4 use warnings FATAL => 'all';
7 *_PERL_LT_5_8_4 = ($] < 5.008004) ? sub(){1} : sub(){0};
10 our $VERSION = '1.005006';
11 $VERSION = eval $VERSION;
16 if (defined $_[1] && eval { $_[0]->UNIVERSAL::VERSION($_[1]); 1}) {
18 unless _PERL_LT_5_8_4;
19 $^H{strictures_enable} = int $_[1];
21 goto &UNIVERSAL::VERSION;
24 our $extra_load_states;
30 my %opts = ref $_[0] ? %{$_[0]} : @_;
31 if (!exists $opts{version}) {
33 = exists $^H{strictures_enable} ? delete $^H{strictures_enable}
36 $class->_enable(\%opts);
40 my ($class, $opts) = @_;
41 my $version = $opts->{version};
44 my $method = "_enable_$version";
45 if (!$class->can($method)) {
46 die "Major version specified as $version - not supported!";
48 $class->$method($opts);
53 warnings->import(FATAL => 'all');
55 my $extra_tests = do {
56 if (exists $ENV{PERL_STRICTURES_EXTRA}) {
57 if (_PERL_LT_5_8_4 and $ENV{PERL_STRICTURES_EXTRA}) {
58 die 'PERL_STRICTURES_EXTRA checks are not available on perls older than 5.8.4: '
59 . "please unset \$ENV{PERL_STRICTURES_EXTRA}\n";
61 $ENV{PERL_STRICTURES_EXTRA};
62 } elsif (! _PERL_LT_5_8_4) {
63 (caller(3))[1] =~ /^(?:t|xt|lib|blib)[\\\/]/
64 and defined $Smells_Like_VCS ? $Smells_Like_VCS
65 : ( $Smells_Like_VCS = !!(
66 -e '.git' || -e '.svn' || -e '.hg'
67 || (-e '../../dist.ini'
68 && (-e '../../.git' || -e '../../.svn' || -e '../../.hg' ))
73 $extra_load_states ||= do {
76 foreach my $mod (qw(indirect multidimensional bareword::filehandles)) {
77 eval "require $mod; \$rv{'$mod'} = 1;" or do {
80 # courtesy of the 5.8 require bug
81 # (we do a copy because 5.16.2 at least uses the same read-only
82 # scalars for the qw() list and it doesn't seem worth a $^V check)
84 (my $file = $mod) =~ s|::|/|g;
85 delete $INC{"${file}.pm"};
90 my $failed = join ' ', @failed;
92 strictures.pm extra testing active but couldn't load all modules. Missing were:
96 Extra testing is auto-enabled in checkouts only, so if you're the author
97 of a strictures-using module you need to run:
99 cpan indirect multidimensional bareword::filehandles
101 but these modules are not required by your users.
108 indirect->unimport(':fatal') if $extra_load_states->{indirect};
109 multidimensional->unimport if $extra_load_states->{multidimensional};
110 bareword::filehandles->unimport if $extra_load_states->{'bareword::filehandles'};
119 strictures - turn on strict and make all warnings fatal
128 use warnings FATAL => 'all';
130 except when called from a file which matches:
132 (caller)[1] =~ /^(?:t|xt|lib|blib)[\\\/]/
134 and when either C<.git>, C<.svn>, or C<.hg> is present in the current directory
135 (with the intention of only forcing extra tests on the author side) -- or when
136 C<.git>, C<.svn>, or C<.hg> is present two directories up along with
137 C<dist.ini> (which would indicate we are in a C<dzil test> operation, via
138 L<Dist::Zilla>) -- or when the C<PERL_STRICTURES_EXTRA> environment variable is
146 use warnings FATAL => 'all';
149 no bareword::filehandles;
151 Note that C<PERL_STRICTURES_EXTRA> may at some point add even more tests, with
152 only a minor version increase, but any changes to the effect of C<use
153 strictures> in normal mode will involve a major version bump.
155 If any of the extra testing modules are not present, L<strictures> will
156 complain loudly, once, via C<warn()>, and then shut up. But you really
157 should consider installing them, they're all great anti-footgun tools.
161 I've been writing the equivalent of this module at the top of my code for
162 about a year now. I figured it was time to make it shorter.
164 Things like the importer in C<use Moose> don't help me because they turn
165 warnings on but don't make them fatal -- which from my point of view is
166 useless because I want an exception to tell me my code isn't warnings-clean.
168 Any time I see a warning from my code, that indicates a mistake.
170 Any time my code encounters a mistake, I want a crash -- not spew to STDERR
171 and then unknown (and probably undesired) subsequent behaviour.
173 I also want to ensure that obvious coding mistakes, like indirect object
174 syntax (and not so obvious mistakes that cause things to accidentally compile
175 as such) get caught, but not at the cost of an XS dependency and not at the
176 cost of blowing things up on another machine.
178 Therefore, L<strictures> turns on additional checking, but only when it thinks
179 it's running in a test file in a VCS checkout -- although if this causes
180 undesired behaviour this can be overridden by setting the
181 C<PERL_STRICTURES_EXTRA> environment variable.
183 If additional useful author side checks come to mind, I'll add them to the
184 C<PERL_STRICTURES_EXTRA> code path only -- this will result in a minor version
185 increase (e.g. 1.000000 to 1.001000 (1.1.0) or similar). Any fixes only to the
186 mechanism of this code will result in a sub-version increase (e.g. 1.000000 to
189 If the behaviour of C<use strictures> in normal mode changes in any way, that
190 will constitute a major version increase -- and the code already checks
191 when its version is tested to ensure that
195 will continue to only introduce the current set of strictures even if 2.0 is
202 This method does the setup work described above in L</DESCRIPTION>
206 This method traps the C<< strictures->VERSION(1) >> call produced by a use line
207 with a version number on it and does the version check.
209 =head1 EXTRA TESTING RATIONALE
211 Every so often, somebody complains that they're deploying via C<git pull>
212 and that they don't want L<strictures> to enable itself in this case -- and that
213 setting C<PERL_STRICTURES_EXTRA> to 0 isn't acceptable (additional ways to
214 disable extra testing would be welcome but the discussion never seems to get
217 In order to allow us to skip a couple of stages and get straight to a
218 productive conversation, here's my current rationale for turning the
219 extra testing on via a heuristic:
221 The extra testing is all stuff that only ever blows up at compile time;
222 this is intentional. So the oft-raised concern that it's different code being
223 tested is only sort of the case -- none of the modules involved affect the
224 final optree to my knowledge, so the author gets some additional compile
225 time crashes which he/she then fixes, and the rest of the testing is
226 completely valid for all environments.
228 The point of the extra testing -- especially C<no indirect> -- is to catch
229 mistakes that newbie users won't even realise are mistakes without
234 where foo is an & prototyped sub that you forgot to import -- this is
235 pernicious to track down since all I<seems> fine until it gets called
236 and you get a crash. Worse still, you can fail to have imported it due
237 to a circular require, at which point you have a load order dependent
238 bug which I've seen before now I<only> show up in production due to tiny
239 differences between the production and the development environment. I wrote
240 L<http://shadow.cat/blog/matt-s-trout/indirect-but-still-fatal/> to explain
241 this particular problem before L<strictures> itself existed.
243 As such, in my experience so far L<strictures>' extra testing has
244 I<avoided> production versus development differences, not caused them.
246 Additionally, L<strictures>' policy is very much "try and provide as much
247 protection as possible for newbies -- who won't think about whether there's
248 an option to turn on or not" -- so having only the environment variable
249 is not sufficient to achieve that (I get to explain that you need to add
250 C<use strict> at least once a week on freenode #perl -- newbies sometimes
251 completely skip steps because they don't understand that that step
254 I make no claims that the heuristic is perfect -- it's already been evolved
255 significantly over time, especially for 1.004 where we changed things to
256 ensure it only fires on files in your checkout (rather than L<strictures>-using
257 modules you happened to have installed, which was just silly). However, I
258 hope the above clarifies why a heuristic approach is not only necessary but
259 desirable from a point of view of providing new users with as much safety as
260 possible, and will allow any future discussion on the subject to focus on "how
261 do we minimise annoyance to people deploying from checkouts intentionally".
277 L<bareword::filehandles>
281 =head1 COMMUNITY AND SUPPORT
285 irc.perl.org #toolchain
287 (or bug 'mst' in query on there or freenode)
289 =head2 Git repository
291 Gitweb is on http://git.shadowcat.co.uk/ and the clone URL is:
293 git clone git://git.shadowcat.co.uk/p5sagit/strictures.git
295 The web interface to the repository is at:
297 http://git.shadowcat.co.uk/gitweb/gitweb.cgi?p=p5sagit/strictures.git
301 mst - Matt S. Trout (cpan:MSTROUT) <mst@shadowcat.co.uk>
305 Karen Etheridge (cpan:ETHER) <ether@cpan.org>
307 Mithaldu - Christian Walde (cpan:MITHALDU) <walde.christian@gmail.com>
309 haarg - Graham Knop (cpan:HAARG) <haarg@haarg.org>
313 Copyright (c) 2010 the strictures L</AUTHOR> and L</CONTRIBUTORS>
318 This library is free software and may be distributed under the same terms