3 # Copyright (c) 1996, 1997, 1998 Malcolm Beattie
5 # You may distribute under the terms of either the GNU General Public
6 # License or the Artistic License, as specified in the README file.
10 our $VERSION = '1.01';
16 # walkoptree_slow comes from B.pm (you are there),
17 # walkoptree comes from B.xs
18 @EXPORT_OK = qw(minus_c ppname save_BEGINs
19 class peekop cast_I32 cstring cchar hash threadsv_names
20 main_root main_start main_cv svref_2object opnumber
21 amagic_generation perlstring
22 walkoptree_slow walkoptree walkoptree_exec walksymtable
23 parents comppadlist sv_undef compile_stats timing_info
24 begin_av init_av end_av regex_padav);
28 @B::SV::ISA = 'B::OBJECT';
29 @B::NULL::ISA = 'B::SV';
30 @B::PV::ISA = 'B::SV';
31 @B::IV::ISA = 'B::SV';
32 @B::NV::ISA = 'B::IV';
33 @B::RV::ISA = 'B::SV';
34 @B::PVIV::ISA = qw(B::PV B::IV);
35 @B::PVNV::ISA = qw(B::PV B::NV);
36 @B::PVMG::ISA = 'B::PVNV';
37 @B::PVLV::ISA = 'B::PVMG';
38 @B::BM::ISA = 'B::PVMG';
39 @B::AV::ISA = 'B::PVMG';
40 @B::GV::ISA = 'B::PVMG';
41 @B::HV::ISA = 'B::PVMG';
42 @B::CV::ISA = 'B::PVMG';
43 @B::IO::ISA = 'B::PVMG';
44 @B::FM::ISA = 'B::CV';
46 @B::OP::ISA = 'B::OBJECT';
47 @B::UNOP::ISA = 'B::OP';
48 @B::BINOP::ISA = 'B::UNOP';
49 @B::LOGOP::ISA = 'B::UNOP';
50 @B::LISTOP::ISA = 'B::BINOP';
51 @B::SVOP::ISA = 'B::OP';
52 @B::PADOP::ISA = 'B::OP';
53 @B::PVOP::ISA = 'B::OP';
54 @B::CVOP::ISA = 'B::OP';
55 @B::LOOP::ISA = 'B::LISTOP';
56 @B::PMOP::ISA = 'B::LISTOP';
57 @B::COP::ISA = 'B::OP';
59 @B::SPECIAL::ISA = 'B::OBJECT';
62 # Stop "-w" from complaining about the lack of a real B::OBJECT class
67 my $name = (shift())->NAME;
69 # The regex below corresponds to the isCONTROLVAR macro
72 $name =~ s/^([\cA-\cZ\c\\c[\c]\c?\c_\c^])/"^".
73 chr( utf8::unicode_to_native( 64 ^ ord($1) ))/e;
75 # When we say unicode_to_native we really mean ascii_to_native,
76 # which matters iff this is a non-ASCII platform (EBCDIC).
81 sub B::IV::int_value {
83 return (($self->FLAGS() & SVf_IVisUV()) ? $self->UVX : $self->IV);
86 sub B::NULL::as_string() {""}
87 sub B::IV::as_string() {goto &B::IV::int_value}
88 sub B::PV::as_string() {goto &B::PV::PV}
95 my ($class, $value) = @_;
97 walkoptree_debug($value);
107 sub parents { \@parents }
112 return sprintf("%s (0x%x) %s", class($op), $$op, $op->name);
115 sub walkoptree_slow {
116 my($op, $method, $level) = @_;
117 $op_count++; # just for statistics
119 warn(sprintf("walkoptree: %d. %s\n", $level, peekop($op))) if $debug;
120 $op->$method($level);
121 if ($$op && ($op->flags & OPf_KIDS)) {
123 unshift(@parents, $op);
124 for ($kid = $op->first; $$kid; $kid = $kid->sibling) {
125 walkoptree_slow($kid, $method, $level + 1);
129 if (class($op) eq 'PMOP' && $op->pmreplroot && ${$op->pmreplroot}) {
130 unshift(@parents, $op);
131 walkoptree_slow($op->pmreplroot, $method, $level + 1);
137 return "Total number of OPs processed: $op_count\n";
141 my ($sec, $min, $hr) = localtime;
142 my ($user, $sys) = times;
143 sprintf("%02d:%02d:%02d user=$user sys=$sys",
144 $hr, $min, $sec, $user, $sys);
154 my ($obj, $value) = @_;
155 # warn(sprintf("savesym: sym_%x => %s\n", $$obj, $value)); # debug
156 $symtable{sprintf("sym_%x", $$obj)} = $value;
161 return $symtable{sprintf("sym_%x", $$obj)};
164 sub walkoptree_exec {
165 my ($op, $method, $level) = @_;
168 my $prefix = " " x $level;
169 for (; $$op; $op = $op->next) {
172 print $prefix, "goto $sym\n";
175 savesym($op, sprintf("%s (0x%lx)", class($op), $$op));
176 $op->$method($level);
179 /^(or|and|mapwhile|grepwhile|entertry|range|cond_expr)$/)
181 print $prefix, uc($1), " => {\n";
182 walkoptree_exec($op->other, $method, $level + 1);
183 print $prefix, "}\n";
184 } elsif ($ppname eq "match" || $ppname eq "subst") {
185 my $pmreplstart = $op->pmreplstart;
187 print $prefix, "PMREPLSTART => {\n";
188 walkoptree_exec($pmreplstart, $method, $level + 1);
189 print $prefix, "}\n";
191 } elsif ($ppname eq "substcont") {
192 print $prefix, "SUBSTCONT => {\n";
193 walkoptree_exec($op->other->pmreplstart, $method, $level + 1);
194 print $prefix, "}\n";
196 } elsif ($ppname eq "enterloop") {
197 print $prefix, "REDO => {\n";
198 walkoptree_exec($op->redoop, $method, $level + 1);
199 print $prefix, "}\n", $prefix, "NEXT => {\n";
200 walkoptree_exec($op->nextop, $method, $level + 1);
201 print $prefix, "}\n", $prefix, "LAST => {\n";
202 walkoptree_exec($op->lastop, $method, $level + 1);
203 print $prefix, "}\n";
204 } elsif ($ppname eq "subst") {
205 my $replstart = $op->pmreplstart;
207 print $prefix, "SUBST => {\n";
208 walkoptree_exec($replstart, $method, $level + 1);
209 print $prefix, "}\n";
216 my ($symref, $method, $recurse, $prefix) = @_;
221 $prefix = '' unless defined $prefix;
222 while (($sym, $ref) = each %$symref) {
223 $fullname = "*main::".$prefix.$sym;
225 $sym = $prefix . $sym;
226 if ($sym ne "main::" && $sym ne "<none>::" && &$recurse($sym)) {
227 walksymtable(\%$fullname, $method, $recurse, $sym);
230 svref_2object(\*$fullname)->$method();
241 my ($class, $section, $symtable, $default) = @_;
242 $output_fh ||= FileHandle->new_tmpfile;
243 my $obj = bless [-1, $section, $symtable, $default], $class;
244 $sections{$section} = $obj;
249 my ($class, $section) = @_;
250 return $sections{$section};
255 while (defined($_ = shift)) {
256 print $output_fh "$section->[1]\t$_\n";
263 return $section->[0];
268 return $section->[1];
273 return $section->[2];
278 return $section->[3];
282 my ($section, $fh, $format) = @_;
283 my $name = $section->name;
284 my $sym = $section->symtable || {};
285 my $default = $section->default;
287 seek($output_fh, 0, 0);
288 while (<$output_fh>) {
293 exists($sym->{$1}) ? $sym->{$1} : $default;
295 printf $fh $format, $_;
309 B - The Perl Compiler
317 The C<B> module supplies classes which allow a Perl program to delve
318 into its own innards. It is the module used to implement the
319 "backends" of the Perl compiler. Usage of the compiler does not
320 require knowledge of this module: see the F<O> module for the
321 user-visible part. The C<B> module is of use to those who want to
322 write new compiler backends. This documentation assumes that the
323 reader knows a fair amount about perl's internals including such
324 things as SVs, OPs and the internal symbol table and syntax tree
329 The C<B> module contains a set of utility functions for querying the
330 current state of the Perl interpreter; typically these functions
331 return objects from the B::SV and B::OP classes, or their derived
332 classes. These classes in turn define methods for querying the
333 resulting objects about their own internal state.
335 =head1 Utility Functions
337 The C<B> module exports a variety of functions: some are simple
338 utility functions, others provide a Perl program with a way to
339 get an initial "handle" on an internal object.
341 =head2 Functions Returning C<B::SV>, C<B::AV>, C<B::HV>, and C<B::CV> objects
343 For descriptions of the class hierachy of these objects and the
344 methods that can be called on them, see below, L<"OVERVIEW OF
345 CLASSES"> and L<"SV-RELATED CLASSES">.
351 Returns the SV object corresponding to the C variable C<sv_undef>.
355 Returns the SV object corresponding to the C variable C<sv_yes>.
359 Returns the SV object corresponding to the C variable C<sv_no>.
361 =item svref_2object(SVREF)
363 Takes a reference to any Perl value, and turns the referred-to value
364 into an object in the appropriate B::OP-derived or B::SV-derived
365 class. Apart from functions such as C<main_root>, this is the primary
366 way to get an initial "handle" on an internal perl data structure
367 which can then be followed with the other access methods.
369 =item amagic_generation
371 Returns the SV object corresponding to the C variable C<amagic_generation>.
375 Returns the AV object (i.e. in class B::AV) representing INIT blocks.
379 Returns the AV object (i.e. in class B::AV) representing BEGIN blocks.
383 Returns the AV object (i.e. in class B::AV) representing END blocks.
387 Returns the AV object (i.e. in class B::AV) of the global comppadlist.
391 Only when perl was compiled with ithreads.
395 Return the (faked) CV corresponding to the main part of the Perl
400 =head2 Functions for Examining the Symbol Table
404 =item walksymtable(SYMREF, METHOD, RECURSE, PREFIX)
406 Walk the symbol table starting at SYMREF and call METHOD on each
407 symbol (a B::GV object) visited. When the walk reaches package
408 symbols (such as "Foo::") it invokes RECURSE, passing in the symbol
409 name, and only recurses into the package if that sub returns true.
411 PREFIX is the name of the SYMREF you're walking.
415 # Walk CGI's symbol table calling print_subs on each symbol.
416 # Recurse only into CGI::Util::
417 walksymtable(\%CGI::, 'print_subs', sub { $_[0] eq 'CGI::Util::' },
420 print_subs() is a B::GV method you have declared. Also see L<"B::GV
425 =head2 Functions Returning C<B::OP> objects or for walking op trees
427 For descriptions of the class hierachy of these objects and the
428 methods that can be called on them, see below, L<"OVERVIEW OF
429 CLASSES"> and L<"OP-RELATED CLASSES">.
435 Returns the root op (i.e. an object in the appropriate B::OP-derived
436 class) of the main part of the Perl program.
440 Returns the starting op of the main part of the Perl program.
442 =item walkoptree(OP, METHOD)
444 Does a tree-walk of the syntax tree based at OP and calls METHOD on
445 each op it visits. Each node is visited before its children. If
446 C<walkoptree_debug> (see below) has been called to turn debugging on then
447 the method C<walkoptree_debug> is called on each op before METHOD is
450 =item walkoptree_debug(DEBUG)
452 Returns the current debugging flag for C<walkoptree>. If the optional
453 DEBUG argument is non-zero, it sets the debugging flag to that. See
454 the description of C<walkoptree> above for what the debugging flag
459 =head2 Miscellaneous Utility Functions
465 Return the PP function name (e.g. "pp_add") of op number OPNUM.
469 Returns a string in the form "0x..." representing the value of the
470 internal hash function used by perl on string STR.
474 Casts I to the internal I32 type used by that perl.
478 Does the equivalent of the C<-c> command-line option. Obviously, this
479 is only useful in a BEGIN block or else the flag is set too late.
483 Returns a double-quote-surrounded escaped version of STR which can
484 be used as a string in C source code.
486 =item perlstring(STR)
488 Returns a double-quote-surrounded escaped version of STR which can
489 be used as a string in Perl source code.
493 Returns the class of an object without the part of the classname
494 preceding the first C<"::">. This is used to turn C<"B::UNOP"> into
495 C<"UNOP"> for example.
499 In a perl compiled for threads, this returns a list of the special
500 per-thread threadsv variables.
507 =head1 OVERVIEW OF CLASSES
509 The C structures used by Perl's internals to hold SV and OP
510 information (PVIV, AV, HV, ..., OP, SVOP, UNOP, ...) are modelled on a
511 class hierarchy and the C<B> module gives access to them via a true
512 object hierarchy. Structure fields which point to other objects
513 (whether types of SV or types of OP) are represented by the C<B>
514 module as Perl objects of the appropriate class.
516 The bulk of the C<B> module is the methods for accessing fields of
519 Note that all access is read-only. You cannot modify the internals by
522 =head2 SV-RELATED CLASSES
524 B::IV, B::NV, B::RV, B::PV, B::PVIV, B::PVNV, B::PVMG, B::BM, B::PVLV,
525 B::AV, B::HV, B::CV, B::GV, B::FM, B::IO. These classes correspond in
526 the obvious way to the underlying C structures of similar names. The
527 inheritance hierarchy mimics the underlying C "inheritance":
531 +--------------+----------------------+
545 +------+-----+----+------+-----+-----+
547 B::PVLV B::BM B::AV B::GV B::HV B::CV B::IO
553 Access methods correspond to the underlying C macros for field access,
554 usually with the leading "class indication" prefix removed (Sv, Av,
555 Hv, ...). The leading prefix is only left in cases where its removal
556 would cause a clash in method name. For example, C<GvREFCNT> stays
557 as-is since its abbreviation would clash with the "superclass" method
558 C<REFCNT> (corresponding to the C function C<SvREFCNT>).
576 Returns the value of the IV, I<interpreted as
577 a signed integer>. This will be misleading
578 if C<FLAGS & SVf_IVisUV>. Perhaps you want the
579 C<int_value> method instead?
587 This method returns the value of the IV as an integer.
588 It differs from C<IV> in that it returns the correct
589 value regardless of whether it's stored signed or
622 This method is the one you usually want. It constructs a
623 string using the length and offset information in the struct:
624 for ordinary scalars it will return the string that you'd see
625 from Perl, even if it contains null characters.
629 Same as B::RV::RV, except that it will die() if the PV isn't
634 This method is less often useful. It assumes that the string
635 stored in the struct is null-terminated, and disregards the
638 It is the appropriate method to use if you need to get the name
639 of a lexical variable from a padname array. Lexical variable names
640 are always stored with a null terminator, and the length field
641 (SvCUR) is overloaded for other purposes and can't be relied on here.
645 =head2 B::PVMG Methods
655 =head2 B::MAGIC Methods
663 Only valid on r-magic, returns the string that generated the regexp.
673 Will die() if called on r-magic.
679 Only valid on r-magic, returns the integer value of the REGEX stored
684 =head2 B::PVLV Methods
718 This method returns TRUE if the GP field of the GV is NULL.
724 This method returns the name of the glob, but if the first
725 character of the name is a control character, then it converts
726 it to ^X first, so that *^G would return "^G" rather than "\cG".
728 It's useful if you want to print out the name of a variable.
729 If you restrict yourself to globs which exist at compile-time
730 then the result ought to be unambiguous, because code like
731 C<${"^G"} = 1> is compiled as two ops - a constant string and
732 a dereference (rv2gv) - so that the glob is created at runtime.
734 If you're working with globs at runtime, and need to disambiguate
735 *^G from *{"^G"}, then you should use the raw NAME method.
799 Takes one arguments ( 'stdin' | 'stdout' | 'stderr' ) and returns true
800 if the IoIFP of the object is equal to the handle whose name was
801 passed as argument ( i.e. $io->IsSTD('stderr') is true if
802 IoIFP($io) == PerlIO_stdin() ).
846 For constant subroutines, returns the constant SV returned by the subroutine.
874 =head2 OP-RELATED CLASSES
876 C<B::OP>, C<B::UNOP>, C<B::BINOP>, C<B::LOGOP>, C<B::LISTOP>, C<B::PMOP>,
877 C<B::SVOP>, C<B::PADOP>, C<B::PVOP>, C<B::CVOP>, C<B::LOOP>, C<B::COP>.
879 These classes correspond in the obvious way to the underlying C
880 structures of similar names. The inheritance hierarchy mimics the
881 underlying C "inheritance":
885 +---------------+--------+--------+------+
887 B::UNOP B::SVOP B::PADOP B::CVOP B::COP
898 Access methods correspond to the underlying C structre field names,
899 with the leading "class indication" prefix (C<"op_">) removed.
911 This returns the op name as a string (e.g. "add", "rv2av").
915 This returns the function name as a string (e.g. "PL_ppaddr[OP_ADD]",
916 "PL_ppaddr[OP_RV2AV]").
920 This returns the op description from the global C PL_op_desc array
921 (e.g. "addition" "array deref").
935 =head2 B::UNOP METHOD
943 =head2 B::BINOP METHOD
951 =head2 B::LOGOP METHOD
959 =head2 B::LISTOP METHOD
967 =head2 B::PMOP Methods
989 Only when perl was compiled with ithreads.
993 =head2 B::SVOP METHOD
1003 =head2 B::PADOP METHOD
1011 =head2 B::PVOP METHOD
1019 =head2 B::LOOP Methods
1031 =head2 B::COP Methods
1052 Malcolm Beattie, C<mbeattie@sable.ox.ac.uk>