, @code_refs ) = @_;
41 # we need to save this here, the eval block will be in scalar context due
43 my $wantarray = wantarray;
45 # work around perl bug by explicitly initializing these, due to the likelyhood
46 # this will be used in global destruction (perl rt#119311)
47 my ( $catch, @finally ) = ();
49 # find labeled blocks in the argument list.
50 # catch and finally tag the blocks by blessing a scalar reference to them.
51 foreach my $code_ref (@code_refs) {
53 if ( ref($code_ref) eq 'Try::Tiny::Catch' ) {
54 croak 'A try() may not be followed by multiple catch() blocks'
56 $catch = ${$code_ref};
57 } elsif ( ref($code_ref) eq 'Try::Tiny::Finally' ) {
58 push @finally, ${$code_ref};
61 'try() encountered an unexpected argument ('
62 . ( defined $code_ref ? $code_ref : 'undef' )
63 . ') - perhaps a missing semi-colon before or'
68 # FIXME consider using local $SIG{__DIE__} to accumulate all errors. It's
69 # not perfect, but we could provide a list of additional errors for
72 # name the blocks if we have Sub::Name installed
74 _subname("${caller}::try {...} " => $try)
77 # set up scope guards to invoke the finally blocks at the end.
78 # this should really be a lexical variable instead of our/local but that
79 # causes issues with perls < 5.20 due to perl rt#119311
80 local @_finally_guards =
81 map { Try::Tiny::ScopeGuard->_new($_) }
84 # save the value of $@ so we can set $@ back to it in the beginning of the eval
85 # and restore $@ after the eval finishes
90 # failed will be true if the eval dies, because 1 will not be returned
92 my $failed = not eval {
95 # evaluate the try block in the correct context
98 } elsif ( defined $wantarray ) {
104 return 1; # properly set $failed to false
107 # preserve the current error and reset the original value of $@
111 # at this point $failed contains a true value if the eval died, even if some
112 # destructor overwrote $@ as the eval was unwinding.
114 # pass $error to the finally blocks
115 push @$_, $error for @_finally_guards;
117 # if we got an error, invoke the catch block.
119 # This works like given($error), but is backwards compatible and
120 # sets $_ in the dynamic scope for the body of C<$catch>
122 return $catch->($error);
125 # in case when() was used without an explicit return, the C<for>
126 # loop will be aborted and there's no useful return value
131 # no failure, $@ is back to what it was, everything is fine
132 return $wantarray ? @ret : $ret[0];
137 my ( $block, @rest ) = @_;
139 croak 'Useless bare catch()' unless wantarray;
142 _subname("${caller}::catch {...} " => $block)
145 bless(\$block, 'Try::Tiny::Catch'),
151 my ( $block, @rest ) = @_;
153 croak 'Useless bare finally()' unless wantarray;
156 _subname("${caller}::finally {...} " => $block)
159 bless(\$block, 'Try::Tiny::Finally'),
165 package # hide from PAUSE
166 Try::Tiny::ScopeGuard;
168 use constant UNSTABLE_DOLLARAT => ($] < '5.013002') ? 1 : 0;
176 my ($code, @args) = @{ $_[0] };
178 local $@ if UNSTABLE_DOLLARAT;
184 "Execution of finally() block $code resulted in an exception, which "
185 . '*CAN NOT BE PROPAGATED* due to fundamental limitations of Perl. '
186 . 'Your program will continue as if this event never took place. '
187 . "Original exception text follows:\n\n"
188 . (defined $@ ? $@ : '$@ left undefined...')
203 You can use Try::Tiny's C<try> and C<catch> to expect and handle exceptional
204 conditions, avoiding quirks in Perl and common mistakes:
206 # handle errors with a catch handler
210 warn "caught error: $_"; # not $@
213 You can also use it like a standalone C<eval> to catch and ignore any error
214 conditions. Obviously, this is an extreme measure not to be undertaken
217 # just silence errors
224 This module provides bare bones C<try>/C<catch>/C<finally> statements that are designed to
225 minimize common mistakes with eval blocks, and NOTHING else.
227 This is unlike L<TryCatch> which provides a nice syntax and avoids adding
228 another call stack layer, and supports calling C<return> from the C<try> block to
229 return from the parent subroutine. These extra features come at a cost of a few
230 dependencies, namely L<Devel::Declare> and L<Scope::Upper> which are
231 occasionally problematic, and the additional catch filtering uses L<Moose>
232 type constraints which may not be desirable either.
234 The main focus of this module is to provide simple and reliable error handling
235 for those having a hard time installing L<TryCatch>, but who still want to
236 write correct C<eval> blocks without 5 lines of boilerplate each time.
238 It's designed to work as correctly as possible in light of the various
239 pathological edge cases (see L</BACKGROUND>) and to be compatible with any style
240 of error values (simple strings, references, objects, overloaded objects, etc).
242 If the C<try> block dies, it returns the value of the last statement executed in
243 the C<catch> block, if there is one. Otherwise, it returns C<undef> in scalar
244 context or the empty list in list context. The following examples all
245 assign C<"bar"> to C<$x>:
247 my $x = try { die "foo" } catch { "bar" };
248 my $x = try { die "foo" } || "bar";
249 my $x = (try { die "foo" }) // "bar";
251 my $x = eval { die "foo" } || "bar";
253 You can add C<finally> blocks, yielding the following:
256 try { die 'foo' } finally { $x = 'bar' };
257 try { die 'foo' } catch { warn "Got a die: $_" } finally { $x = 'bar' };
259 C<finally> blocks are always executed making them suitable for cleanup code
260 which cannot be handled using local. You can add as many C<finally> blocks to a
261 given C<try> block as you like.
263 Note that adding a C<finally> block without a preceding C<catch> block
264 suppresses any errors. This behaviour is consistent with using a standalone
265 C<eval>, but it is not consistent with C<try>/C<finally> patterns found in
266 other programming languages, such as Java, Python, Javascript or C#. If you
267 learnt the C<try>/C<finally> pattern from one of these languages, watch out for
272 All functions are exported by default using L<Exporter>.
274 If you need to rename the C<try>, C<catch> or C<finally> keyword consider using
275 L<Sub::Import> to get L<Sub::Exporter>'s flexibility.
281 Takes one mandatory C<try> subroutine, an optional C<catch> subroutine and C<finally>
284 The mandatory subroutine is evaluated in the context of an C<eval> block.
286 If no error occurred the value from the first block is returned, preserving
289 If there was an error and the second subroutine was given it will be invoked
290 with the error in C<$_> (localized) and as that block's first and only
293 C<$@> does B<not> contain the error. Inside the C<catch> block it has the same
294 value it had before the C<try> block was executed.
296 Note that the error may be false, but if that happens the C<catch> block will
299 Once all execution is finished then the C<finally> block, if given, will execute.
303 Intended to be used in the second argument position of C<try>.
305 Returns a reference to the subroutine it was given but blessed as
306 C<Try::Tiny::Catch> which allows try to decode correctly what to do
307 with this code reference.
311 Inside the C<catch> block the caught error is stored in C<$_>, while previous
312 value of C<$@> is still available for use. This value may or may not be
313 meaningful depending on what happened before the C<try>, but it might be a good
314 idea to preserve it in an error stack.
316 For code that captures C<$@> when throwing new errors (i.e.
317 L<Class::Throwable>), you'll need to do:
338 Intended to be the second or third element of C<try>. C<finally> blocks are always
339 executed in the event of a successful C<try> or if C<catch> is run. This allows
340 you to locate cleanup code which cannot be done via C<local()> e.g. closing a file
343 When invoked, the C<finally> block is passed the error that was caught. If no
344 error was caught, it is passed nothing. (Note that the C<finally> block does not
345 localize C<$_> with the error, since unlike in a C<catch> block, there is no way
346 to know if C<$_ == undef> implies that there were no errors.) In other words,
347 the following code does just what you would expect:
352 # ...code run in case of error
355 print "The try block died with: @_\n";
357 print "The try block ran without error.\n";
361 B<You must always do your own error handling in the C<finally> block>. C<Try::Tiny> will
362 not do anything about handling possible errors coming from code located in these
365 Furthermore B<exceptions in C<finally> blocks are not trappable and are unable
366 to influence the execution of your program>. This is due to limitation of
367 C<DESTROY>-based scope guards, which C<finally> is implemented on top of. This
368 may change in a future version of Try::Tiny.
370 In the same way C<catch()> blesses the code reference this subroutine does the same
371 except it bless them as C<Try::Tiny::Finally>.
377 There are a number of issues with C<eval>.
381 When you run an C<eval> block and it succeeds, C<$@> will be cleared, potentially
382 clobbering an error that is currently being caught.
384 This causes action at a distance, clearing previous errors your caller may have
387 C<$@> must be properly localized before invoking C<eval> in order to avoid this
390 More specifically, C<$@> is clobbered at the beginning of the C<eval>, which
391 also makes it impossible to capture the previous error before you die (for
392 instance when making exception objects with error stacks).
394 For this reason C<try> will actually set C<$@> to its previous value (the one
395 available before entering the C<try> block) in the beginning of the C<eval>
398 =head2 Localizing $@ silently masks errors
400 Inside an C<eval> block, C<die> behaves sort of like:
404 return_undef_from_eval();
407 This means that if you were polite and localized C<$@> you can't die in that
408 scope, or your error will be discarded (printing "Something's wrong" instead).
410 The workaround is very ugly:
421 =head2 $@ might not be a true value
429 because due to the previous caveats it may have been unset.
431 C<$@> could also be an overloaded error object that evaluates to false, but
432 that's asking for trouble anyway.
434 The classic failure mode is:
436 sub Object::DESTROY {
441 my $obj = Object->new;
450 In this case since C<Object::DESTROY> is not localizing C<$@> but still uses
451 C<eval>, it will set C<$@> to C<"">.
453 The destructor is called when the stack is unwound, after C<die> sets C<$@> to
454 C<"foo at Foo.pm line 42\n">, so by the time C<if ( $@ )> is evaluated it has
455 been cleared by C<eval> in the destructor.
457 The workaround for this is even uglier than the previous ones. Even though we
458 can't save the value of C<$@> from code that doesn't localize, we can at least
459 be sure the C<eval> was aborted due to an error:
461 my $failed = not eval {
467 This is because an C<eval> that caught a C<die> will always return a false
472 Using Perl 5.10 you can use L<perlsyn/"Switch statements">.
474 =for stopwords topicalizer
476 The C<catch> block is invoked in a topicalizer context (like a C<given> block),
477 but note that you can't return a useful value from C<catch> using the C<when>
478 blocks without an explicit C<return>.
480 This is somewhat similar to Perl 6's C<CATCH> blocks. You can use it to
481 concisely match errors:
486 when (/^Can't locate .*?\.pm in \@INC/) { } # ignore
496 C<@_> is not available within the C<try> block, so you need to copy your
497 argument list. In case you want to work with argument values directly via C<@_>
498 aliasing (i.e. allow C<$_[1] = "foo">), you need to pass C<@_> by reference:
501 my ( $self, @args ) = @_;
502 try { $self->bar(@args) }
510 try { $_ = $self->bar($_) for @$args }
515 C<return> returns from the C<try> block, not from the parent sub (note that
516 this is also how C<eval> works, but not how L<TryCatch> works):
526 say "this text WILL be displayed, even though an exception is thrown";
529 Instead, you should capture the return value:
536 return unless $success;
538 say "This text WILL NEVER appear!";
541 sub parent_sub_with_catch {
547 # do something with $_
548 return undef; #see note
550 return unless $success;
552 say "This text WILL NEVER appear!";
555 Note that if you have a C<catch> block, it must return C<undef> for this to work,
556 since if a C<catch> block exists, its return value is returned in place of C<undef>
557 when an exception is thrown.
561 C<try> introduces another caller stack frame. L<Sub::Uplevel> is not used. L<Carp>
562 will not report this when using full stack traces, though, because
563 C<%Carp::Internal> is used. This lack of magic is considered a feature.
565 =for stopwords unhygienically
569 The value of C<$_> in the C<catch> block is not guaranteed to be the value of
570 the exception thrown (C<$@>) in the C<try> block. There is no safe way to
571 ensure this, since C<eval> may be used unhygienically in destructors. The only
572 guarantee is that the C<catch> will be called if an exception is thrown.
576 The return value of the C<catch> block is not ignored, so if testing the result
577 of the expression for truth on success, be sure to return a false value from
585 return; # avoid returning a true value;
592 C<$SIG{__DIE__}> is still in effect.
594 Though it can be argued that C<$SIG{__DIE__}> should be disabled inside of
595 C<eval> blocks, since it isn't people have grown to rely on it. Therefore in
596 the interests of compatibility, C<try> does not disable C<$SIG{__DIE__}> for
597 the scope of the error throwing code.
601 Lexical C<$_> may override the one set by C<catch>.
603 For example Perl 5.10's C<given> form uses a lexical C<$_>, creating some
611 warn $_; # will print $foo, not the error
612 warn $_[0]; # instead, get the error like this
617 Note that this behavior was changed once again in L<Perl5 version 18
618 |https://metacpan.org/module/perldelta#given-now-aliases-the-global-_>.
619 However, since the entirety of lexical C<$_> is now L<considered experimental
620 |https://metacpan.org/module/perldelta#Lexical-_-is-now-experimental>, it
621 is unclear whether the new version 18 behavior is final.
631 Much more feature complete, more convenient semantics, but at the cost of
632 implementation complexity.
636 Automatic error throwing for builtin functions and more. Also designed to
637 work well with C<given>/C<when>.
641 A lightweight role for rolling your own exception classes.
645 Exception object implementation with a C<try> statement. Does not localize
648 =item L<Exception::Class::TryCatch>
650 Provides a C<catch> statement, but properly calling C<eval> is your
653 The C<try> keyword pushes C<$@> onto an error stack, avoiding some of the
654 issues with C<$@>, but you still need to localize to prevent clobbering.
658 =head1 LIGHTNING TALK
660 I gave a lightning talk about this module, you can see the slides (Firefox
663 L<http://web.archive.org/web/20100628040134/http://nothingmuch.woobling.org/talks/takahashi.xul>
667 L<http://web.archive.org/web/20100305133605/http://nothingmuch.woobling.org/talks/yapc_asia_2009/try_tiny.yml>
669 =head1 VERSION CONTROL
671 L<http://github.com/doy/try-tiny/>