1 |
=head1 NAME |
2 |
|
3 |
common::sense - save a tree AND a kitten, use common::sense! |
4 |
|
5 |
=head1 SYNOPSIS |
6 |
|
7 |
use common::sense; |
8 |
|
9 |
# Supposed to be mostly the same, with much lower memory usage, as: |
10 |
|
11 |
# use utf8; |
12 |
# use strict qw(vars subs); |
13 |
# use feature qw(say state switch); |
14 |
# use feature qw(unicode_strings unicode_eval current_sub fc evalbytes); |
15 |
# no feature qw(array_base); |
16 |
# no warnings; |
17 |
# use warnings qw(FATAL closed threads internal debugging pack |
18 |
# portable prototype inplace io pipe unpack malloc |
19 |
# deprecated glob digit printf layer |
20 |
# reserved taint closure semicolon); |
21 |
# no warnings qw(exec newline unopened); |
22 |
|
23 |
=head1 DESCRIPTION |
24 |
|
25 |
“Nothing is more fairly distributed than common sense: no one thinks |
26 |
he needs more of it than he already has.” |
27 |
|
28 |
– René Descartes |
29 |
|
30 |
This module implements some sane defaults for Perl programs, as defined by |
31 |
two typical (or not so typical - use your common sense) specimens of Perl |
32 |
coders. In fact, after working out details on which warnings and strict |
33 |
modes to enable and make fatal, we found that we (and our code written so |
34 |
far, and others) fully agree on every option, even though we never used |
35 |
warnings before, so it seems this module indeed reflects a "common" sense |
36 |
among some long-time Perl coders. |
37 |
|
38 |
The basic philosophy behind the choices made in common::sense can be |
39 |
summarised as: "enforcing strict policies to catch as many bugs as |
40 |
possible, while at the same time, not limiting the expressive power |
41 |
available to the programmer". |
42 |
|
43 |
Two typical examples of how this philosophy is applied in practise is the |
44 |
handling of uninitialised and malloc warnings: |
45 |
|
46 |
=over 4 |
47 |
|
48 |
=item I<uninitialised> |
49 |
|
50 |
C<undef> is a well-defined feature of perl, and enabling warnings for |
51 |
using it rarely catches any bugs, but considerably limits you in what you |
52 |
can do, so uninitialised warnings are disabled. |
53 |
|
54 |
=item I<malloc> |
55 |
|
56 |
Freeing something twice on the C level is a serious bug, usually causing |
57 |
memory corruption. It often leads to side effects much later in the |
58 |
program and there are no advantages to not reporting this, so malloc |
59 |
warnings are fatal by default. |
60 |
|
61 |
=back |
62 |
|
63 |
Unfortunately, there is no fine-grained warning control in perl, so often |
64 |
whole groups of useful warnings had to be excluded because of a single |
65 |
useless warning (for example, perl puts an arbitrary limit on the length |
66 |
of text you can match with some regexes before emitting a warning, making |
67 |
the whole C<regexp> category useless). |
68 |
|
69 |
What follows is a more thorough discussion of what this module does, |
70 |
and why it does it, and what the advantages (and disadvantages) of this |
71 |
approach are. |
72 |
|
73 |
=head1 RATIONALE |
74 |
|
75 |
=over 4 |
76 |
|
77 |
=item use utf8 |
78 |
|
79 |
While it's not common sense to write your programs in UTF-8, it's quickly |
80 |
becoming the most common encoding, is the designated future default |
81 |
encoding for perl sources, and the most convenient encoding available |
82 |
(you can do really nice quoting tricks...). Experience has shown that our |
83 |
programs were either all pure ascii or utf-8, both of which will stay the |
84 |
same. |
85 |
|
86 |
There are few drawbacks to enabling UTF-8 source code by default (mainly |
87 |
some speed hits due to bugs in older versions of perl), so this module |
88 |
enables UTF-8 source code encoding by default. |
89 |
|
90 |
|
91 |
=item use strict qw(subs vars) |
92 |
|
93 |
Using C<use strict> is definitely common sense, but C<use strict |
94 |
'refs'> definitely overshoots its usefulness. After almost two |
95 |
decades of Perl hacking, we decided that it does more harm than being |
96 |
useful. Specifically, constructs like these: |
97 |
|
98 |
@{ $var->[0] } |
99 |
|
100 |
Must be written like this (or similarly), when C<use strict 'refs'> is in |
101 |
scope, and C<$var> can legally be C<undef>: |
102 |
|
103 |
@{ $var->[0] || [] } |
104 |
|
105 |
This is annoying, and doesn't shield against obvious mistakes such as |
106 |
using C<"">, so one would even have to write (at least for the time |
107 |
being): |
108 |
|
109 |
@{ defined $var->[0] ? $var->[0] : [] } |
110 |
|
111 |
... which nobody with a bit of common sense would consider |
112 |
writing: clear code is clearly something else. |
113 |
|
114 |
Curiously enough, sometimes perl is not so strict, as this works even with |
115 |
C<use strict> in scope: |
116 |
|
117 |
for (@{ $var->[0] }) { ... |
118 |
|
119 |
If that isn't hypocrisy! And all that from a mere program! |
120 |
|
121 |
|
122 |
=item use feature qw(say state given ...) |
123 |
|
124 |
We found it annoying that we always have to enable extra features. If |
125 |
something breaks because it didn't anticipate future changes, so be |
126 |
it. 5.10 broke almost all our XS modules and nobody cared either (or at |
127 |
least I know of nobody who really complained about gratuitous changes - |
128 |
as opposed to bugs). |
129 |
|
130 |
Few modules that are not actively maintained work with newer versions of |
131 |
Perl, regardless of use feature or not, so a new major perl release means |
132 |
changes to many modules - new keywords are just the tip of the iceberg. |
133 |
|
134 |
If your code isn't alive, it's dead, Jim - be an active maintainer. |
135 |
|
136 |
But nobody forces you to use those extra features in modules meant for |
137 |
older versions of perl - common::sense of course works there as well. |
138 |
There is also an important other mode where having additional features by |
139 |
default is useful: commandline hacks and internal use scripts: See "much |
140 |
reduced typing", below. |
141 |
|
142 |
There is one notable exception: C<unicode_eval> is not enabled by |
143 |
default. In our opinion, C<use feature> had one main effect - newer perl |
144 |
versions don't value backwards compatibility and the ability to write |
145 |
modules for multiple perl versions much, after all, you can use feature. |
146 |
|
147 |
C<unicode_eval> doesn't add a new feature, it breaks an existing function. |
148 |
|
149 |
=item no warnings, but a lot of new errors |
150 |
|
151 |
Ah, the dreaded warnings. Even worse, the horribly dreaded C<-w> |
152 |
switch: Even though we don't care if other people use warnings (and |
153 |
certainly there are useful ones), a lot of warnings simply go against the |
154 |
spirit of Perl. |
155 |
|
156 |
Most prominently, the warnings related to C<undef>. There is nothing wrong |
157 |
with C<undef>: it has well-defined semantics, it is useful, and spitting |
158 |
out warnings you never asked for is just evil. |
159 |
|
160 |
The result was that every one of our modules did C<no warnings> in the |
161 |
past, to avoid somebody accidentally using and forcing his bad standards |
162 |
on our code. Of course, this switched off all warnings, even the useful |
163 |
ones. Not a good situation. Really, the C<-w> switch should only enable |
164 |
warnings for the main program only. |
165 |
|
166 |
Funnily enough, L<perllexwarn> explicitly mentions C<-w> (and not in a |
167 |
favourable way, calling it outright "wrong"), but standard utilities, such |
168 |
as L<prove>, or MakeMaker when running C<make test>, still enable them |
169 |
blindly. |
170 |
|
171 |
For version 2 of common::sense, we finally sat down a few hours and went |
172 |
through I<every single warning message>, identifying - according to |
173 |
common sense - all the useful ones. |
174 |
|
175 |
This resulted in the rather impressive list in the SYNOPSIS. When we |
176 |
weren't sure, we didn't include the warning, so the list might grow in |
177 |
the future (we might have made a mistake, too, so the list might shrink |
178 |
as well). |
179 |
|
180 |
Note the presence of C<FATAL> in the list: we do not think that the |
181 |
conditions caught by these warnings are worthy of a warning, we I<insist> |
182 |
that they are worthy of I<stopping> your program, I<instantly>. They are |
183 |
I<bugs>! |
184 |
|
185 |
Therefore we consider C<common::sense> to be much stricter than C<use |
186 |
warnings>, which is good if you are into strict things (we are not, |
187 |
actually, but these things tend to be subjective). |
188 |
|
189 |
After deciding on the list, we ran the module against all of our code that |
190 |
uses C<common::sense> (that is almost all of our code), and found only one |
191 |
occurrence where one of them caused a problem: one of elmex's (unreleased) |
192 |
modules contained: |
193 |
|
194 |
$fmt =~ s/([^\s\[]*)\[( [^\]]* )\]/\x0$1\x1$2\x0/xgo; |
195 |
|
196 |
We quickly agreed that indeed the code should be changed, even though it |
197 |
happened to do the right thing when the warning was switched off. |
198 |
|
199 |
|
200 |
=item much reduced typing |
201 |
|
202 |
Especially with version 2.0 of common::sense, the amount of boilerplate |
203 |
code you need to add to get I<this> policy is daunting. Nobody would write |
204 |
this out in throwaway scripts, commandline hacks or in quick internal-use |
205 |
scripts. |
206 |
|
207 |
By using common::sense you get a defined set of policies (ours, but maybe |
208 |
yours, too, if you accept them), and they are easy to apply to your |
209 |
scripts: typing C<use common::sense;> is even shorter than C<use warnings; |
210 |
use strict; use feature ...>. |
211 |
|
212 |
And you can immediately use the features of your installed perl, which |
213 |
is more difficult in code you release, but not usually an issue for |
214 |
internal-use code (downgrades of your production perl should be rare, |
215 |
right?). |
216 |
|
217 |
|
218 |
=item mucho reduced memory usage |
219 |
|
220 |
Just using all those pragmas mentioned in the SYNOPSIS together wastes |
221 |
<blink>I<< B<776> kilobytes >></blink> of precious memory in my perl, for |
222 |
I<every single perl process using our code>, which on our machines, is a |
223 |
lot. In comparison, this module only uses I<< B<four> >> kilobytes (I even |
224 |
had to write it out so it looks like more) of memory on the same platform. |
225 |
|
226 |
The money/time/effort/electricity invested in these gigabytes (probably |
227 |
petabytes globally!) of wasted memory could easily save 42 trees, and a |
228 |
kitten! |
229 |
|
230 |
Unfortunately, until everybody applies more common sense, there will still |
231 |
often be modules that pull in the monster pragmas. But one can hope... |
232 |
|
233 |
=back |
234 |
|
235 |
=head1 THERE IS NO 'no common::sense'!!!! !!!! !! |
236 |
|
237 |
This module doesn't offer an unimport. First of all, it wastes even more |
238 |
memory, second, and more importantly, who with even a bit of common sense |
239 |
would want no common sense? |
240 |
|
241 |
=head1 STABILITY AND FUTURE VERSIONS |
242 |
|
243 |
Future versions might change just about everything in this module. We |
244 |
might test our modules and upload new ones working with newer versions of |
245 |
this module, and leave you standing in the rain because we didn't tell |
246 |
you. In fact, we did so when switching from 1.0 to 2.0, which enabled gobs |
247 |
of warnings, and made them FATAL on top. |
248 |
|
249 |
Maybe we will load some nifty modules that try to emulate C<say> or so |
250 |
with perls older than 5.10 (this module, of course, should work with older |
251 |
perl versions - supporting 5.8 for example is just common sense at this |
252 |
time. Maybe not in the future, but of course you can trust our common |
253 |
sense to be consistent with, uhm, our opinion). |
254 |
|
255 |
=head1 WHAT OTHER PEOPLE HAD TO SAY ABOUT THIS MODULE |
256 |
|
257 |
apeiron |
258 |
|
259 |
"... wow" |
260 |
"I hope common::sense is a joke." |
261 |
|
262 |
crab |
263 |
|
264 |
"i wonder how it would be if joerg schilling wrote perl modules." |
265 |
|
266 |
Adam Kennedy |
267 |
|
268 |
"Very interesting, efficient, and potentially something I'd use all the time." |
269 |
[...] |
270 |
"So no common::sense for me, alas." |
271 |
|
272 |
H.Merijn Brand |
273 |
|
274 |
"Just one more reason to drop JSON::XS from my distribution list" |
275 |
|
276 |
Pista Palo |
277 |
|
278 |
"Something in short supply these days..." |
279 |
|
280 |
Steffen Schwigon |
281 |
|
282 |
"This module is quite for sure *not* just a repetition of all the other |
283 |
'use strict, use warnings'-approaches, and it's also not the opposite. |
284 |
[...] And for its chosen middle-way it's also not the worst name ever. |
285 |
And everything is documented." |
286 |
|
287 |
BKB |
288 |
|
289 |
"[Deleted - thanks to Steffen Schwigon for pointing out this review was |
290 |
in error.]" |
291 |
|
292 |
Somni |
293 |
|
294 |
"the arrogance of the guy" |
295 |
"I swear he tacked somenoe else's name onto the module |
296 |
just so he could use the royal 'we' in the documentation" |
297 |
|
298 |
Anonymous Monk |
299 |
|
300 |
"You just gotta love this thing, its got META.json!!!" |
301 |
|
302 |
dngor |
303 |
|
304 |
"Heh. '"<elmex at ta-sa.org>"' The quotes are semantic |
305 |
distancing from that e-mail address." |
306 |
|
307 |
Jerad Pierce |
308 |
|
309 |
"Awful name (not a proper pragma), and the SYNOPSIS doesn't tell you |
310 |
anything either. Nor is it clear what features have to do with "common |
311 |
sense" or discipline." |
312 |
|
313 |
acme |
314 |
|
315 |
"THERE IS NO 'no common::sense'!!!! !!!! !!" |
316 |
|
317 |
apeiron (meta-comment about us commenting^Wquoting his comment) |
318 |
|
319 |
"How about quoting this: get a clue, you fucktarded amoeba." |
320 |
|
321 |
quanth |
322 |
|
323 |
"common sense is beautiful, json::xs is fast, Anyevent, EV are fast and |
324 |
furious. I love mlehmannware ;)" |
325 |
|
326 |
apeiron |
327 |
|
328 |
"... it's mlehmann's view of what common sense is. His view of common |
329 |
sense is certainly uncommon, insofar as anyone with a clue disagrees |
330 |
with him." |
331 |
|
332 |
apeiron (another meta-comment) |
333 |
|
334 |
"apeiron wonders if his little informant is here to steal more quotes" |
335 |
|
336 |
ew73 |
337 |
|
338 |
"... I never got past the SYNOPSIS before calling it shit." |
339 |
[...] |
340 |
How come no one ever quotes me. :(" |
341 |
|
342 |
chip (not willing to explain his cryptic questions about links in Changes files) |
343 |
|
344 |
"I'm willing to ask the question I've asked. I'm not willing to go |
345 |
through the whole dance you apparently have choreographed. Either |
346 |
answer the completely obvious question, or tell me to fuck off again." |
347 |
|
348 |
=head1 FREQUENTLY ASKED QUESTIONS |
349 |
|
350 |
Or frequently-come-up confusions. |
351 |
|
352 |
=over 4 |
353 |
|
354 |
=item Is this module meant to be serious? |
355 |
|
356 |
Yes, we would have put it under the C<Acme::> namespace otherwise. |
357 |
|
358 |
=item But the manpage is written in a funny/stupid/... way? |
359 |
|
360 |
This was meant to make it clear that our common sense is a subjective |
361 |
thing and other people can use their own notions, taking the steam out |
362 |
of anybody who might be offended (as some people are always offended no |
363 |
matter what you do). |
364 |
|
365 |
This was a failure. |
366 |
|
367 |
But we hope the manpage still is somewhat entertaining even though it |
368 |
explains boring rationale. |
369 |
|
370 |
=item Why do you impose your conventions on my code? |
371 |
|
372 |
For some reason people keep thinking that C<common::sense> imposes |
373 |
process-wide limits, even though the SYNOPSIS makes it clear that it works |
374 |
like other similar modules - i.e. only within the scope that C<use>s them. |
375 |
|
376 |
So, no, we don't - nobody is forced to use this module, and using a module |
377 |
that relies on common::sense does not impose anything on you. |
378 |
|
379 |
=item Why do you think only your notion of common::sense is valid? |
380 |
|
381 |
Well, we don't, and have clearly written this in the documentation to |
382 |
every single release. We were just faster than anybody else w.r.t. to |
383 |
grabbing the namespace. |
384 |
|
385 |
=item But everybody knows that you have to use strict and use warnings, |
386 |
why do you disable them? |
387 |
|
388 |
Well, we don't do this either - we selectively disagree with the |
389 |
usefulness of some warnings over others. This module is aimed at |
390 |
experienced Perl programmers, not people migrating from other languages |
391 |
who might be surprised about stuff such as C<undef>. On the other hand, |
392 |
this does not exclude the usefulness of this module for total newbies, due |
393 |
to its strictness in enforcing policy, while at the same time not limiting |
394 |
the expressive power of perl. |
395 |
|
396 |
This module is considerably I<more> strict than the canonical C<use |
397 |
strict; use warnings>, as it makes all its warnings fatal in nature, so |
398 |
you can not get away with as many things as with the canonical approach. |
399 |
|
400 |
This was not implemented in version 1.0 because of the daunting number |
401 |
of warning categories and the difficulty in getting exactly the set of |
402 |
warnings you wish (i.e. look at the SYNOPSIS in how complicated it is to |
403 |
get a specific set of warnings - it is not reasonable to put this into |
404 |
every module, the maintenance effort would be enormous). |
405 |
|
406 |
=item But many modules C<use strict> or C<use warnings>, so the memory |
407 |
savings do not apply? |
408 |
|
409 |
I suddenly feel sad... |
410 |
|
411 |
But yes, that's true. Fortunately C<common::sense> still uses only a |
412 |
miniscule amount of RAM. |
413 |
|
414 |
=item But it adds another dependency to your modules! |
415 |
|
416 |
It's a fact, yeah. But it's trivial to install, most popular modules have |
417 |
many more dependencies and we consider dependencies a good thing - it |
418 |
leads to better APIs, more thought about interworking of modules and so |
419 |
on. |
420 |
|
421 |
=item Why do you use JSON and not YAML for your META.yml? |
422 |
|
423 |
This is not true - YAML supports a large subset of JSON, and this subset |
424 |
is what META.yml is written in, so it would be correct to say "the |
425 |
META.yml is written in a common subset of YAML and JSON". |
426 |
|
427 |
The META.yml follows the YAML, JSON and META.yml specifications, and is |
428 |
correctly parsed by CPAN, so if you have trouble with it, the problem is |
429 |
likely on your side. |
430 |
|
431 |
=item But! But! |
432 |
|
433 |
Yeah, we know. |
434 |
|
435 |
=back |
436 |
|
437 |
=head1 AUTHOR |
438 |
|
439 |
Marc Lehmann <schmorp@schmorp.de> |
440 |
http://home.schmorp.de/ |
441 |
|
442 |
Robin Redeker, "<elmex at ta-sa.org>". |
443 |
|
444 |
=cut |
445 |
|