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 = '1.05';
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::IV';
35 @B::RV::ISA = 'B::SV';
36 @B::PVIV::ISA = qw(B::PV B::IV);
37 @B::PVNV::ISA = qw(B::PV B::NV);
38 @B::PVMG::ISA = 'B::PVNV';
39 @B::PVLV::ISA = 'B::GV';
40 @B::BM::ISA = 'B::PVMG';
41 @B::AV::ISA = 'B::PVMG';
42 @B::GV::ISA = 'B::PVMG';
43 @B::HV::ISA = 'B::PVMG';
44 @B::CV::ISA = 'B::PVMG';
45 @B::IO::ISA = 'B::PVMG';
46 @B::FM::ISA = 'B::CV';
48 @B::OP::ISA = 'B::OBJECT';
49 @B::UNOP::ISA = 'B::OP';
50 @B::BINOP::ISA = 'B::UNOP';
51 @B::LOGOP::ISA = 'B::UNOP';
52 @B::LISTOP::ISA = 'B::BINOP';
53 @B::SVOP::ISA = 'B::OP';
54 @B::PADOP::ISA = 'B::OP';
55 @B::PVOP::ISA = 'B::OP';
56 @B::LOOP::ISA = 'B::LISTOP';
57 @B::PMOP::ISA = 'B::LISTOP';
58 @B::COP::ISA = 'B::OP';
60 @B::SPECIAL::ISA = 'B::OBJECT';
63 # Stop "-w" from complaining about the lack of a real B::OBJECT class
68 my $name = (shift())->NAME;
70 # The regex below corresponds to the isCONTROLVAR macro
73 $name =~ s/^([\cA-\cZ\c\\c[\c]\c?\c_\c^])/"^".
74 chr( utf8::unicode_to_native( 64 ^ ord($1) ))/e;
76 # When we say unicode_to_native we really mean ascii_to_native,
77 # which matters iff this is a non-ASCII platform (EBCDIC).
82 sub B::IV::int_value {
84 return (($self->FLAGS() & SVf_IVisUV()) ? $self->UVX : $self->IV);
87 sub B::NULL::as_string() {""}
88 sub B::IV::as_string() {goto &B::IV::int_value}
89 sub B::PV::as_string() {goto &B::PV::PV}
96 my ($class, $value) = @_;
98 walkoptree_debug($value);
108 sub parents { \@parents }
113 return sprintf("%s (0x%x) %s", class($op), $$op, $op->name);
116 sub walkoptree_slow {
117 my($op, $method, $level) = @_;
118 $op_count++; # just for statistics
120 warn(sprintf("walkoptree: %d. %s\n", $level, peekop($op))) if $debug;
121 $op->$method($level);
122 if ($$op && ($op->flags & OPf_KIDS)) {
124 unshift(@parents, $op);
125 for ($kid = $op->first; $$kid; $kid = $kid->sibling) {
126 walkoptree_slow($kid, $method, $level + 1);
130 if (class($op) eq 'PMOP' && $op->pmreplroot && ${$op->pmreplroot}) {
131 unshift(@parents, $op);
132 walkoptree_slow($op->pmreplroot, $method, $level + 1);
138 return "Total number of OPs processed: $op_count\n";
142 my ($sec, $min, $hr) = localtime;
143 my ($user, $sys) = times;
144 sprintf("%02d:%02d:%02d user=$user sys=$sys",
145 $hr, $min, $sec, $user, $sys);
155 my ($obj, $value) = @_;
156 # warn(sprintf("savesym: sym_%x => %s\n", $$obj, $value)); # debug
157 $symtable{sprintf("sym_%x", $$obj)} = $value;
162 return $symtable{sprintf("sym_%x", $$obj)};
165 sub walkoptree_exec {
166 my ($op, $method, $level) = @_;
169 my $prefix = " " x $level;
170 for (; $$op; $op = $op->next) {
173 print $prefix, "goto $sym\n";
176 savesym($op, sprintf("%s (0x%lx)", class($op), $$op));
177 $op->$method($level);
180 /^(d?or(assign)?|and(assign)?|mapwhile|grepwhile|entertry|range|cond_expr)$/)
182 print $prefix, uc($1), " => {\n";
183 walkoptree_exec($op->other, $method, $level + 1);
184 print $prefix, "}\n";
185 } elsif ($ppname eq "match" || $ppname eq "subst") {
186 my $pmreplstart = $op->pmreplstart;
188 print $prefix, "PMREPLSTART => {\n";
189 walkoptree_exec($pmreplstart, $method, $level + 1);
190 print $prefix, "}\n";
192 } elsif ($ppname eq "substcont") {
193 print $prefix, "SUBSTCONT => {\n";
194 walkoptree_exec($op->other->pmreplstart, $method, $level + 1);
195 print $prefix, "}\n";
197 } elsif ($ppname eq "enterloop") {
198 print $prefix, "REDO => {\n";
199 walkoptree_exec($op->redoop, $method, $level + 1);
200 print $prefix, "}\n", $prefix, "NEXT => {\n";
201 walkoptree_exec($op->nextop, $method, $level + 1);
202 print $prefix, "}\n", $prefix, "LAST => {\n";
203 walkoptree_exec($op->lastop, $method, $level + 1);
204 print $prefix, "}\n";
205 } elsif ($ppname eq "subst") {
206 my $replstart = $op->pmreplstart;
208 print $prefix, "SUBST => {\n";
209 walkoptree_exec($replstart, $method, $level + 1);
210 print $prefix, "}\n";
217 my ($symref, $method, $recurse, $prefix) = @_;
222 $prefix = '' unless defined $prefix;
223 while (($sym, $ref) = each %$symref) {
224 $fullname = "*main::".$prefix.$sym;
226 $sym = $prefix . $sym;
227 if ($sym ne "main::" && $sym ne "<none>::" && &$recurse($sym)) {
228 walksymtable(\%$fullname, $method, $recurse, $sym);
231 svref_2object(\*$fullname)->$method();
242 my ($class, $section, $symtable, $default) = @_;
243 $output_fh ||= FileHandle->new_tmpfile;
244 my $obj = bless [-1, $section, $symtable, $default], $class;
245 $sections{$section} = $obj;
250 my ($class, $section) = @_;
251 return $sections{$section};
256 while (defined($_ = shift)) {
257 print $output_fh "$section->[1]\t$_\n";
264 return $section->[0];
269 return $section->[1];
274 return $section->[2];
279 return $section->[3];
283 my ($section, $fh, $format) = @_;
284 my $name = $section->name;
285 my $sym = $section->symtable || {};
286 my $default = $section->default;
288 seek($output_fh, 0, 0);
289 while (<$output_fh>) {
294 exists($sym->{$1}) ? $sym->{$1} : $default;
296 printf $fh $format, $_;
310 B - The Perl Compiler
318 The C<B> module supplies classes which allow a Perl program to delve
319 into its own innards. It is the module used to implement the
320 "backends" of the Perl compiler. Usage of the compiler does not
321 require knowledge of this module: see the F<O> module for the
322 user-visible part. The C<B> module is of use to those who want to
323 write new compiler backends. This documentation assumes that the
324 reader knows a fair amount about perl's internals including such
325 things as SVs, OPs and the internal symbol table and syntax tree
330 The C<B> module contains a set of utility functions for querying the
331 current state of the Perl interpreter; typically these functions
332 return objects from the B::SV and B::OP classes, or their derived
333 classes. These classes in turn define methods for querying the
334 resulting objects about their own internal state.
336 =head1 Utility Functions
338 The C<B> module exports a variety of functions: some are simple
339 utility functions, others provide a Perl program with a way to
340 get an initial "handle" on an internal object.
342 =head2 Functions Returning C<B::SV>, C<B::AV>, C<B::HV>, and C<B::CV> objects
344 For descriptions of the class hierachy of these objects and the
345 methods that can be called on them, see below, L<"OVERVIEW OF
346 CLASSES"> and L<"SV-RELATED CLASSES">.
352 Returns the SV object corresponding to the C variable C<sv_undef>.
356 Returns the SV object corresponding to the C variable C<sv_yes>.
360 Returns the SV object corresponding to the C variable C<sv_no>.
362 =item svref_2object(SVREF)
364 Takes a reference to any Perl value, and turns the referred-to value
365 into an object in the appropriate B::OP-derived or B::SV-derived
366 class. Apart from functions such as C<main_root>, this is the primary
367 way to get an initial "handle" on an internal perl data structure
368 which can then be followed with the other access methods.
370 =item amagic_generation
372 Returns the SV object corresponding to the C variable C<amagic_generation>.
376 Returns the AV object (i.e. in class B::AV) representing INIT blocks.
380 Returns the AV object (i.e. in class B::AV) representing CHECK blocks.
384 Returns the AV object (i.e. in class B::AV) representing BEGIN blocks.
388 Returns the AV object (i.e. in class B::AV) representing END blocks.
392 Returns the AV object (i.e. in class B::AV) of the global comppadlist.
396 Only when perl was compiled with ithreads.
400 Return the (faked) CV corresponding to the main part of the Perl
405 =head2 Functions for Examining the Symbol Table
409 =item walksymtable(SYMREF, METHOD, RECURSE, PREFIX)
411 Walk the symbol table starting at SYMREF and call METHOD on each
412 symbol (a B::GV object) visited. When the walk reaches package
413 symbols (such as "Foo::") it invokes RECURSE, passing in the symbol
414 name, and only recurses into the package if that sub returns true.
416 PREFIX is the name of the SYMREF you're walking.
420 # Walk CGI's symbol table calling print_subs on each symbol.
421 # Recurse only into CGI::Util::
422 walksymtable(\%CGI::, 'print_subs', sub { $_[0] eq 'CGI::Util::' },
425 print_subs() is a B::GV method you have declared. Also see L<"B::GV
430 =head2 Functions Returning C<B::OP> objects or for walking op trees
432 For descriptions of the class hierachy of these objects and the
433 methods that can be called on them, see below, L<"OVERVIEW OF
434 CLASSES"> and L<"OP-RELATED CLASSES">.
440 Returns the root op (i.e. an object in the appropriate B::OP-derived
441 class) of the main part of the Perl program.
445 Returns the starting op of the main part of the Perl program.
447 =item walkoptree(OP, METHOD)
449 Does a tree-walk of the syntax tree based at OP and calls METHOD on
450 each op it visits. Each node is visited before its children. If
451 C<walkoptree_debug> (see below) has been called to turn debugging on then
452 the method C<walkoptree_debug> is called on each op before METHOD is
455 =item walkoptree_debug(DEBUG)
457 Returns the current debugging flag for C<walkoptree>. If the optional
458 DEBUG argument is non-zero, it sets the debugging flag to that. See
459 the description of C<walkoptree> above for what the debugging flag
464 =head2 Miscellaneous Utility Functions
470 Return the PP function name (e.g. "pp_add") of op number OPNUM.
474 Returns a string in the form "0x..." representing the value of the
475 internal hash function used by perl on string STR.
479 Casts I to the internal I32 type used by that perl.
483 Does the equivalent of the C<-c> command-line option. Obviously, this
484 is only useful in a BEGIN block or else the flag is set too late.
488 Returns a double-quote-surrounded escaped version of STR which can
489 be used as a string in C source code.
491 =item perlstring(STR)
493 Returns a double-quote-surrounded escaped version of STR which can
494 be used as a string in Perl source code.
498 Returns the class of an object without the part of the classname
499 preceding the first C<"::">. This is used to turn C<"B::UNOP"> into
500 C<"UNOP"> for example.
504 In a perl compiled for threads, this returns a list of the special
505 per-thread threadsv variables.
512 =head1 OVERVIEW OF CLASSES
514 The C structures used by Perl's internals to hold SV and OP
515 information (PVIV, AV, HV, ..., OP, SVOP, UNOP, ...) are modelled on a
516 class hierarchy and the C<B> module gives access to them via a true
517 object hierarchy. Structure fields which point to other objects
518 (whether types of SV or types of OP) are represented by the C<B>
519 module as Perl objects of the appropriate class.
521 The bulk of the C<B> module is the methods for accessing fields of
524 Note that all access is read-only. You cannot modify the internals by
527 =head2 SV-RELATED CLASSES
529 B::IV, B::NV, B::RV, B::PV, B::PVIV, B::PVNV, B::PVMG, B::BM, B::PVLV,
530 B::AV, B::HV, B::CV, B::GV, B::FM, B::IO. These classes correspond in
531 the obvious way to the underlying C structures of similar names. The
532 inheritance hierarchy mimics the underlying C "inheritance":
536 +--------------+----------------------+
550 +-----+----+------+-----+-----+
552 B::BM B::AV B::GV B::HV B::CV B::IO
558 Access methods correspond to the underlying C macros for field access,
559 usually with the leading "class indication" prefix removed (Sv, Av,
560 Hv, ...). The leading prefix is only left in cases where its removal
561 would cause a clash in method name. For example, C<GvREFCNT> stays
562 as-is since its abbreviation would clash with the "superclass" method
563 C<REFCNT> (corresponding to the C function C<SvREFCNT>).
575 Returns a reference to the regular scalar corresponding to this
576 B::SV object. In other words, this method is the inverse operation
577 to the svref_2object() subroutine. This scalar and other data it points
578 at should be considered read-only: modifying them is neither safe nor
579 guaranteed to have a sensible effect.
589 Returns the value of the IV, I<interpreted as
590 a signed integer>. This will be misleading
591 if C<FLAGS & SVf_IVisUV>. Perhaps you want the
592 C<int_value> method instead?
600 This method returns the value of the IV as an integer.
601 It differs from C<IV> in that it returns the correct
602 value regardless of whether it's stored signed or
635 This method is the one you usually want. It constructs a
636 string using the length and offset information in the struct:
637 for ordinary scalars it will return the string that you'd see
638 from Perl, even if it contains null characters.
642 Same as B::RV::RV, except that it will die() if the PV isn't
647 This method is less often useful. It assumes that the string
648 stored in the struct is null-terminated, and disregards the
651 It is the appropriate method to use if you need to get the name
652 of a lexical variable from a padname array. Lexical variable names
653 are always stored with a null terminator, and the length field
654 (SvCUR) is overloaded for other purposes and can't be relied on here.
658 =head2 B::PVMG Methods
668 =head2 B::MAGIC Methods
676 Only valid on r-magic, returns the string that generated the regexp.
686 Will die() if called on r-magic.
692 Only valid on r-magic, returns the integer value of the REGEX stored
697 =head2 B::PVLV Methods
731 This method returns TRUE if the GP field of the GV is NULL.
737 This method returns the name of the glob, but if the first
738 character of the name is a control character, then it converts
739 it to ^X first, so that *^G would return "^G" rather than "\cG".
741 It's useful if you want to print out the name of a variable.
742 If you restrict yourself to globs which exist at compile-time
743 then the result ought to be unambiguous, because code like
744 C<${"^G"} = 1> is compiled as two ops - a constant string and
745 a dereference (rv2gv) - so that the glob is created at runtime.
747 If you're working with globs at runtime, and need to disambiguate
748 *^G from *{"^G"}, then you should use the raw NAME method.
812 Takes one arguments ( 'stdin' | 'stdout' | 'stderr' ) and returns true
813 if the IoIFP of the object is equal to the handle whose name was
814 passed as argument ( i.e. $io->IsSTD('stderr') is true if
815 IoIFP($io) == PerlIO_stdin() ).
833 Like C<ARRAY>, but takes an index as an argument to get only one element,
834 rather than a list of all of them.
866 For constant subroutines, returns the constant SV returned by the subroutine.
894 =head2 OP-RELATED CLASSES
896 C<B::OP>, C<B::UNOP>, C<B::BINOP>, C<B::LOGOP>, C<B::LISTOP>, C<B::PMOP>,
897 C<B::SVOP>, C<B::PADOP>, C<B::PVOP>, C<B::LOOP>, C<B::COP>.
899 These classes correspond in the obvious way to the underlying C
900 structures of similar names. The inheritance hierarchy mimics the
901 underlying C "inheritance":
905 +---------------+--------+--------+
907 B::UNOP B::SVOP B::PADOP B::COP
918 Access methods correspond to the underlying C structre field names,
919 with the leading "class indication" prefix (C<"op_">) removed.
931 This returns the op name as a string (e.g. "add", "rv2av").
935 This returns the function name as a string (e.g. "PL_ppaddr[OP_ADD]",
936 "PL_ppaddr[OP_RV2AV]").
940 This returns the op description from the global C PL_op_desc array
941 (e.g. "addition" "array deref").
955 =head2 B::UNOP METHOD
963 =head2 B::BINOP METHOD
971 =head2 B::LOGOP METHOD
979 =head2 B::LISTOP METHOD
987 =head2 B::PMOP Methods
1009 Only when perl was compiled with ithreads.
1013 =head2 B::SVOP METHOD
1023 =head2 B::PADOP METHOD
1031 =head2 B::PVOP METHOD
1039 =head2 B::LOOP Methods
1051 =head2 B::COP Methods
1078 Malcolm Beattie, C<mbeattie@sable.ox.ac.uk>