4 use warnings FATAL => 'all';
7 *_PERL_LT_5_8_4 = ($] < 5.008004) ? sub(){1} : sub(){0};
10 our $VERSION = '2.000000';
11 $VERSION = eval $VERSION;
13 our @WARNING_CATEGORIES = grep { exists $warnings::Offsets{$_} } qw(
18 experimental::autoderef
19 experimental::lexical_subs
20 experimental::lexical_topic
21 experimental::postderef
22 experimental::regex_sets
23 experimental::signatures
24 experimental::smartmatch
79 if (defined $_[1] && eval { $_[0]->UNIVERSAL::VERSION($_[1]); 1}) {
81 unless _PERL_LT_5_8_4;
82 $^H{strictures_enable} = int $_[1];
84 goto &UNIVERSAL::VERSION;
87 our %extra_load_states;
93 my %opts = ref $_[0] ? %{$_[0]} : @_;
94 if (!exists $opts{version}) {
96 = exists $^H{strictures_enable} ? delete $^H{strictures_enable}
99 $opts{file} = (caller)[1];
100 $class->_enable(\%opts);
104 my ($class, $opts) = @_;
105 my $version = $opts->{version};
107 if !defined $version;
108 my $method = "_enable_$version";
109 if (!$class->can($method)) {
111 Carp::croak("Major version specified as $version - not supported!");
113 $class->$method($opts);
117 my ($class, $opts) = @_;
119 warnings->import(FATAL => 'all');
121 if (_want_extra($opts->{file})) {
122 _load_extras(qw(indirect multidimensional bareword::filehandles));
123 indirect->unimport(':fatal')
124 if $extra_load_states{indirect};
125 multidimensional->unimport
126 if $extra_load_states{multidimensional};
127 bareword::filehandles->unimport
128 if $extra_load_states{'bareword::filehandles'};
132 our @V2_NONFATAL = grep { exists $warnings::Offsets{$_} } (
133 'exec', # not safe to catch
134 'recursion', # will be caught by other mechanisms
135 'internal', # not safe to catch
136 'malloc', # not safe to catch
137 'newline', # stat on nonexistent file with a newline in it
138 'experimental', # no reason for these to be fatal
139 'deprecated', # unfortunately can't make these fatal
140 'portable', # everything worked fine here, just may not elsewhere
142 our @V2_DISABLE = grep { exists $warnings::Offsets{$_} } (
143 'once' # triggers inconsistently, can't be fatalized
147 my ($class, $opts) = @_;
150 warnings->import(FATAL => @WARNING_CATEGORIES);
151 warnings->unimport(FATAL => @V2_NONFATAL);
152 warnings->import(@V2_NONFATAL);
153 warnings->unimport(@V2_DISABLE);
155 if (_want_extra($opts->{file})) {
156 _load_extras(qw(indirect multidimensional bareword::filehandles));
157 indirect->unimport(':fatal')
158 if $extra_load_states{indirect};
159 multidimensional->unimport
160 if $extra_load_states{multidimensional};
161 bareword::filehandles->unimport
162 if $extra_load_states{'bareword::filehandles'};
166 sub _want_extra_env {
167 if (exists $ENV{PERL_STRICTURES_EXTRA}) {
168 if (_PERL_LT_5_8_4 and $ENV{PERL_STRICTURES_EXTRA}) {
169 die 'PERL_STRICTURES_EXTRA checks are not available on perls older'
170 . "than 5.8.4: please unset \$ENV{PERL_STRICTURES_EXTRA}\n";
172 return $ENV{PERL_STRICTURES_EXTRA} ? 1 : 0;
179 my $want_env = _want_extra_env();
181 if defined $want_env;
184 and $file =~ /^(?:t|xt|lib|blib)[\\\/]/
185 and defined $Smells_Like_VCS ? $Smells_Like_VCS
186 : ( $Smells_Like_VCS = !!(
187 -e '.git' || -e '.svn' || -e '.hg'
188 || (-e '../../dist.ini'
189 && (-e '../../.git' || -e '../../.svn' || -e '../../.hg' ))
197 foreach my $mod (@extras) {
199 if exists $extra_load_states{$mod};
201 $extra_load_states{$mod} = eval "require $mod; 1;" or do {
204 #work around 5.8 require bug
205 (my $file = $mod) =~ s|::|/|g;
206 delete $INC{"${file}.pm"};
211 my $failed = join ' ', @failed;
212 my $extras = join ' ', @extras;
214 strictures.pm extra testing active but couldn't load all modules. Missing were:
218 Extra testing is auto-enabled in checkouts only, so if you're the author
219 of a strictures-using module you need to run:
223 but these modules are not required by your users.
233 strictures - turn on strict and make all warnings fatal
242 use warnings FATAL => 'all';
243 use warnings NONFATAL => qw(
255 except when called from a file which matches:
257 (caller)[1] =~ /^(?:t|xt|lib|blib)[\\\/]/
259 and when either C<.git>, C<.svn>, or C<.hg> is present in the current directory
260 (with the intention of only forcing extra tests on the author side) -- or when
261 C<.git>, C<.svn>, or C<.hg> is present two directories up along with
262 C<dist.ini> (which would indicate we are in a C<dzil test> operation, via
263 L<Dist::Zilla>) -- or when the C<PERL_STRICTURES_EXTRA> environment variable is
271 use warnings FATAL => 'all';
272 use warnings NONFATAL => qw(
285 no bareword::filehandles;
287 Note that C<PERL_STRICTURES_EXTRA> may at some point add even more tests, with
288 only a minor version increase, but any changes to the effect of C<use
289 strictures> in normal mode will involve a major version bump.
291 If any of the extra testing modules are not present, L<strictures> will
292 complain loudly, once, via C<warn()>, and then shut up. But you really
293 should consider installing them, they're all great anti-footgun tools.
297 I've been writing the equivalent of this module at the top of my code for
298 about a year now. I figured it was time to make it shorter.
300 Things like the importer in C<use Moose> don't help me because they turn
301 warnings on but don't make them fatal -- which from my point of view is
302 useless because I want an exception to tell me my code isn't warnings-clean.
304 Any time I see a warning from my code, that indicates a mistake.
306 Any time my code encounters a mistake, I want a crash -- not spew to STDERR
307 and then unknown (and probably undesired) subsequent behaviour.
309 I also want to ensure that obvious coding mistakes, like indirect object
310 syntax (and not so obvious mistakes that cause things to accidentally compile
311 as such) get caught, but not at the cost of an XS dependency and not at the
312 cost of blowing things up on another machine.
314 Therefore, L<strictures> turns on additional checking, but only when it thinks
315 it's running in a test file in a VCS checkout -- although if this causes
316 undesired behaviour this can be overridden by setting the
317 C<PERL_STRICTURES_EXTRA> environment variable.
319 If additional useful author side checks come to mind, I'll add them to the
320 C<PERL_STRICTURES_EXTRA> code path only -- this will result in a minor version
321 increase (e.g. 1.000000 to 1.001000 (1.1.0) or similar). Any fixes only to the
322 mechanism of this code will result in a sub-version increase (e.g. 1.000000 to
327 Depending on the version of strictures requested, different warnings will be
328 enabled. If no specific version is requested, the current version's behavior
329 will be used. Versions can be requested using perl's standard mechanism:
333 Or, by passing in a C<version> option:
335 use strictures version => 2;
342 use warnings FATAL => 'all';
343 use warnings NONFATAL => 'deprecated', 'experimental';
344 # and if in dev mode:
347 no bareword::filehandles;
354 use warnings FATAL => 'all';
355 # and if in dev mode:
358 no bareword::filehandles;
364 This method does the setup work described above in L</DESCRIPTION>. Optionally
365 accepts a C<version> option to request a specific version's behavior.
369 This method traps the C<< strictures->VERSION(1) >> call produced by a use line
370 with a version number on it and does the version check.
372 =head1 EXTRA TESTING RATIONALE
374 Every so often, somebody complains that they're deploying via C<git pull>
375 and that they don't want L<strictures> to enable itself in this case -- and that
376 setting C<PERL_STRICTURES_EXTRA> to 0 isn't acceptable (additional ways to
377 disable extra testing would be welcome but the discussion never seems to get
380 In order to allow us to skip a couple of stages and get straight to a
381 productive conversation, here's my current rationale for turning the
382 extra testing on via a heuristic:
384 The extra testing is all stuff that only ever blows up at compile time;
385 this is intentional. So the oft-raised concern that it's different code being
386 tested is only sort of the case -- none of the modules involved affect the
387 final optree to my knowledge, so the author gets some additional compile
388 time crashes which he/she then fixes, and the rest of the testing is
389 completely valid for all environments.
391 The point of the extra testing -- especially C<no indirect> -- is to catch
392 mistakes that newbie users won't even realise are mistakes without
397 where foo is an & prototyped sub that you forgot to import -- this is
398 pernicious to track down since all I<seems> fine until it gets called
399 and you get a crash. Worse still, you can fail to have imported it due
400 to a circular require, at which point you have a load order dependent
401 bug which I've seen before now I<only> show up in production due to tiny
402 differences between the production and the development environment. I wrote
403 L<http://shadow.cat/blog/matt-s-trout/indirect-but-still-fatal/> to explain
404 this particular problem before L<strictures> itself existed.
406 As such, in my experience so far L<strictures>' extra testing has
407 I<avoided> production versus development differences, not caused them.
409 Additionally, L<strictures>' policy is very much "try and provide as much
410 protection as possible for newbies -- who won't think about whether there's
411 an option to turn on or not" -- so having only the environment variable
412 is not sufficient to achieve that (I get to explain that you need to add
413 C<use strict> at least once a week on freenode #perl -- newbies sometimes
414 completely skip steps because they don't understand that that step
417 I make no claims that the heuristic is perfect -- it's already been evolved
418 significantly over time, especially for 1.004 where we changed things to
419 ensure it only fires on files in your checkout (rather than L<strictures>-using
420 modules you happened to have installed, which was just silly). However, I
421 hope the above clarifies why a heuristic approach is not only necessary but
422 desirable from a point of view of providing new users with as much safety as
423 possible, and will allow any future discussion on the subject to focus on "how
424 do we minimise annoyance to people deploying from checkouts intentionally".
440 L<bareword::filehandles>
444 =head1 COMMUNITY AND SUPPORT
448 irc.perl.org #toolchain
450 (or bug 'mst' in query on there or freenode)
452 =head2 Git repository
454 Gitweb is on http://git.shadowcat.co.uk/ and the clone URL is:
456 git clone git://git.shadowcat.co.uk/p5sagit/strictures.git
458 The web interface to the repository is at:
460 http://git.shadowcat.co.uk/gitweb/gitweb.cgi?p=p5sagit/strictures.git
464 mst - Matt S. Trout (cpan:MSTROUT) <mst@shadowcat.co.uk>
468 Karen Etheridge (cpan:ETHER) <ether@cpan.org>
470 Mithaldu - Christian Walde (cpan:MITHALDU) <walde.christian@gmail.com>
472 haarg - Graham Knop (cpan:HAARG) <haarg@haarg.org>
476 Copyright (c) 2010 the strictures L</AUTHOR> and L</CONTRIBUTORS>
481 This library is free software and may be distributed under the same terms