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 = '1.09';
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 check_av end_av regex_padav dowarn
25 defstash curstash warnhook diehook inc_gv
30 @B::SV::ISA = 'B::OBJECT';
31 @B::NULL::ISA = 'B::SV';
32 @B::PV::ISA = 'B::SV';
33 @B::IV::ISA = 'B::SV';
34 @B::NV::ISA = 'B::SV';
35 @B::RV::ISA = 'B::SV';
36 @B::PVIV::ISA = qw(B::PV B::IV);
37 @B::PVNV::ISA = qw(B::PVIV B::NV);
38 @B::PVMG::ISA = 'B::PVNV';
39 # Change in the inheritance hierarchy post 5.9.0
40 @B::PVLV::ISA = $] > 5.009 ? 'B::GV' : 'B::PVMG';
41 @B::BM::ISA = 'B::PVMG';
42 @B::AV::ISA = 'B::PVMG';
43 @B::GV::ISA = 'B::PVMG';
44 @B::HV::ISA = 'B::PVMG';
45 @B::CV::ISA = 'B::PVMG';
46 @B::IO::ISA = 'B::PVMG';
47 @B::FM::ISA = 'B::CV';
49 @B::OP::ISA = 'B::OBJECT';
50 @B::UNOP::ISA = 'B::OP';
51 @B::BINOP::ISA = 'B::UNOP';
52 @B::LOGOP::ISA = 'B::UNOP';
53 @B::LISTOP::ISA = 'B::BINOP';
54 @B::SVOP::ISA = 'B::OP';
55 @B::PADOP::ISA = 'B::OP';
56 @B::PVOP::ISA = 'B::OP';
57 @B::LOOP::ISA = 'B::LISTOP';
58 @B::PMOP::ISA = 'B::LISTOP';
59 @B::COP::ISA = 'B::OP';
61 @B::SPECIAL::ISA = 'B::OBJECT';
64 # Stop "-w" from complaining about the lack of a real B::OBJECT class
69 my $name = (shift())->NAME;
71 # The regex below corresponds to the isCONTROLVAR macro
74 $name =~ s/^([\cA-\cZ\c\\c[\c]\c?\c_\c^])/"^".
75 chr( utf8::unicode_to_native( 64 ^ ord($1) ))/e;
77 # When we say unicode_to_native we really mean ascii_to_native,
78 # which matters iff this is a non-ASCII platform (EBCDIC).
83 sub B::IV::int_value {
85 return (($self->FLAGS() & SVf_IVisUV()) ? $self->UVX : $self->IV);
88 sub B::NULL::as_string() {""}
89 sub B::IV::as_string() {goto &B::IV::int_value}
90 sub B::PV::as_string() {goto &B::PV::PV}
97 my ($class, $value) = @_;
99 walkoptree_debug($value);
109 sub parents { \@parents }
114 return sprintf("%s (0x%x) %s", class($op), $$op, $op->name);
117 sub walkoptree_slow {
118 my($op, $method, $level) = @_;
119 $op_count++; # just for statistics
121 warn(sprintf("walkoptree: %d. %s\n", $level, peekop($op))) if $debug;
122 $op->$method($level);
123 if ($$op && ($op->flags & OPf_KIDS)) {
125 unshift(@parents, $op);
126 for ($kid = $op->first; $$kid; $kid = $kid->sibling) {
127 walkoptree_slow($kid, $method, $level + 1);
131 if (class($op) eq 'PMOP' && ref($op->pmreplroot) && ${$op->pmreplroot}) {
132 unshift(@parents, $op);
133 walkoptree_slow($op->pmreplroot, $method, $level + 1);
139 return "Total number of OPs processed: $op_count\n";
143 my ($sec, $min, $hr) = localtime;
144 my ($user, $sys) = times;
145 sprintf("%02d:%02d:%02d user=$user sys=$sys",
146 $hr, $min, $sec, $user, $sys);
156 my ($obj, $value) = @_;
157 # warn(sprintf("savesym: sym_%x => %s\n", $$obj, $value)); # debug
158 $symtable{sprintf("sym_%x", $$obj)} = $value;
163 return $symtable{sprintf("sym_%x", $$obj)};
166 sub walkoptree_exec {
167 my ($op, $method, $level) = @_;
170 my $prefix = " " x $level;
171 for (; $$op; $op = $op->next) {
174 print $prefix, "goto $sym\n";
177 savesym($op, sprintf("%s (0x%lx)", class($op), $$op));
178 $op->$method($level);
181 /^(d?or(assign)?|and(assign)?|mapwhile|grepwhile|entertry|range|cond_expr)$/)
183 print $prefix, uc($1), " => {\n";
184 walkoptree_exec($op->other, $method, $level + 1);
185 print $prefix, "}\n";
186 } elsif ($ppname eq "match" || $ppname eq "subst") {
187 my $pmreplstart = $op->pmreplstart;
189 print $prefix, "PMREPLSTART => {\n";
190 walkoptree_exec($pmreplstart, $method, $level + 1);
191 print $prefix, "}\n";
193 } elsif ($ppname eq "substcont") {
194 print $prefix, "SUBSTCONT => {\n";
195 walkoptree_exec($op->other->pmreplstart, $method, $level + 1);
196 print $prefix, "}\n";
198 } elsif ($ppname eq "enterloop") {
199 print $prefix, "REDO => {\n";
200 walkoptree_exec($op->redoop, $method, $level + 1);
201 print $prefix, "}\n", $prefix, "NEXT => {\n";
202 walkoptree_exec($op->nextop, $method, $level + 1);
203 print $prefix, "}\n", $prefix, "LAST => {\n";
204 walkoptree_exec($op->lastop, $method, $level + 1);
205 print $prefix, "}\n";
206 } elsif ($ppname eq "subst") {
207 my $replstart = $op->pmreplstart;
209 print $prefix, "SUBST => {\n";
210 walkoptree_exec($replstart, $method, $level + 1);
211 print $prefix, "}\n";
218 my ($symref, $method, $recurse, $prefix) = @_;
223 $prefix = '' unless defined $prefix;
224 while (($sym, $ref) = each %$symref) {
225 $fullname = "*main::".$prefix.$sym;
227 $sym = $prefix . $sym;
228 if ($sym ne "main::" && $sym ne "<none>::" && &$recurse($sym)) {
229 walksymtable(\%$fullname, $method, $recurse, $sym);
232 svref_2object(\*$fullname)->$method();
243 my ($class, $section, $symtable, $default) = @_;
244 $output_fh ||= FileHandle->new_tmpfile;
245 my $obj = bless [-1, $section, $symtable, $default], $class;
246 $sections{$section} = $obj;
251 my ($class, $section) = @_;
252 return $sections{$section};
257 while (defined($_ = shift)) {
258 print $output_fh "$section->[1]\t$_\n";
265 return $section->[0];
270 return $section->[1];
275 return $section->[2];
280 return $section->[3];
284 my ($section, $fh, $format) = @_;
285 my $name = $section->name;
286 my $sym = $section->symtable || {};
287 my $default = $section->default;
289 seek($output_fh, 0, 0);
290 while (<$output_fh>) {
295 exists($sym->{$1}) ? $sym->{$1} : $default;
297 printf $fh $format, $_;
311 B - The Perl Compiler
319 The C<B> module supplies classes which allow a Perl program to delve
320 into its own innards. It is the module used to implement the
321 "backends" of the Perl compiler. Usage of the compiler does not
322 require knowledge of this module: see the F<O> module for the
323 user-visible part. The C<B> module is of use to those who want to
324 write new compiler backends. This documentation assumes that the
325 reader knows a fair amount about perl's internals including such
326 things as SVs, OPs and the internal symbol table and syntax tree
331 The C<B> module contains a set of utility functions for querying the
332 current state of the Perl interpreter; typically these functions
333 return objects from the B::SV and B::OP classes, or their derived
334 classes. These classes in turn define methods for querying the
335 resulting objects about their own internal state.
337 =head1 Utility Functions
339 The C<B> module exports a variety of functions: some are simple
340 utility functions, others provide a Perl program with a way to
341 get an initial "handle" on an internal object.
343 =head2 Functions Returning C<B::SV>, C<B::AV>, C<B::HV>, and C<B::CV> objects
345 For descriptions of the class hierarchy of these objects and the
346 methods that can be called on them, see below, L<"OVERVIEW OF
347 CLASSES"> and L<"SV-RELATED CLASSES">.
353 Returns the SV object corresponding to the C variable C<sv_undef>.
357 Returns the SV object corresponding to the C variable C<sv_yes>.
361 Returns the SV object corresponding to the C variable C<sv_no>.
363 =item svref_2object(SVREF)
365 Takes a reference to any Perl value, and turns the referred-to value
366 into an object in the appropriate B::OP-derived or B::SV-derived
367 class. Apart from functions such as C<main_root>, this is the primary
368 way to get an initial "handle" on an internal perl data structure
369 which can then be followed with the other access methods.
371 The returned object will only be valid as long as the underlying OPs
372 and SVs continue to exist. Do not attempt to use the object after the
373 underlying structures are freed.
375 =item amagic_generation
377 Returns the SV object corresponding to the C variable C<amagic_generation>.
381 Returns the AV object (i.e. in class B::AV) representing INIT blocks.
385 Returns the AV object (i.e. in class B::AV) representing CHECK blocks.
389 Returns the AV object (i.e. in class B::AV) representing BEGIN blocks.
393 Returns the AV object (i.e. in class B::AV) representing END blocks.
397 Returns the AV object (i.e. in class B::AV) of the global comppadlist.
401 Only when perl was compiled with ithreads.
405 Return the (faked) CV corresponding to the main part of the Perl
410 =head2 Functions for Examining the Symbol Table
414 =item walksymtable(SYMREF, METHOD, RECURSE, PREFIX)
416 Walk the symbol table starting at SYMREF and call METHOD on each
417 symbol (a B::GV object) visited. When the walk reaches package
418 symbols (such as "Foo::") it invokes RECURSE, passing in the symbol
419 name, and only recurses into the package if that sub returns true.
421 PREFIX is the name of the SYMREF you're walking.
425 # Walk CGI's symbol table calling print_subs on each symbol.
426 # Recurse only into CGI::Util::
427 walksymtable(\%CGI::, 'print_subs', sub { $_[0] eq 'CGI::Util::' },
430 print_subs() is a B::GV method you have declared. Also see L<"B::GV
435 =head2 Functions Returning C<B::OP> objects or for walking op trees
437 For descriptions of the class hierarchy of these objects and the
438 methods that can be called on them, see below, L<"OVERVIEW OF
439 CLASSES"> and L<"OP-RELATED CLASSES">.
445 Returns the root op (i.e. an object in the appropriate B::OP-derived
446 class) of the main part of the Perl program.
450 Returns the starting op of the main part of the Perl program.
452 =item walkoptree(OP, METHOD)
454 Does a tree-walk of the syntax tree based at OP and calls METHOD on
455 each op it visits. Each node is visited before its children. If
456 C<walkoptree_debug> (see below) has been called to turn debugging on then
457 the method C<walkoptree_debug> is called on each op before METHOD is
460 =item walkoptree_debug(DEBUG)
462 Returns the current debugging flag for C<walkoptree>. If the optional
463 DEBUG argument is non-zero, it sets the debugging flag to that. See
464 the description of C<walkoptree> above for what the debugging flag
469 =head2 Miscellaneous Utility Functions
475 Return the PP function name (e.g. "pp_add") of op number OPNUM.
479 Returns a string in the form "0x..." representing the value of the
480 internal hash function used by perl on string STR.
484 Casts I to the internal I32 type used by that perl.
488 Does the equivalent of the C<-c> command-line option. Obviously, this
489 is only useful in a BEGIN block or else the flag is set too late.
493 Returns a double-quote-surrounded escaped version of STR which can
494 be used as a string in C source code.
496 =item perlstring(STR)
498 Returns a double-quote-surrounded escaped version of STR which can
499 be used as a string in Perl source code.
503 Returns the class of an object without the part of the classname
504 preceding the first C<"::">. This is used to turn C<"B::UNOP"> into
505 C<"UNOP"> for example.
509 In a perl compiled for threads, this returns a list of the special
510 per-thread threadsv variables.
517 =head1 OVERVIEW OF CLASSES
519 The C structures used by Perl's internals to hold SV and OP
520 information (PVIV, AV, HV, ..., OP, SVOP, UNOP, ...) are modelled on a
521 class hierarchy and the C<B> module gives access to them via a true
522 object hierarchy. Structure fields which point to other objects
523 (whether types of SV or types of OP) are represented by the C<B>
524 module as Perl objects of the appropriate class.
526 The bulk of the C<B> module is the methods for accessing fields of
529 Note that all access is read-only. You cannot modify the internals by
530 using this module. Also, note that the B::OP and B::SV objects created
531 by this module are only valid for as long as the underlying objects
532 exist; their creation doesn't increase the reference counts of the
533 underlying objects. Trying to access the fields of a freed object will
534 give incomprehensible results, or worse.
536 =head2 SV-RELATED CLASSES
538 B::IV, B::NV, B::RV, B::PV, B::PVIV, B::PVNV, B::PVMG, B::BM, B::PVLV,
539 B::AV, B::HV, B::CV, B::GV, B::FM, B::IO. These classes correspond in
540 the obvious way to the underlying C structures of similar names. The
541 inheritance hierarchy mimics the underlying C "inheritance". For 5.9.1
546 +--------------+----------+------------+
548 B::PV B::IV B::NV B::RV
560 +-----+----+------+-----+-----+
562 B::BM B::AV B::GV B::HV B::CV B::IO
568 For 5.9.0 and earlier, PVLV is a direct subclass of PVMG, so the base
574 +------+-----+----+------+-----+-----+
576 B::PVLV B::BM B::AV B::GV B::HV B::CV B::IO
582 Access methods correspond to the underlying C macros for field access,
583 usually with the leading "class indication" prefix removed (Sv, Av,
584 Hv, ...). The leading prefix is only left in cases where its removal
585 would cause a clash in method name. For example, C<GvREFCNT> stays
586 as-is since its abbreviation would clash with the "superclass" method
587 C<REFCNT> (corresponding to the C function C<SvREFCNT>).
599 Returns a reference to the regular scalar corresponding to this
600 B::SV object. In other words, this method is the inverse operation
601 to the svref_2object() subroutine. This scalar and other data it points
602 at should be considered read-only: modifying them is neither safe nor
603 guaranteed to have a sensible effect.
613 Returns the value of the IV, I<interpreted as
614 a signed integer>. This will be misleading
615 if C<FLAGS & SVf_IVisUV>. Perhaps you want the
616 C<int_value> method instead?
624 This method returns the value of the IV as an integer.
625 It differs from C<IV> in that it returns the correct
626 value regardless of whether it's stored signed or
659 This method is the one you usually want. It constructs a
660 string using the length and offset information in the struct:
661 for ordinary scalars it will return the string that you'd see
662 from Perl, even if it contains null characters.
666 Same as B::RV::RV, except that it will die() if the PV isn't
671 This method is less often useful. It assumes that the string
672 stored in the struct is null-terminated, and disregards the
675 It is the appropriate method to use if you need to get the name
676 of a lexical variable from a padname array. Lexical variable names
677 are always stored with a null terminator, and the length field
678 (SvCUR) is overloaded for other purposes and can't be relied on here.
682 =head2 B::PVMG Methods
692 =head2 B::MAGIC Methods
700 Only valid on r-magic, returns the string that generated the regexp.
710 Will die() if called on r-magic.
716 Only valid on r-magic, returns the integer value of the REGEX stored
721 =head2 B::PVLV Methods
755 This method returns TRUE if the GP field of the GV is NULL.
761 This method returns the name of the glob, but if the first
762 character of the name is a control character, then it converts
763 it to ^X first, so that *^G would return "^G" rather than "\cG".
765 It's useful if you want to print out the name of a variable.
766 If you restrict yourself to globs which exist at compile-time
767 then the result ought to be unambiguous, because code like
768 C<${"^G"} = 1> is compiled as two ops - a constant string and
769 a dereference (rv2gv) - so that the glob is created at runtime.
771 If you're working with globs at runtime, and need to disambiguate
772 *^G from *{"^G"}, then you should use the raw NAME method.
836 Takes one arguments ( 'stdin' | 'stdout' | 'stderr' ) and returns true
837 if the IoIFP of the object is equal to the handle whose name was
838 passed as argument ( i.e. $io->IsSTD('stderr') is true if
839 IoIFP($io) == PerlIO_stdin() ).
857 Like C<ARRAY>, but takes an index as an argument to get only one element,
858 rather than a list of all of them.
890 For constant subroutines, returns the constant SV returned by the subroutine.
918 =head2 OP-RELATED CLASSES
920 C<B::OP>, C<B::UNOP>, C<B::BINOP>, C<B::LOGOP>, C<B::LISTOP>, C<B::PMOP>,
921 C<B::SVOP>, C<B::PADOP>, C<B::PVOP>, C<B::LOOP>, C<B::COP>.
923 These classes correspond in the obvious way to the underlying C
924 structures of similar names. The inheritance hierarchy mimics the
925 underlying C "inheritance":
929 +---------------+--------+--------+
931 B::UNOP B::SVOP B::PADOP B::COP
942 Access methods correspond to the underlying C structre field names,
943 with the leading "class indication" prefix (C<"op_">) removed.
947 These methods get the values of similarly named fields within the OP
948 data structure. See top of C<op.h> for more info.
958 This returns the op name as a string (e.g. "add", "rv2av").
962 This returns the function name as a string (e.g. "PL_ppaddr[OP_ADD]",
963 "PL_ppaddr[OP_RV2AV]").
967 This returns the op description from the global C PL_op_desc array
968 (e.g. "addition" "array deref").
986 =head2 B::UNOP METHOD
994 =head2 B::BINOP METHOD
1002 =head2 B::LOGOP METHOD
1010 =head2 B::LISTOP METHOD
1018 =head2 B::PMOP Methods
1040 Only when perl was compiled with ithreads.
1044 =head2 B::SVOP METHOD
1054 =head2 B::PADOP METHOD
1062 =head2 B::PVOP METHOD
1070 =head2 B::LOOP Methods
1082 =head2 B::COP Methods
1109 Malcolm Beattie, C<mbeattie@sable.ox.ac.uk>