1perlfilter(3) User Contributed Perl Documentation perlfilter(3)
2
6 perlfilter - Source Filters
7
9 This article is about a little-known feature of Perl called source
10 filters. Source filters alter the program text of a module before Perl
11 sees it, much as a C preprocessor alters the source text of a C program
12 before the compiler sees it. This article tells you more about what
13 source filters are, how they work, and how to write your own.
14 The original purpose of source filters was to let you encrypt your
16 program source to prevent casual piracy. This isn't all they can do, as
17 you'll soon learn. But first, the basics.
18
20 Before the Perl interpreter can execute a Perl script, it must first
21 read it from a file into memory for parsing and compilation. If that
22 script itself includes other scripts with a "use" or "require"
23 statement, then each of those scripts will have to be read from their
24 respective files as well.
25 Now think of each logical connection between the Perl parser and an
27 individual file as a source stream. A source stream is created when the
28 Perl parser opens a file, it continues to exist as the source code is
29 read into memory, and it is destroyed when Perl is finished parsing the
30 file. If the parser encounters a "require" or "use" statement in a
31 source stream, a new and distinct stream is created just for that file.
32 The diagram below represents a single source stream, with the flow of
34 source from a Perl script file on the left into the Perl parser on the
35 right. This is how Perl normally operates.
36 file -------> parser
38 There are two important points to remember:
40 1. Although there can be any number of source streams in existence at
42 any given time, only one will be active.
43 2. Every source stream is associated with only one file.
45 A source filter is a special kind of Perl module that intercepts and
47 modifies a source stream before it reaches the parser. A source filter
48 changes our diagram like this:
49 file ----> filter ----> parser
51 If that doesn't make much sense, consider the analogy of a command
53 pipeline. Say you have a shell script stored in the compressed file
54 trial.gz. The simple pipeline command below runs the script without
55 needing to create a temporary file to hold the uncompressed file.
56 gunzip -c trial.gz | sh
58 In this case, the data flow from the pipeline can be represented as
60 follows:
61 trial.gz ----> gunzip ----> sh
63 With source filters, you can store the text of your script compressed
65 and use a source filter to uncompress it for Perl's parser:
66 compressed gunzip
68 Perl program ---> source filter ---> parser
69
71 So how do you use a source filter in a Perl script? Above, I said that
72 a source filter is just a special kind of module. Like all Perl
73 modules, a source filter is invoked with a use statement.
74 Say you want to pass your Perl source through the C preprocessor before
76 execution. As it happens, the source filters distribution comes with a
77 C preprocessor filter module called Filter::cpp.
78 Below is an example program, "cpp_test", which makes use of this
80 filter. Line numbers have been added to allow specific lines to be
81 referenced easily.
82 1: use Filter::cpp;
84 2: #define TRUE 1
85 3: $a = TRUE;
86 4: print "a = $a\n";
87 When you execute this script, Perl creates a source stream for the
89 file. Before the parser processes any of the lines from the file, the
90 source stream looks like this:
91 cpp_test ---------> parser
93 Line 1, "use Filter::cpp", includes and installs the "cpp" filter
95 module. All source filters work this way. The use statement is compiled
96 and executed at compile time, before any more of the file is read, and
97 it attaches the cpp filter to the source stream behind the scenes. Now
98 the data flow looks like this:
99 cpp_test ----> cpp filter ----> parser
101 As the parser reads the second and subsequent lines from the source
103 stream, it feeds those lines through the "cpp" source filter before
104 processing them. The "cpp" filter simply passes each line through the
105 real C preprocessor. The output from the C preprocessor is then
106 inserted back into the source stream by the filter.
107 .-> cpp --.
109 | |
110 | |
111 | <-'
112 cpp_test ----> cpp filter ----> parser
113 The parser then sees the following code:
115 use Filter::cpp;
117 $a = 1;
118 print "a = $a\n";
119 Let's consider what happens when the filtered code includes another
121 module with use:
122 1: use Filter::cpp;
124 2: #define TRUE 1
125 3: use Fred;
126 4: $a = TRUE;
127 5: print "a = $a\n";
128 The "cpp" filter does not apply to the text of the Fred module, only to
130 the text of the file that used it ("cpp_test"). Although the use
131 statement on line 3 will pass through the cpp filter, the module that
132 gets included ("Fred") will not. The source streams look like this
133 after line 3 has been parsed and before line 4 is parsed:
134 cpp_test ---> cpp filter ---> parser (INACTIVE)
136 Fred.pm ----> parser
138 As you can see, a new stream has been created for reading the source
140 from "Fred.pm". This stream will remain active until all of "Fred.pm"
141 has been parsed. The source stream for "cpp_test" will still exist, but
142 is inactive. Once the parser has finished reading Fred.pm, the source
143 stream associated with it will be destroyed. The source stream for
144 "cpp_test" then becomes active again and the parser reads line 4 and
145 subsequent lines from "cpp_test".
146 You can use more than one source filter on a single file. Similarly,
148 you can reuse the same filter in as many files as you like.
149 For example, if you have a uuencoded and compressed source file, it is
151 possible to stack a uudecode filter and an uncompression filter like
152 this:
153 use Filter::uudecode; use Filter::uncompress;
155 M'XL(".H<US4''V9I;F%L')Q;>7/;1I;_>_I3=&E=%:F*I"T?22Q/
156 M6]9*<IQCO*XFT"0[PL%%'Y+IG?WN^ZYN-$'J.[.JE$,20/?K=_[>
157 ...
158 Once the first line has been processed, the flow will look like this:
160 file ---> uudecode ---> uncompress ---> parser
162 filter filter
163 Data flows through filters in the same order they appear in the source
165 file. The uudecode filter appeared before the uncompress filter, so the
166 source file will be uudecoded before it's uncompressed.
167
169 There are three ways to write your own source filter. You can write it
170 in C, use an external program as a filter, or write the filter in Perl.
171 I won't cover the first two in any great detail, so I'll get them out
172 of the way first. Writing the filter in Perl is most convenient, so
173 I'll devote the most space to it.
174
176 The first of the three available techniques is to write the filter
177 completely in C. The external module you create interfaces directly
178 with the source filter hooks provided by Perl.
179 The advantage of this technique is that you have complete control over
181 the implementation of your filter. The big disadvantage is the
182 increased complexity required to write the filter - not only do you
183 need to understand the source filter hooks, but you also need a
184 reasonable knowledge of Perl guts. One of the few times it is worth
185 going to this trouble is when writing a source scrambler. The "decrypt"
186 filter (which unscrambles the source before Perl parses it) included
187 with the source filter distribution is an example of a C source filter
188 (see Decryption Filters, below).
189 Decryption Filters
191 All decryption filters work on the principle of "security through
192 obscurity." Regardless of how well you write a decryption filter
193 and how strong your encryption algorithm is, anyone determined
194 enough can retrieve the original source code. The reason is quite
195 simple - once the decryption filter has decrypted the source back
196 to its original form, fragments of it will be stored in the
197 computer's memory as Perl parses it. The source might only be in
198 memory for a short period of time, but anyone possessing a
199 debugger, skill, and lots of patience can eventually reconstruct
200 your program.
201 That said, there are a number of steps that can be taken to make
203 life difficult for the potential cracker. The most important:
204 Write your decryption filter in C and statically link the
205 decryption module into the Perl binary. For further tips to make
206 life difficult for the potential cracker, see the file decrypt.pm
207 in the source filters distribution.
208
210 An alternative to writing the filter in C is to create a separate
211 executable in the language of your choice. The separate executable
212 reads from standard input, does whatever processing is necessary, and
213 writes the filtered data to standard output. "Filter::cpp" is an
214 example of a source filter implemented as a separate executable - the
215 executable is the C preprocessor bundled with your C compiler.
216 The source filter distribution includes two modules that simplify this
218 task: "Filter::exec" and "Filter::sh". Both allow you to run any
219 external executable. Both use a coprocess to control the flow of data
220 into and out of the external executable. (For details on coprocesses,
221 see Stephens, W.R., "Advanced Programming in the UNIX Environment."
222 Addison-Wesley, ISBN 0-210-56317-7, pages 441-445.) The difference
223 between them is that "Filter::exec" spawns the external command
224 directly, while "Filter::sh" spawns a shell to execute the external
225 command. (Unix uses the Bourne shell; NT uses the cmd shell.) Spawning
226 a shell allows you to make use of the shell metacharacters and
227 redirection facilities.
228 Here is an example script that uses "Filter::sh":
230 use Filter::sh 'tr XYZ PQR';
232 $a = 1;
233 print "XYZ a = $a\n";
234 The output you'll get when the script is executed:
236 PQR a = 1
238 Writing a source filter as a separate executable works fine, but a
240 small performance penalty is incurred. For example, if you execute the
241 small example above, a separate subprocess will be created to run the
242 Unix "tr" command. Each use of the filter requires its own subprocess.
243 If creating subprocesses is expensive on your system, you might want to
244 consider one of the other options for creating source filters.
245
247 The easiest and most portable option available for creating your own
248 source filter is to write it completely in Perl. To distinguish this
249 from the previous two techniques, I'll call it a Perl source filter.
250 To help understand how to write a Perl source filter we need an example
252 to study. Here is a complete source filter that performs rot13
253 decoding. (Rot13 is a very simple encryption scheme used in Usenet
254 postings to hide the contents of offensive posts. It moves every letter
255 forward thirteen places, so that A becomes N, B becomes O, and Z
256 becomes M.)
257 package Rot13;
259 use Filter::Util::Call;
261 sub import {
263 my ($type) = @_;
264 my ($ref) = [];
265 filter_add(bless $ref);
266 }
267 sub filter {
269 my ($self) = @_;
270 my ($status);
271 tr/n-za-mN-ZA-M/a-zA-Z/
273 if ($status = filter_read()) > 0;
274 $status;
275 }
276 1;
278 All Perl source filters are implemented as Perl classes and have the
280 same basic structure as the example above.
281 First, we include the "Filter::Util::Call" module, which exports a
283 number of functions into your filter's namespace. The filter shown
284 above uses two of these functions, "filter_add()" and "filter_read()".
285 Next, we create the filter object and associate it with the source
287 stream by defining the "import" function. If you know Perl well enough,
288 you know that "import" is called automatically every time a module is
289 included with a use statement. This makes "import" the ideal place to
290 both create and install a filter object.
291 In the example filter, the object ($ref) is blessed just like any other
293 Perl object. Our example uses an anonymous array, but this isn't a
294 requirement. Because this example doesn't need to store any context
295 information, we could have used a scalar or hash reference just as
296 well. The next section demonstrates context data.
297 The association between the filter object and the source stream is made
299 with the "filter_add()" function. This takes a filter object as a
300 parameter ($ref in this case) and installs it in the source stream.
301 Finally, there is the code that actually does the filtering. For this
303 type of Perl source filter, all the filtering is done in a method
304 called "filter()". (It is also possible to write a Perl source filter
305 using a closure. See the "Filter::Util::Call" manual page for more
306 details.) It's called every time the Perl parser needs another line of
307 source to process. The "filter()" method, in turn, reads lines from the
308 source stream using the "filter_read()" function.
309 If a line was available from the source stream, "filter_read()" returns
311 a status value greater than zero and appends the line to $_. A status
312 value of zero indicates end-of-file, less than zero means an error. The
313 filter function itself is expected to return its status in the same
314 way, and put the filtered line it wants written to the source stream in
315 $_. The use of $_ accounts for the brevity of most Perl source filters.
316 In order to make use of the rot13 filter we need some way of encoding
318 the source file in rot13 format. The script below, "mkrot13", does just
319 that.
320 die "usage mkrot13 filename\n" unless @ARGV;
322 my $in = $ARGV[0];
323 my $out = "$in.tmp";
324 open(IN, "<$in") or die "Cannot open file $in: $!\n";
325 open(OUT, ">$out") or die "Cannot open file $out: $!\n";
326 print OUT "use Rot13;\n";
328 while (<IN>) {
329 tr/a-zA-Z/n-za-mN-ZA-M/;
330 print OUT;
331 }
332 close IN;
334 close OUT;
335 unlink $in;
336 rename $out, $in;
337 If we encrypt this with "mkrot13":
339 print " hello fred \n";
341 the result will be this:
343 use Rot13;
345 cevag "uryyb serq\a";
346 Running it produces this output:
348 hello fred
350
352 The rot13 example was a trivial example. Here's another demonstration
353 that shows off a few more features.
354 Say you wanted to include a lot of debugging code in your Perl script
356 during development, but you didn't want it available in the released
357 product. Source filters offer a solution. In order to keep the example
358 simple, let's say you wanted the debugging output to be controlled by
359 an environment variable, "DEBUG". Debugging code is enabled if the
360 variable exists, otherwise it is disabled.
361 Two special marker lines will bracket debugging code, like this:
363 ## DEBUG_BEGIN
365 if ($year > 1999) {
366 warn "Debug: millennium bug in year $year\n";
367 }
368 ## DEBUG_END
369 The filter ensures that Perl parses the code between the <DEBUG_BEGIN>
371 and "DEBUG_END" markers only when the "DEBUG" environment variable
372 exists. That means that when "DEBUG" does exist, the code above should
373 be passed through the filter unchanged. The marker lines can also be
374 passed through as-is, because the Perl parser will see them as comment
375 lines. When "DEBUG" isn't set, we need a way to disable the debug code.
376 A simple way to achieve that is to convert the lines between the two
377 markers into comments:
378 ## DEBUG_BEGIN
380 #if ($year > 1999) {
381 # warn "Debug: millennium bug in year $year\n";
382 #}
383 ## DEBUG_END
384 Here is the complete Debug filter:
386 package Debug;
388 use strict;
390 use warnings;
391 use Filter::Util::Call;
392 use constant TRUE => 1;
394 use constant FALSE => 0;
395 sub import {
397 my ($type) = @_;
398 my (%context) = (
399 Enabled => defined $ENV{DEBUG},
400 InTraceBlock => FALSE,
401 Filename => (caller)[1],
402 LineNo => 0,
403 LastBegin => 0,
404 );
405 filter_add(bless \%context);
406 }
407 sub Die {
409 my ($self) = shift;
410 my ($message) = shift;
411 my ($line_no) = shift || $self->{LastBegin};
412 die "$message at $self->{Filename} line $line_no.\n"
413 }
414 sub filter {
416 my ($self) = @_;
417 my ($status);
418 $status = filter_read();
419 ++ $self->{LineNo};
420 # deal with EOF/error first
422 if ($status <= 0) {
423 $self->Die("DEBUG_BEGIN has no DEBUG_END")
424 if $self->{InTraceBlock};
425 return $status;
426 }
427 if ($self->{InTraceBlock}) {
429 if (/^\s*##\s*DEBUG_BEGIN/ ) {
430 $self->Die("Nested DEBUG_BEGIN", $self->{LineNo})
431 } elsif (/^\s*##\s*DEBUG_END/) {
432 $self->{InTraceBlock} = FALSE;
433 }
434 # comment out the debug lines when the filter is disabled
436 s/^/#/ if ! $self->{Enabled};
437 } elsif ( /^\s*##\s*DEBUG_BEGIN/ ) {
438 $self->{InTraceBlock} = TRUE;
439 $self->{LastBegin} = $self->{LineNo};
440 } elsif ( /^\s*##\s*DEBUG_END/ ) {
441 $self->Die("DEBUG_END has no DEBUG_BEGIN", $self->{LineNo});
442 }
443 return $status;
444 }
445 1;
447 The big difference between this filter and the previous example is the
449 use of context data in the filter object. The filter object is based on
450 a hash reference, and is used to keep various pieces of context
451 information between calls to the filter function. All but two of the
452 hash fields are used for error reporting. The first of those two,
453 Enabled, is used by the filter to determine whether the debugging code
454 should be given to the Perl parser. The second, InTraceBlock, is true
455 when the filter has encountered a "DEBUG_BEGIN" line, but has not yet
456 encountered the following "DEBUG_END" line.
457 If you ignore all the error checking that most of the code does, the
459 essence of the filter is as follows:
460 sub filter {
462 my ($self) = @_;
463 my ($status);
464 $status = filter_read();
465 # deal with EOF/error first
467 return $status if $status <= 0;
468 if ($self->{InTraceBlock}) {
469 if (/^\s*##\s*DEBUG_END/) {
470 $self->{InTraceBlock} = FALSE
471 }
472 # comment out debug lines when the filter is disabled
474 s/^/#/ if ! $self->{Enabled};
475 } elsif ( /^\s*##\s*DEBUG_BEGIN/ ) {
476 $self->{InTraceBlock} = TRUE;
477 }
478 return $status;
479 }
480 Be warned: just as the C-preprocessor doesn't know C, the Debug filter
482 doesn't know Perl. It can be fooled quite easily:
483 print <<EOM;
485 ##DEBUG_BEGIN
486 EOM
487 Such things aside, you can see that a lot can be achieved with a modest
489 amount of code.
490
492 You now have better understanding of what a source filter is, and you
493 might even have a possible use for them. If you feel like playing with
494 source filters but need a bit of inspiration, here are some extra
495 features you could add to the Debug filter.
496 First, an easy one. Rather than having debugging code that is all-or-
498 nothing, it would be much more useful to be able to control which
499 specific blocks of debugging code get included. Try extending the
500 syntax for debug blocks to allow each to be identified. The contents of
501 the "DEBUG" environment variable can then be used to control which
502 blocks get included.
503 Once you can identify individual blocks, try allowing them to be
505 nested. That isn't difficult either.
506 Here is an interesting idea that doesn't involve the Debug filter.
508 Currently Perl subroutines have fairly limited support for formal
509 parameter lists. You can specify the number of parameters and their
510 type, but you still have to manually take them out of the @_ array
511 yourself. Write a source filter that allows you to have a named
512 parameter list. Such a filter would turn this:
513 sub MySub ($first, $second, @rest) { ... }
515 into this:
517 sub MySub($$@) {
519 my ($first) = shift;
520 my ($second) = shift;
521 my (@rest) = @_;
522 ...
523 }
524 Finally, if you feel like a real challenge, have a go at writing a
526 full-blown Perl macro preprocessor as a source filter. Borrow the
527 useful features from the C preprocessor and any other macro processors
528 you know. The tricky bit will be choosing how much knowledge of Perl's
529 syntax you want your filter to have.
530
532 Source filters only work on the string level, thus are highly limited
533 in its ability to change source code on the fly. It cannot detect
534 comments, quoted strings, heredocs, it is no replacement for a real
535 parser. The only stable usage for source filters are encryption,
536 compression, or the byteloader, to translate binary code back to source
537 code.
538 See for example the limitations in Switch, which uses source filters,
540 and thus is does not work inside a string eval, the presence of regexes
541 with embedded newlines that are specified with raw "/.../" delimiters
542 and don't have a modifier "//x" are indistinguishable from code chunks
543 beginning with the division operator "/". As a workaround you must use
544 "m/.../" or "m?...?" for such patterns. Also, the presence of regexes
545 specified with raw "?...?" delimiters may cause mysterious errors. The
546 workaround is to use "m?...?" instead. See
547 <https://metacpan.org/pod/Switch#LIMITATIONS>.
548 Currently the content of the "__DATA__" block is not filtered.
550 Currently internal buffer lengths are limited to 32-bit only.
552
554 Some Filters Clobber the "DATA" Handle
555 Some source filters use the "DATA" handle to read the calling
556 program. When using these source filters you cannot rely on this
557 handle, nor expect any particular kind of behavior when operating
558 on it. Filters based on Filter::Util::Call (and therefore
559 Filter::Simple) do not alter the "DATA" filehandle, but on the
560 other hand totally ignore the text after "__DATA__".
561
563 The Source Filters distribution is available on CPAN, in
564 CPAN/modules/by-module/Filter
566 Starting from Perl 5.8 Filter::Util::Call (the core part of the Source
568 Filters distribution) is part of the standard Perl distribution. Also
569 included is a friendlier interface called Filter::Simple, by Damian
570 Conway.
571
573 Paul Marquess <Paul.Marquess@btinternet.com>
574 Reini Urban <rurban@cpan.org>
576
578 The first version of this article originally appeared in The Perl
579 Journal #11, and is copyright 1998 The Perl Journal. It appears
580 courtesy of Jon Orwant and The Perl Journal. This document may be
581 distributed under the same terms as Perl itself.
582perl v5.32.1 2021-01-27 perlfilter(3)