5 Consistent formatting of this file is achieved with:
6 perl ./Porting/podtidy pod/perlhacktips.pod
10 perlhacktips - Tips for Perl core C code hacking
14 This document will help you learn the best way to go about hacking on
15 the Perl core C code. It covers common problems, debugging, profiling,
18 If you haven't read L<perlhack> and L<perlhacktut> yet, you might want
21 =head1 COMMON PROBLEMS
23 Perl source plays by ANSI C89 rules: no C99 (or C++) extensions.
24 You don't care about some particular platform having broken Perl? I
25 hear there is still a strong demand for J2EE programmers.
27 =head2 Perl environment problems
33 Not compiling with threading
35 Compiling with threading (-Duseithreads) completely rewrites the
36 function prototypes of Perl. You better try your changes with that.
37 Related to this is the difference between "Perl_-less" and "Perl_-ly"
40 Perl_sv_setiv(aTHX_ ...);
43 The first one explicitly passes in the context, which is needed for
44 e.g. threaded builds. The second one does that implicitly; do not get
45 them mixed. If you are not passing in a aTHX_, you will need to do a
46 dTHX (or a dVAR) as the first thing in the function.
48 See L<perlguts/"How multiple interpreters and concurrency are
49 supported"> for further discussion about context.
53 Not compiling with -DDEBUGGING
55 The DEBUGGING define exposes more code to the compiler, therefore more
56 ways for things to go wrong. You should try it.
60 Introducing (non-read-only) globals
62 Do not introduce any modifiable globals, truly global or file static.
63 They are bad form and complicate multithreading and other forms of
64 concurrency. The right way is to introduce them as new interpreter
65 variables, see F<intrpvar.h> (at the very end for binary
68 Introducing read-only (const) globals is okay, as long as you verify
69 with e.g. C<nm libperl.a|egrep -v ' [TURtr] '> (if your C<nm> has
70 BSD-style output) that the data you added really is read-only. (If it
71 is, it shouldn't show up in the output of that command.)
73 If you want to have static strings, make them constant:
75 static const char etc[] = "...";
77 If you want to have arrays of constant strings, note carefully the
78 right combination of C<const>s:
80 static const char * const yippee[] =
81 {"hi", "ho", "silver"};
83 There is a way to completely hide any modifiable globals (they are all
84 moved to heap), the compilation setting
85 C<-DPERL_GLOBAL_STRUCT_PRIVATE>. It is not normally used, but can be
86 used for testing, read more about it in L<perlguts/"Background and
87 PERL_IMPLICIT_CONTEXT">.
91 Not exporting your new function
93 Some platforms (Win32, AIX, VMS, OS/2, to name a few) require any
94 function that is part of the public API (the shared Perl library) to be
95 explicitly marked as exported. See the discussion about F<embed.pl> in
100 Exporting your new function
102 The new shiny result of either genuine new functionality or your
103 arduous refactoring is now ready and correctly exported. So what could
106 Maybe simply that your function did not need to be exported in the
107 first place. Perl has a long and not so glorious history of exporting
108 functions that it should not have.
110 If the function is used only inside one source code file, make it
111 static. See the discussion about F<embed.pl> in L<perlguts>.
113 If the function is used across several files, but intended only for
114 Perl's internal use (and this should be the common case), do not export
115 it to the public API. See the discussion about F<embed.pl> in
120 =head2 Portability problems
122 The following are common causes of compilation and/or execution
123 failures, not common to Perl as such. The C FAQ is good bedtime
124 reading. Please test your changes with as many C compilers and
125 platforms as possible; we will, anyway, and it's nice to save oneself
126 from public embarrassment.
128 If using gcc, you can add the C<-std=c89> option which will hopefully
129 catch most of these unportabilities. (However it might also catch
130 incompatibilities in your system's header files.)
132 Use the Configure C<-Dgccansipedantic> flag to enable the gcc C<-ansi
133 -pedantic> flags which enforce stricter ANSI rules.
135 If using the C<gcc -Wall> note that not all the possible warnings (like
136 C<-Wuninitialized>) are given unless you also compile with C<-O>.
138 Note that if using gcc, starting from Perl 5.9.5 the Perl core source
139 code files (the ones at the top level of the source code distribution,
140 but not e.g. the extensions under ext/) are automatically compiled with
141 as many as possible of the C<-std=c89>, C<-ansi>, C<-pedantic>, and a
142 selection of C<-W> flags (see cflags.SH).
144 Also study L<perlport> carefully to avoid any bad assumptions about the
145 operating system, filesystems, character set, and so forth.
147 You may once in a while try a "make microperl" to see whether we can
148 still compile Perl with just the bare minimum of interfaces. (See
151 Do not assume an operating system indicates a certain compiler.
157 Casting pointers to integers or casting integers to pointers
169 Both are bad, and broken, and unportable. Use the PTR2IV() macro that
170 does it right. (Likewise, there are PTR2UV(), PTR2NV(), INT2PTR(), and
175 Casting between function pointers and data pointers
177 Technically speaking casting between function pointers and data
178 pointers is unportable and undefined, but practically speaking it seems
179 to work, but you should use the FPTR2DPTR() and DPTR2FPTR() macros.
180 Sometimes you can also play games with unions.
184 Assuming sizeof(int) == sizeof(long)
186 There are platforms where longs are 64 bits, and platforms where ints
187 are 64 bits, and while we are out to shock you, even platforms where
188 shorts are 64 bits. This is all legal according to the C standard. (In
189 other words, "long long" is not a portable way to specify 64 bits, and
190 "long long" is not even guaranteed to be any wider than "long".)
192 Instead, use the definitions IV, UV, IVSIZE, I32SIZE, and so forth.
193 Avoid things like I32 because they are B<not> guaranteed to be
194 I<exactly> 32 bits, they are I<at least> 32 bits, nor are they
195 guaranteed to be B<int> or B<long>. If you really explicitly need
196 64-bit variables, use I64 and U64, but only if guarded by HAS_QUAD.
200 Assuming one can dereference any type of pointer for any type of data
203 long pony = *(long *)p; /* BAD */
205 Many platforms, quite rightly so, will give you a core dump instead of
206 a pony if the p happens not to be correctly aligned.
212 (int)*p = ...; /* BAD */
214 Simply not portable. Get your lvalue to be of the right type, or maybe
215 use temporary variables, or dirty tricks with unions.
219 Assume B<anything> about structs (especially the ones you don't
220 control, like the ones coming from the system headers)
226 That a certain field exists in a struct
230 That no other fields exist besides the ones you know of
234 That a field is of certain signedness, sizeof, or type
238 That the fields are in a certain order
244 While C guarantees the ordering specified in the struct definition,
245 between different platforms the definitions might differ
251 That the sizeof(struct) or the alignments are the same everywhere
257 There might be padding bytes between the fields to align the fields -
258 the bytes can be anything
262 Structs are required to be aligned to the maximum alignment required by
263 the fields - which for native types is for usually equivalent to
264 sizeof() of the field
272 Assuming the character set is ASCIIish
274 Perl can compile and run under EBCDIC platforms. See L<perlebcdic>.
275 This is transparent for the most part, but because the character sets
276 differ, you shouldn't use numeric (decimal, octal, nor hex) constants
277 to refer to characters. You can safely say C<'A'>, but not C<0x41>.
278 You can safely say C<'\n'>, but not C<\012>. However, you can use
279 macros defined in F<utf8.h> to specify any code point portably.
280 C<LATIN1_TO_NATIVE(0xDF)> is going to be the code point that means
281 LATIN SMALL LETTER SHARP S on whatever platform you are running on (on
282 ASCII platforms it compiles without adding any extra code, so there is
283 zero performance hit on those). The acceptable inputs to
284 C<LATIN1_TO_NATIVE> are from C<0x00> through C<0xFF>. If your input
285 isn't guaranteed to be in that range, use C<UNICODE_TO_NATIVE> instead.
286 C<NATIVE_TO_LATIN1> and C<NATIVE_TO_UNICODE> translate the opposite
289 If you need the string representation of a character that doesn't have a
290 mnemonic name in C, you should add it to the list in
291 F<regen/unicode_constants.pl>, and have Perl create C<#define>'s for you,
292 based on the current platform.
294 Note that the C<isI<FOO>> and C<toI<FOO>> macros in F<handy.h> work
295 properly on native code points and strings.
297 Also, the range 'A' - 'Z' in ASCII is an unbroken sequence of 26 upper
298 case alphabetic characters. That is not true in EBCDIC. Nor for 'a' to
299 'z'. But '0' - '9' is an unbroken range in both systems. Don't assume
300 anything about other ranges. (Note that special handling of ranges in
301 regular expression patterns and transliterations makes it appear to Perl
302 code that the aforementioned ranges are all unbroken.)
304 Many of the comments in the existing code ignore the possibility of
305 EBCDIC, and may be wrong therefore, even if the code works. This is
306 actually a tribute to the successful transparent insertion of being
307 able to handle EBCDIC without having to change pre-existing code.
309 UTF-8 and UTF-EBCDIC are two different encodings used to represent
310 Unicode code points as sequences of bytes. Macros with the same names
311 (but different definitions) in F<utf8.h> and F<utfebcdic.h> are used to
312 allow the calling code to think that there is only one such encoding.
313 This is almost always referred to as C<utf8>, but it means the EBCDIC
314 version as well. Again, comments in the code may well be wrong even if
315 the code itself is right. For example, the concept of UTF-8 C<invariant
316 characters> differs between ASCII and EBCDIC. On ASCII platforms, only
317 characters that do not have the high-order bit set (i.e. whose ordinals
318 are strict ASCII, 0 - 127) are invariant, and the documentation and
319 comments in the code may assume that, often referring to something
320 like, say, C<hibit>. The situation differs and is not so simple on
321 EBCDIC machines, but as long as the code itself uses the
322 C<NATIVE_IS_INVARIANT()> macro appropriately, it works, even if the
325 As noted in L<perlhack/TESTING>, when writing test scripts, the file
326 F<t/charset_tools.pl> contains some helpful functions for writing tests
327 valid on both ASCII and EBCDIC platforms. Sometimes, though, a test
328 can't use a function and it's inconvenient to have different test
329 versions depending on the platform. There are 20 code points that are
330 the same in all 4 character sets currently recognized by Perl (the 3
331 EBCDIC code pages plus ISO 8859-1 (ASCII/Latin1)). These can be used in
332 such tests, though there is a small possibility that Perl will become
333 available in yet another character set, breaking your test. All but one
334 of these code points are C0 control characters. The most significant
335 controls that are the same are C<\0>, C<\r>, and C<\N{VT}> (also
336 specifiable as C<\cK>, C<\x0B>, C<\N{U+0B}>, or C<\013>). The single
337 non-control is U+00B6 PILCROW SIGN. The controls that are the same have
338 the same bit pattern in all 4 character sets, regardless of the UTF8ness
339 of the string containing them. The bit pattern for U+B6 is the same in
340 all 4 for non-UTF8 strings, but differs in each when its containing
341 string is UTF-8 encoded. The only other code points that have some sort
342 of sameness across all 4 character sets are the pair 0xDC and 0xFC.
343 Together these represent upper- and lowercase LATIN LETTER U WITH
344 DIAERESIS, but which is upper and which is lower may be reversed: 0xDC
345 is the capital in Latin1 and 0xFC is the small letter, while 0xFC is the
346 capital in EBCDIC and 0xDC is the small one. This factoid may be
347 exploited in writing case insensitive tests that are the same across all
352 Assuming the character set is just ASCII
354 ASCII is a 7 bit encoding, but bytes have 8 bits in them. The 128 extra
355 characters have different meanings depending on the locale. Absent a
356 locale, currently these extra characters are generally considered to be
357 unassigned, and this has presented some problems. This has being
358 changed starting in 5.12 so that these characters can be considered to
359 be Latin-1 (ISO-8859-1).
363 Mixing #define and #ifdef
365 #define BURGLE(x) ... \
366 #ifdef BURGLE_OLD_STYLE /* BAD */
367 ... do it the old way ... \
369 ... do it the new way ... \
372 You cannot portably "stack" cpp directives. For example in the above
373 you need two separate BURGLE() #defines, one for each #ifdef branch.
377 Adding non-comment stuff after #endif or #else
381 #else !SNOSH /* BAD */
383 #endif SNOSH /* BAD */
385 The #endif and #else cannot portably have anything non-comment after
386 them. If you want to document what is going (which is a good idea
387 especially if the branches are long), use (C) comments:
395 The gcc option C<-Wendif-labels> warns about the bad variant (by
396 default on starting from Perl 5.9.4).
400 Having a comma after the last element of an enum list
408 is not portable. Leave out the last comma.
410 Also note that whether enums are implicitly morphable to ints varies
411 between compilers, you might need to (int).
417 // This function bamfoodles the zorklator. /* BAD */
419 That is C99 or C++. Perl is C89. Using the //-comments is silently
420 allowed by many C compilers but cranking up the ANSI C89 strictness
421 (which we like to do) causes the compilation to fail.
425 Mixing declarations and code
430 set_zorkmids(n); /* BAD */
433 That is C99 or C++. Some C compilers allow that, but you shouldn't.
435 The gcc option C<-Wdeclaration-after-statements> scans for such
436 problems (by default on starting from Perl 5.9.4).
440 Introducing variables inside for()
442 for(int i = ...; ...; ...) { /* BAD */
444 That is C99 or C++. While it would indeed be awfully nice to have that
445 also in C89, to limit the scope of the loop variable, alas, we cannot.
449 Mixing signed char pointers with unsigned char pointers
451 int foo(char *s) { ... }
453 unsigned char *t = ...; /* Or U8* t = ... */
456 While this is legal practice, it is certainly dubious, and downright
457 fatal in at least one platform: for example VMS cc considers this a
458 fatal error. One cause for people often making this mistake is that a
459 "naked char" and therefore dereferencing a "naked char pointer" have an
460 undefined signedness: it depends on the compiler and the flags of the
461 compiler and the underlying platform whether the result is signed or
462 unsigned. For this very same reason using a 'char' as an array index is
467 Macros that have string constants and their arguments as substrings of
470 #define FOO(n) printf("number = %d\n", n) /* BAD */
473 Pre-ANSI semantics for that was equivalent to
475 printf("10umber = %d\10");
477 which is probably not what you were expecting. Unfortunately at least
478 one reasonably common and modern C compiler does "real backward
479 compatibility" here, in AIX that is what still happens even though the
480 rest of the AIX compiler is very happily C89.
484 Using printf formats for non-basic C types
487 printf("i = %d\n", i); /* BAD */
489 While this might by accident work in some platform (where IV happens to
490 be an C<int>), in general it cannot. IV might be something larger. Even
491 worse the situation is with more specific types (defined by Perl's
492 configuration step in F<config.h>):
495 printf("who = %d\n", who); /* BAD */
497 The problem here is that Uid_t might be not only not C<int>-wide but it
498 might also be unsigned, in which case large uids would be printed as
501 There is no simple solution to this because of printf()'s limited
502 intelligence, but for many types the right format is available as with
503 either 'f' or '_f' suffix, for example:
505 IVdf /* IV in decimal */
506 UVxf /* UV is hexadecimal */
508 printf("i = %"IVdf"\n", i); /* The IVdf is a string constant. */
510 Uid_t_f /* Uid_t in decimal */
512 printf("who = %"Uid_t_f"\n", who);
514 Or you can try casting to a "wide enough" type:
516 printf("i = %"IVdf"\n", (IV)something_very_small_and_signed);
518 See L<perlguts/Formatted Printing of Size_t and SSize_t> for how to
521 Also remember that the C<%p> format really does require a void pointer:
524 printf("p = %p\n", (void*)p);
526 The gcc option C<-Wformat> scans for such problems.
530 Blindly using variadic macros
532 gcc has had them for a while with its own syntax, and C99 brought them
533 with a standardized syntax. Don't use the former, and use the latter
534 only if the HAS_C99_VARIADIC_MACROS is defined.
538 Blindly passing va_list
540 Not all platforms support passing va_list to further varargs (stdarg)
541 functions. The right thing to do is to copy the va_list using the
542 Perl_va_copy() if the NEED_VA_COPY is defined.
546 Using gcc statement expressions
548 val = ({...;...;...}); /* BAD */
550 While a nice extension, it's not portable. The Perl code does
551 admittedly use them if available to gain some extra speed (essentially
552 as a funky form of inlining), but you shouldn't.
556 Binding together several statements in a macro
558 Use the macros STMT_START and STMT_END.
566 Testing for operating systems or versions when should be testing for
569 #ifdef __FOONIX__ /* BAD */
573 Unless you know with 100% certainty that quux() is only ever available
574 for the "Foonix" operating system B<and> that is available B<and>
575 correctly working for B<all> past, present, B<and> future versions of
576 "Foonix", the above is very wrong. This is more correct (though still
577 not perfect, because the below is a compile-time check):
583 How does the HAS_QUUX become defined where it needs to be? Well, if
584 Foonix happens to be Unixy enough to be able to run the Configure
585 script, and Configure has been taught about detecting and testing
586 quux(), the HAS_QUUX will be correctly defined. In other platforms, the
587 corresponding configuration step will hopefully do the same.
589 In a pinch, if you cannot wait for Configure to be educated, or if you
590 have a good hunch of where quux() might be available, you can
591 temporarily try the following:
593 #if (defined(__FOONIX__) || defined(__BARNIX__))
603 But in any case, try to keep the features and operating systems
606 A good resource on the predefined macros for various operating
607 systems, compilers, and so forth is
608 L<http://sourceforge.net/p/predef/wiki/Home/>
612 Assuming the contents of static memory pointed to by the return values
613 of Perl wrappers for C library functions doesn't change. Many C library
614 functions return pointers to static storage that can be overwritten by
615 subsequent calls to the same or related functions. Perl has
616 light-weight wrappers for some of these functions, and which don't make
617 copies of the static memory. A good example is the interface to the
618 environment variables that are in effect for the program. Perl has
619 C<PerlEnv_getenv> to get values from the environment. But the return is
620 a pointer to static memory in the C library. If you are using the value
621 to immediately test for something, that's fine, but if you save the
622 value and expect it to be unchanged by later processing, you would be
623 wrong, but perhaps you wouldn't know it because different C library
624 implementations behave differently, and the one on the platform you're
625 testing on might work for your situation. But on some platforms, a
626 subsequent call to C<PerlEnv_getenv> or related function WILL overwrite
627 the memory that your first call points to. This has led to some
628 hard-to-debug problems. Do a L<perlapi/savepv> to make a copy, thus
629 avoiding these problems. You will have to free the copy when you're
630 done to avoid memory leaks. If you don't have control over when it gets
631 freed, you'll need to make the copy in a mortal scalar, like so:
633 if ((s = PerlEnv_getenv("foo") == NULL) {
634 ... /* handle NULL case */
637 s = SvPVX(sv_2mortal(newSVpv(s, 0)));
640 The above example works only if C<"s"> is C<NUL>-terminated; otherwise
641 you have to pass its length to C<newSVpv>.
645 =head2 Problematic System Interfaces
651 malloc(0), realloc(0), calloc(0, 0) are non-portable. To be portable
652 allocate at least one byte. (In general you should rarely need to work
653 at this low level, but instead use the various malloc wrappers.)
657 snprintf() - the return type is unportable. Use my_snprintf() instead.
661 =head2 Security problems
663 Last but not least, here are various tips for safer coding.
664 See also L<perlclib> for libc/stdio replacements one should use.
672 Or we will publicly ridicule you. Seriously.
678 Use mkstemp() instead.
682 Do not use strcpy() or strcat() or strncpy() or strncat()
684 Use my_strlcpy() and my_strlcat() instead: they either use the native
685 implementation, or Perl's own implementation (borrowed from the public
686 domain implementation of INN).
690 Do not use sprintf() or vsprintf()
692 If you really want just plain byte strings, use my_snprintf() and
693 my_vsnprintf() instead, which will try to use snprintf() and
694 vsnprintf() if those safer APIs are available. If you want something
695 fancier than a plain byte string, use
696 L<C<Perl_form>()|perlapi/form> or SVs and
697 L<C<Perl_sv_catpvf()>|perlapi/sv_catpvf>.
699 Note that glibc C<printf()>, C<sprintf()>, etc. are buggy before glibc
700 version 2.17. They won't allow a C<%.s> format with a precision to
701 create a string that isn't valid UTF-8 if the current underlying locale
702 of the program is UTF-8. What happens is that the C<%s> and its operand are
703 simply skipped without any notice.
704 L<https://sourceware.org/bugzilla/show_bug.cgi?id=6530>.
710 Use grok_atoUV() instead. atoi() has ill-defined behavior on overflows,
711 and cannot be used for incremental parsing. It is also affected by locale,
716 Do not use strtol() or strtoul()
718 Use grok_atoUV() instead. strtol() or strtoul() (or their IV/UV-friendly
719 macro disguises, Strtol() and Strtoul(), or Atol() and Atoul() are
720 affected by locale, which is bad.
726 You can compile a special debugging version of Perl, which allows you
727 to use the C<-D> option of Perl to tell more about what Perl is doing.
728 But sometimes there is no alternative than to dive in with a debugger,
729 either to see the stack trace of a core dump (very useful in a bug
730 report), or trying to figure out what went wrong before the core dump
731 happened, or how did we end up having wrong or unexpected results.
733 =head2 Poking at Perl
735 To really poke around with Perl, you'll probably want to build Perl for
736 debugging, like this:
738 ./Configure -d -DDEBUGGING
741 C<-DDEBUGGING> turns on the C compiler's C<-g> flag to have it produce
742 debugging information which will allow us to step through a running
743 program, and to see in which C function we are at (without the debugging
744 information we might see only the numerical addresses of the functions,
745 which is not very helpful). It will also turn on the C<DEBUGGING>
746 compilation symbol which enables all the internal debugging code in Perl.
747 There are a whole bunch of things you can debug with this: L<perlrun>
748 lists them all, and the best way to find out about them is to play about
749 with them. The most useful options are probably
751 l Context (loop) stack processing
752 s Stack snapshots (with v, displays all stacks)
754 o Method and overloading resolution
755 c String/numeric conversions
759 $ perl -Dst -e '$a + 1'
769 Some of the functionality of the debugging code can be achieved with a
770 non-debugging perl by using XS modules:
772 -Dr => use re 'debug'
775 =head2 Using a source-level debugger
777 If the debugging output of C<-D> doesn't help you, it's time to step
778 through perl's execution with a source-level debugger.
784 We'll use C<gdb> for our examples here; the principles will apply to
785 any debugger (many vendors call their debugger C<dbx>), but check the
786 manual of the one you're using.
790 To fire up the debugger, type
794 Or if you have a core dump:
798 You'll want to do that in your Perl source tree so the debugger can
799 read the source code. You should see the copyright message, followed by
804 C<help> will get you into the documentation, but here are the most
811 Run the program with the given arguments.
813 =item * break function_name
815 =item * break source.c:xxx
817 Tells the debugger that we'll want to pause execution when we reach
818 either the named function (but see L<perlguts/Internal Functions>!) or
819 the given line in the named source file.
823 Steps through the program a line at a time.
827 Steps through the program a line at a time, without descending into
832 Run until the next breakpoint.
836 Run until the end of the current function, then stop again.
840 Just pressing Enter will do the most recent operation again - it's a
841 blessing when stepping through miles of source code.
845 Prints the C definition of the argument given.
851 OP *(*op_ppaddr)(void);
853 unsigned int op_type : 9;
854 unsigned int op_opt : 1;
855 unsigned int op_slabbed : 1;
856 unsigned int op_savefree : 1;
857 unsigned int op_static : 1;
858 unsigned int op_folded : 1;
859 unsigned int op_spare : 2;
866 Execute the given C code and print its results. B<WARNING>: Perl makes
867 heavy use of macros, and F<gdb> does not necessarily support macros
868 (see later L</"gdb macro support">). You'll have to substitute them
869 yourself, or to invoke cpp on the source code files (see L</"The .i
870 Targets">) So, for instance, you can't say
876 print Perl_sv_2pv_nolen(sv)
880 You may find it helpful to have a "macro dictionary", which you can
881 produce by saying C<cpp -dM perl.c | sort>. Even then, F<cpp> won't
882 recursively apply those macros for you.
884 =head2 gdb macro support
886 Recent versions of F<gdb> have fairly good macro support, but in order
887 to use it you'll need to compile perl with macro definitions included
888 in the debugging information. Using F<gcc> version 3.1, this means
889 configuring with C<-Doptimize=-g3>. Other compilers might use a
890 different switch (if they support debugging macros at all).
892 =head2 Dumping Perl Data Structures
894 One way to get around this macro hell is to use the dumping functions
895 in F<dump.c>; these work a little like an internal
896 L<Devel::Peek|Devel::Peek>, but they also cover OPs and other
897 structures that you can't get at from Perl. Let's take an example.
898 We'll use the C<$a = $b + $c> we used before, but give it a bit of
899 context: C<$b = "6XXXX"; $c = 2.3;>. Where's a good place to stop and
902 What about C<pp_add>, the function we examined earlier to implement the
905 (gdb) break Perl_pp_add
906 Breakpoint 1 at 0x46249f: file pp_hot.c, line 309.
908 Notice we use C<Perl_pp_add> and not C<pp_add> - see
909 L<perlguts/Internal Functions>. With the breakpoint in place, we can
912 (gdb) run -e '$b = "6XXXX"; $c = 2.3; $a = $b + $c'
914 Lots of junk will go past as gdb reads in the relevant source files and
917 Breakpoint 1, Perl_pp_add () at pp_hot.c:309
918 1396 dSP; dATARGET; bool useleft; SV *svl, *svr;
923 We looked at this bit of code before, and we said that
924 C<dPOPTOPnnrl_ul> arranges for two C<NV>s to be placed into C<left> and
925 C<right> - let's slightly expand it:
927 #define dPOPTOPnnrl_ul NV right = POPn; \
929 NV left = USE_LEFT(leftsv) ? SvNV(leftsv) : 0.0
931 C<POPn> takes the SV from the top of the stack and obtains its NV
932 either directly (if C<SvNOK> is set) or by calling the C<sv_2nv>
933 function. C<TOPs> takes the next SV from the top of the stack - yes,
934 C<POPn> uses C<TOPs> - but doesn't remove it. We then use C<SvNV> to
935 get the NV from C<leftsv> in the same way as before - yes, C<POPn> uses
938 Since we don't have an NV for C<$b>, we'll have to use C<sv_2nv> to
939 convert it. If we step again, we'll find ourselves there:
942 Perl_sv_2nv (sv=0xa0675d0) at sv.c:1669
946 We can now use C<Perl_sv_dump> to investigate the SV:
948 (gdb) print Perl_sv_dump(sv)
949 SV = PV(0xa057cc0) at 0xa0675d0
952 PV = 0xa06a510 "6XXXX"\0
957 We know we're going to get C<6> from this, so let's finish the
961 Run till exit from #0 Perl_sv_2nv (sv=0xa0675d0) at sv.c:1671
962 0x462669 in Perl_pp_add () at pp_hot.c:311
965 We can also dump out this op: the current op is always stored in
966 C<PL_op>, and we can dump it with C<Perl_op_dump>. This'll give us
967 similar output to CPAN module B::Debug.
969 (gdb) print Perl_op_dump(PL_op)
971 13 TYPE = add ===> 14
973 FLAGS = (SCALAR,KIDS)
975 TYPE = null ===> (12)
977 FLAGS = (SCALAR,KIDS)
979 11 TYPE = gvsv ===> 12
985 # finish this later #
987 =head2 Using gdb to look at specific parts of a program
989 With the example above, you knew to look for C<Perl_pp_add>, but what if
990 there were multiple calls to it all over the place, or you didn't know what
991 the op was you were looking for?
993 One way to do this is to inject a rare call somewhere near what you're looking
994 for. For example, you could add C<study> before your method:
1000 (gdb) break Perl_pp_study
1002 And then step until you hit what you're
1003 looking for. This works well in a loop
1004 if you want to only break at certain iterations:
1006 for my $c (1..100) {
1010 =head2 Using gdb to look at what the parser/lexer are doing
1012 If you want to see what perl is doing when parsing/lexing your code, you can
1021 (gdb) break Perl_pp_study
1023 If you want to see what the parser/lexer is doing inside of C<if> blocks and
1024 the like you need to be a little trickier:
1026 if ($a && $b && do { BEGIN { study } 1 } && $c) { ... }
1028 =head1 SOURCE CODE STATIC ANALYSIS
1030 Various tools exist for analysing C source code B<statically>, as
1031 opposed to B<dynamically>, that is, without executing the code. It is
1032 possible to detect resource leaks, undefined behaviour, type
1033 mismatches, portability problems, code paths that would cause illegal
1034 memory accesses, and other similar problems by just parsing the C code
1035 and looking at the resulting graph, what does it tell about the
1036 execution and data flows. As a matter of fact, this is exactly how C
1037 compilers know to give warnings about dubious code.
1041 The good old C code quality inspector, C<lint>, is available in several
1042 platforms, but please be aware that there are several different
1043 implementations of it by different vendors, which means that the flags
1044 are not identical across different platforms.
1046 There is a C<lint> target in Makefile, but you may have to
1047 diddle with the flags (see above).
1051 Coverity (L<http://www.coverity.com/>) is a product similar to lint and as
1052 a testbed for their product they periodically check several open source
1053 projects, and they give out accounts to open source developers to the
1056 There is Coverity setup for the perl5 project:
1057 L<https://scan.coverity.com/projects/perl5>
1059 =head2 HP-UX cadvise (Code Advisor)
1061 HP has a C/C++ static analyzer product for HP-UX caller Code Advisor.
1062 (Link not given here because the URL is horribly long and seems horribly
1063 unstable; use the search engine of your choice to find it.) The use of
1064 the C<cadvise_cc> recipe with C<Configure ... -Dcc=./cadvise_cc>
1065 (see cadvise "User Guide") is recommended; as is the use of C<+wall>.
1067 =head2 cpd (cut-and-paste detector)
1069 The cpd tool detects cut-and-paste coding. If one instance of the
1070 cut-and-pasted code changes, all the other spots should probably be
1071 changed, too. Therefore such code should probably be turned into a
1072 subroutine or a macro.
1074 cpd (L<http://pmd.sourceforge.net/cpd.html>) is part of the pmd project
1075 (L<http://pmd.sourceforge.net/>). pmd was originally written for static
1076 analysis of Java code, but later the cpd part of it was extended to
1077 parse also C and C++.
1079 Download the pmd-bin-X.Y.zip () from the SourceForge site, extract the
1080 pmd-X.Y.jar from it, and then run that on source code thusly:
1082 java -cp pmd-X.Y.jar net.sourceforge.pmd.cpd.CPD \
1083 --minimum-tokens 100 --files /some/where/src --language c > cpd.txt
1085 You may run into memory limits, in which case you should use the -Xmx
1092 Though much can be written about the inconsistency and coverage
1093 problems of gcc warnings (like C<-Wall> not meaning "all the warnings",
1094 or some common portability problems not being covered by C<-Wall>, or
1095 C<-ansi> and C<-pedantic> both being a poorly defined collection of
1096 warnings, and so forth), gcc is still a useful tool in keeping our
1099 The C<-Wall> is by default on.
1101 The C<-ansi> (and its sidekick, C<-pedantic>) would be nice to be on
1102 always, but unfortunately they are not safe on all platforms, they can
1103 for example cause fatal conflicts with the system headers (Solaris
1104 being a prime example). If Configure C<-Dgccansipedantic> is used, the
1105 C<cflags> frontend selects C<-ansi -pedantic> for the platforms where
1106 they are known to be safe.
1108 The following extra flags are added:
1130 C<-Werror=declaration-after-statement>
1134 C<-Werror=pointer-arith>
1138 The following flags would be nice to have but they would first need
1139 their own Augean stablemaster:
1149 C<-Wstrict-prototypes>
1153 The C<-Wtraditional> is another example of the annoying tendency of gcc
1154 to bundle a lot of warnings under one switch (it would be impossible to
1155 deploy in practice because it would complain a lot) but it does contain
1156 some warnings that would be beneficial to have available on their own,
1157 such as the warning about string constants inside macros containing the
1158 macro arguments: this behaved differently pre-ANSI than it does in
1159 ANSI, and some C compilers are still in transition, AIX being an
1162 =head2 Warnings of other C compilers
1164 Other C compilers (yes, there B<are> other C compilers than gcc) often
1165 have their "strict ANSI" or "strict ANSI with some portability
1166 extensions" modes on, like for example the Sun Workshop has its C<-Xa>
1167 mode on (though implicitly), or the DEC (these days, HP...) has its
1170 =head1 MEMORY DEBUGGERS
1172 B<NOTE 1>: Running under older memory debuggers such as Purify,
1173 valgrind or Third Degree greatly slows down the execution: seconds
1174 become minutes, minutes become hours. For example as of Perl 5.8.1, the
1175 ext/Encode/t/Unicode.t takes extraordinarily long to complete under
1176 e.g. Purify, Third Degree, and valgrind. Under valgrind it takes more
1177 than six hours, even on a snappy computer. The said test must be doing
1178 something that is quite unfriendly for memory debuggers. If you don't
1179 feel like waiting, that you can simply kill away the perl process.
1180 Roughly valgrind slows down execution by factor 10, AddressSanitizer by
1183 B<NOTE 2>: To minimize the number of memory leak false alarms (see
1184 L</PERL_DESTRUCT_LEVEL> for more information), you have to set the
1185 environment variable PERL_DESTRUCT_LEVEL to 2. For example, like this:
1187 env PERL_DESTRUCT_LEVEL=2 valgrind ./perl -Ilib ...
1189 B<NOTE 3>: There are known memory leaks when there are compile-time
1190 errors within eval or require, seeing C<S_doeval> in the call stack is
1191 a good sign of these. Fixing these leaks is non-trivial, unfortunately,
1192 but they must be fixed eventually.
1194 B<NOTE 4>: L<DynaLoader> will not clean up after itself completely
1195 unless Perl is built with the Configure option
1196 C<-Accflags=-DDL_UNLOAD_ALL_AT_EXIT>.
1200 The valgrind tool can be used to find out both memory leaks and illegal
1201 heap memory accesses. As of version 3.3.0, Valgrind only supports Linux
1202 on x86, x86-64 and PowerPC and Darwin (OS X) on x86 and x86-64. The
1203 special "test.valgrind" target can be used to run the tests under
1204 valgrind. Found errors and memory leaks are logged in files named
1205 F<testfile.valgrind> and by default output is displayed inline.
1211 Since valgrind adds significant overhead, tests will take much longer to
1212 run. The valgrind tests support being run in parallel to help with this:
1214 TEST_JOBS=9 make test.valgrind
1216 Note that the above two invocations will be very verbose as reachable
1217 memory and leak-checking is enabled by default. If you want to just see
1220 VG_OPTS='-q --leak-check=no --show-reachable=no' TEST_JOBS=9 \
1223 Valgrind also provides a cachegrind tool, invoked on perl as:
1225 VG_OPTS=--tool=cachegrind make test.valgrind
1227 As system libraries (most notably glibc) are also triggering errors,
1228 valgrind allows to suppress such errors using suppression files. The
1229 default suppression file that comes with valgrind already catches a lot
1230 of them. Some additional suppressions are defined in F<t/perl.supp>.
1232 To get valgrind and for more information see
1234 http://valgrind.org/
1236 =head2 AddressSanitizer
1238 AddressSanitizer is a clang and gcc extension, included in clang since
1239 v3.1 and gcc since v4.8. It checks illegal heap pointers, global
1240 pointers, stack pointers and use after free errors, and is fast enough
1241 that you can easily compile your debugging or optimized perl with it.
1242 It does not check memory leaks though. AddressSanitizer is available
1243 for Linux, Mac OS X and soon on Windows.
1245 To build perl with AddressSanitizer, your Configure invocation should
1248 sh Configure -des -Dcc=clang \
1249 -Accflags=-faddress-sanitizer -Aldflags=-faddress-sanitizer \
1250 -Alddlflags=-shared\ -faddress-sanitizer
1252 where these arguments mean:
1258 This should be replaced by the full path to your clang executable if it
1259 is not in your path.
1261 =item * -Accflags=-faddress-sanitizer
1263 Compile perl and extensions sources with AddressSanitizer.
1265 =item * -Aldflags=-faddress-sanitizer
1267 Link the perl executable with AddressSanitizer.
1269 =item * -Alddlflags=-shared\ -faddress-sanitizer
1271 Link dynamic extensions with AddressSanitizer. You must manually
1272 specify C<-shared> because using C<-Alddlflags=-shared> will prevent
1273 Configure from setting a default value for C<lddlflags>, which usually
1274 contains C<-shared> (at least on Linux).
1279 L<http://code.google.com/p/address-sanitizer/wiki/AddressSanitizer>.
1284 Depending on your platform there are various ways of profiling Perl.
1286 There are two commonly used techniques of profiling executables:
1287 I<statistical time-sampling> and I<basic-block counting>.
1289 The first method takes periodically samples of the CPU program counter,
1290 and since the program counter can be correlated with the code generated
1291 for functions, we get a statistical view of in which functions the
1292 program is spending its time. The caveats are that very small/fast
1293 functions have lower probability of showing up in the profile, and that
1294 periodically interrupting the program (this is usually done rather
1295 frequently, in the scale of milliseconds) imposes an additional
1296 overhead that may skew the results. The first problem can be alleviated
1297 by running the code for longer (in general this is a good idea for
1298 profiling), the second problem is usually kept in guard by the
1299 profiling tools themselves.
1301 The second method divides up the generated code into I<basic blocks>.
1302 Basic blocks are sections of code that are entered only in the
1303 beginning and exited only at the end. For example, a conditional jump
1304 starts a basic block. Basic block profiling usually works by
1305 I<instrumenting> the code by adding I<enter basic block #nnnn>
1306 book-keeping code to the generated code. During the execution of the
1307 code the basic block counters are then updated appropriately. The
1308 caveat is that the added extra code can skew the results: again, the
1309 profiling tools usually try to factor their own effects out of the
1312 =head2 Gprof Profiling
1314 I<gprof> is a profiling tool available in many Unix platforms which
1315 uses I<statistical time-sampling>. You can build a profiled version of
1316 F<perl> by compiling using gcc with the flag C<-pg>. Either edit
1317 F<config.sh> or re-run F<Configure>. Running the profiled version of
1318 Perl will create an output file called F<gmon.out> which contains the
1319 profiling data collected during the execution.
1323 $ sh Configure -des -Dusedevel -Accflags='-pg' \
1324 -Aldflags='-pg' -Alddlflags='-pg -shared' \
1326 $ ./perl ... # creates gmon.out in current directory
1327 $ gprof ./perl > out
1330 (you probably need to add C<-shared> to the <-Alddlflags> line until RT
1331 #118199 is resolved)
1333 The F<gprof> tool can then display the collected data in various ways.
1334 Usually F<gprof> understands the following options:
1340 Suppress statically defined functions from the profile.
1344 Suppress the verbose descriptions in the profile.
1348 Exclude the given routine and its descendants from the profile.
1352 Display only the given routine and its descendants in the profile.
1356 Generate a summary file called F<gmon.sum> which then may be given to
1357 subsequent gprof runs to accumulate data over several runs.
1361 Display routines that have zero usage.
1365 For more detailed explanation of the available commands and output
1366 formats, see your own local documentation of F<gprof>.
1368 =head2 GCC gcov Profiling
1370 I<basic block profiling> is officially available in gcc 3.0 and later.
1371 You can build a profiled version of F<perl> by compiling using gcc with
1372 the flags C<-fprofile-arcs -ftest-coverage>. Either edit F<config.sh>
1373 or re-run F<Configure>.
1377 $ sh Configure -des -Dusedevel -Doptimize='-g' \
1378 -Accflags='-fprofile-arcs -ftest-coverage' \
1379 -Aldflags='-fprofile-arcs -ftest-coverage' \
1380 -Alddlflags='-fprofile-arcs -ftest-coverage -shared' \
1382 $ rm -f regexec.c.gcov regexec.gcda
1385 $ less regexec.c.gcov
1387 (you probably need to add C<-shared> to the <-Alddlflags> line until RT
1388 #118199 is resolved)
1390 Running the profiled version of Perl will cause profile output to be
1391 generated. For each source file an accompanying F<.gcda> file will be
1394 To display the results you use the I<gcov> utility (which should be
1395 installed if you have gcc 3.0 or newer installed). F<gcov> is run on
1396 source code files, like this
1400 which will cause F<sv.c.gcov> to be created. The F<.gcov> files contain
1401 the source code annotated with relative frequencies of execution
1402 indicated by "#" markers. If you want to generate F<.gcov> files for
1403 all profiled object files, you can run something like this:
1405 for file in `find . -name \*.gcno`
1406 do sh -c "cd `dirname $file` && gcov `basename $file .gcno`"
1409 Useful options of F<gcov> include C<-b> which will summarise the basic
1410 block, branch, and function call coverage, and C<-c> which instead of
1411 relative frequencies will use the actual counts. For more information
1412 on the use of F<gcov> and basic block profiling with gcc, see the
1413 latest GNU CC manual. As of gcc 4.8, this is at
1414 L<http://gcc.gnu.org/onlinedocs/gcc/Gcov-Intro.html#Gcov-Intro>
1416 =head1 MISCELLANEOUS TRICKS
1418 =head2 PERL_DESTRUCT_LEVEL
1420 If you want to run any of the tests yourself manually using e.g.
1421 valgrind, please note that by default perl B<does not> explicitly
1422 cleanup all the memory it has allocated (such as global memory arenas)
1423 but instead lets the exit() of the whole program "take care" of such
1424 allocations, also known as "global destruction of objects".
1426 There is a way to tell perl to do complete cleanup: set the environment
1427 variable PERL_DESTRUCT_LEVEL to a non-zero value. The t/TEST wrapper
1428 does set this to 2, and this is what you need to do too, if you don't
1429 want to see the "global leaks": For example, for running under valgrind
1431 env PERL_DESTRUCT_LEVEL=2 valgrind ./perl -Ilib t/foo/bar.t
1433 (Note: the mod_perl apache module uses also this environment variable
1434 for its own purposes and extended its semantics. Refer to the mod_perl
1435 documentation for more information. Also, spawned threads do the
1436 equivalent of setting this variable to the value 1.)
1438 If, at the end of a run you get the message I<N scalars leaked>, you
1439 can recompile with C<-DDEBUG_LEAKING_SCALARS>,
1440 (C<Configure -Accflags=-DDEBUG_LEAKING_SCALARS>), which will cause the
1441 addresses of all those leaked SVs to be dumped along with details as to
1442 where each SV was originally allocated. This information is also
1443 displayed by Devel::Peek. Note that the extra details recorded with
1444 each SV increases memory usage, so it shouldn't be used in production
1445 environments. It also converts C<new_SV()> from a macro into a real
1446 function, so you can use your favourite debugger to discover where
1447 those pesky SVs were allocated.
1449 If you see that you're leaking memory at runtime, but neither valgrind
1450 nor C<-DDEBUG_LEAKING_SCALARS> will find anything, you're probably
1451 leaking SVs that are still reachable and will be properly cleaned up
1452 during destruction of the interpreter. In such cases, using the C<-Dm>
1453 switch can point you to the source of the leak. If the executable was
1454 built with C<-DDEBUG_LEAKING_SCALARS>, C<-Dm> will output SV
1455 allocations in addition to memory allocations. Each SV allocation has a
1456 distinct serial number that will be written on creation and destruction
1457 of the SV. So if you're executing the leaking code in a loop, you need
1458 to look for SVs that are created, but never destroyed between each
1459 cycle. If such an SV is found, set a conditional breakpoint within
1460 C<new_SV()> and make it break only when C<PL_sv_serial> is equal to the
1461 serial number of the leaking SV. Then you will catch the interpreter in
1462 exactly the state where the leaking SV is allocated, which is
1463 sufficient in many cases to find the source of the leak.
1465 As C<-Dm> is using the PerlIO layer for output, it will by itself
1466 allocate quite a bunch of SVs, which are hidden to avoid recursion. You
1467 can bypass the PerlIO layer if you use the SV logging provided by
1468 C<-DPERL_MEM_LOG> instead.
1472 If compiled with C<-DPERL_MEM_LOG> (C<-Accflags=-DPERL_MEM_LOG>), both
1473 memory and SV allocations go through logging functions, which is
1474 handy for breakpoint setting.
1476 Unless C<-DPERL_MEM_LOG_NOIMPL> (C<-Accflags=-DPERL_MEM_LOG_NOIMPL>) is
1477 also compiled, the logging functions read $ENV{PERL_MEM_LOG} to
1478 determine whether to log the event, and if so how:
1480 $ENV{PERL_MEM_LOG} =~ /m/ Log all memory ops
1481 $ENV{PERL_MEM_LOG} =~ /s/ Log all SV ops
1482 $ENV{PERL_MEM_LOG} =~ /t/ include timestamp in Log
1483 $ENV{PERL_MEM_LOG} =~ /^(\d+)/ write to FD given (default is 2)
1485 Memory logging is somewhat similar to C<-Dm> but is independent of
1486 C<-DDEBUGGING>, and at a higher level; all uses of Newx(), Renew(), and
1487 Safefree() are logged with the caller's source code file and line
1488 number (and C function name, if supported by the C compiler). In
1489 contrast, C<-Dm> is directly at the point of C<malloc()>. SV logging is
1492 Since the logging doesn't use PerlIO, all SV allocations are logged and
1493 no extra SV allocations are introduced by enabling the logging. If
1494 compiled with C<-DDEBUG_LEAKING_SCALARS>, the serial number for each SV
1495 allocation is also logged.
1499 Those debugging perl with the DDD frontend over gdb may find the
1502 You can extend the data conversion shortcuts menu, so for example you
1503 can display an SV's IV value with one click, without doing any typing.
1504 To do that simply edit ~/.ddd/init file and add after:
1506 ! Display shortcuts.
1507 Ddd*gdbDisplayShortcuts: \
1508 /t () // Convert to Bin\n\
1509 /d () // Convert to Dec\n\
1510 /x () // Convert to Hex\n\
1511 /o () // Convert to Oct(\n\
1513 the following two lines:
1515 ((XPV*) (())->sv_any )->xpv_pv // 2pvx\n\
1516 ((XPVIV*) (())->sv_any )->xiv_iv // 2ivx
1518 so now you can do ivx and pvx lookups or you can plug there the sv_peek
1521 Perl_sv_peek(my_perl, (SV*)()) // sv_peek
1523 (The my_perl is for threaded builds.) Just remember that every line,
1524 but the last one, should end with \n\
1526 Alternatively edit the init file interactively via: 3rd mouse button ->
1527 New Display -> Edit Menu
1529 Note: you can define up to 20 conversion shortcuts in the gdb section.
1533 On some platforms Perl supports retrieving the C level backtrace
1534 (similar to what symbolic debuggers like gdb do).
1536 The backtrace returns the stack trace of the C call frames,
1537 with the symbol names (function names), the object names (like "perl"),
1538 and if it can, also the source code locations (file:line).
1540 The supported platforms are Linux, and OS X (some *BSD might
1541 work at least partly, but they have not yet been tested).
1543 This feature hasn't been tested with multiple threads, but it will
1544 only show the backtrace of the thread doing the backtracing.
1546 The feature needs to be enabled with C<Configure -Dusecbacktrace>.
1548 The C<-Dusecbacktrace> also enables keeping the debug information when
1549 compiling/linking (often: C<-g>). Many compilers/linkers do support
1550 having both optimization and keeping the debug information. The debug
1551 information is needed for the symbol names and the source locations.
1553 Static functions might not be visible for the backtrace.
1555 Source code locations, even if available, can often be missing or
1556 misleading if the compiler has e.g. inlined code. Optimizer can
1557 make matching the source code and the object code quite challenging.
1563 You B<must> have the BFD (-lbfd) library installed, otherwise C<perl> will
1564 fail to link. The BFD is usually distributed as part of the GNU binutils.
1566 Summary: C<Configure ... -Dusecbacktrace>
1567 and you need C<-lbfd>.
1571 The source code locations are supported B<only> if you have
1572 the Developer Tools installed. (BFD is B<not> needed.)
1574 Summary: C<Configure ... -Dusecbacktrace>
1575 and installing the Developer Tools would be good.
1579 Optionally, for trying out the feature, you may want to enable
1580 automatic dumping of the backtrace just before a warning or croak (die)
1581 message is emitted, by adding C<-Accflags=-DUSE_C_BACKTRACE_ON_ERROR>
1584 Unless the above additional feature is enabled, nothing about the
1585 backtrace functionality is visible, except for the Perl/XS level.
1587 Furthermore, even if you have enabled this feature to be compiled,
1588 you need to enable it in runtime with an environment variable:
1589 C<PERL_C_BACKTRACE_ON_ERROR=10>. It must be an integer higher
1590 than zero, telling the desired frame count.
1592 Retrieving the backtrace from Perl level (using for example an XS
1593 extension) would be much less exciting than one would hope: normally
1594 you would see C<runops>, C<entersub>, and not much else. This API is
1595 intended to be called B<from within> the Perl implementation, not from
1596 Perl level execution.
1598 The C API for the backtrace is as follows:
1602 =item get_c_backtrace
1604 =item free_c_backtrace
1606 =item get_c_backtrace_dump
1608 =item dump_c_backtrace
1614 If you see in a debugger a memory area mysteriously full of 0xABABABAB
1615 or 0xEFEFEFEF, you may be seeing the effect of the Poison() macros, see
1618 =head2 Read-only optrees
1620 Under ithreads the optree is read only. If you want to enforce this, to
1621 check for write accesses from buggy code, compile with
1622 C<-Accflags=-DPERL_DEBUG_READONLY_OPS>
1623 to enable code that allocates op memory
1624 via C<mmap>, and sets it read-only when it is attached to a subroutine.
1625 Any write access to an op results in a C<SIGBUS> and abort.
1627 This code is intended for development only, and may not be portable
1628 even to all Unix variants. Also, it is an 80% solution, in that it
1629 isn't able to make all ops read only. Specifically it does not apply to
1630 op slabs belonging to C<BEGIN> blocks.
1632 However, as an 80% solution it is still effective, as it has caught
1635 =head2 When is a bool not a bool?
1637 On pre-C99 compilers, C<bool> is defined as equivalent to C<char>.
1638 Consequently assignment of any larger type to a C<bool> is unsafe and may be
1639 truncated. The C<cBOOL> macro exists to cast it correctly; you may also find
1640 that using it is shorter and clearer than writing out the equivalent
1641 conditional expression longhand.
1643 On those platforms and compilers where C<bool> really is a boolean (C++,
1644 C99), it is easy to forget the cast. You can force C<bool> to be a C<char>
1645 by compiling with C<-Accflags=-DPERL_BOOL_AS_CHAR>. You may also wish to
1646 run C<Configure> with something like
1648 -Accflags='-Wconversion -Wno-sign-conversion -Wno-shorten-64-to-32'
1650 or your compiler's equivalent to make it easier to spot any unsafe truncations
1653 The C<TRUE> and C<FALSE> macros are available for situations where using them
1654 would clarify intent. (But they always just mean the same as the integers 1 and
1655 0 regardless, so using them isn't compulsory.)
1657 =head2 The .i Targets
1659 You can expand the macros in a F<foo.c> file by saying
1663 which will expand the macros using cpp. Don't be scared by the
1668 This document was originally written by Nathan Torkington, and is
1669 maintained by the perl5-porters mailing list.