6 BEGIN { *DEBUG = sub () {0} unless defined &DEBUG }
8 use Pod::Escapes 1.03 ();
9 use Pod::Simple::LinkSection ();
10 use Pod::Simple::BlackBox ();
15 @Known_formatting_codes @Known_directives
16 %Known_formatting_codes %Known_directives
20 @ISA = ('Pod::Simple::BlackBox');
23 @Known_formatting_codes = qw(I B C L E F S X Z);
24 %Known_formatting_codes = map(($_=>1), @Known_formatting_codes);
25 @Known_directives = qw(head1 head2 head3 head4 item over back);
26 %Known_directives = map(($_=>'Plain'), @Known_directives);
27 $NL = $/ unless defined $NL;
29 #-----------------------------------------------------------------------------
30 # Set up some constants:
33 if(defined &ASCII) { }
34 elsif(chr(65) eq 'A') { *ASCII = sub () {1} }
35 else { *ASCII = sub () {''} }
37 unless(defined &MANY_LINES) { *MANY_LINES = sub () {20} }
38 DEBUG > 4 and print "MANY_LINES is ", MANY_LINES(), "\n";
39 unless(MANY_LINES() >= 1) {
40 die "MANY_LINES is too small (", MANY_LINES(), ")!\nAborting";
42 if(defined &UNICODE) { }
43 elsif($] >= 5.008) { *UNICODE = sub() {1} }
44 else { *UNICODE = sub() {''} }
47 print "# We are ", ASCII ? '' : 'not ', "in ASCII-land\n";
48 print "# We are under a Unicode-safe Perl.\n";
52 # This is a parser for Pod. It is not a parser for the set of Pod-like
53 # languages which happens to contain Pod -- it is just for Pod, plus possibly
56 # @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @
57 #@ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @ @
58 #@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
60 __PACKAGE__->_accessorize(
61 'nbsp_for_S', # Whether to map S<...>'s to \xA0 characters
62 'source_filename', # Filename of the source, for use in warnings
63 'source_dead', # Whether to consider this parser's source dead
65 'output_fh', # The filehandle we're writing to, if applicable.
66 # Used only in some derived classes.
68 'hide_line_numbers', # For some dumping subclasses: whether to pointedly
69 # suppress the start_line attribute
71 'line_count', # the current line number
72 'pod_para_count', # count of pod paragraphs seen so far
74 'no_whining', # whether to suppress whining
75 'no_errata_section', # whether to suppress the errata section
76 'complain_stderr', # whether to complain to stderr
78 'doc_has_started', # whether we've fired the open-Document event yet
80 'bare_output', # For some subclasses: whether to prepend
81 # header-code and postpend footer-code
83 'fullstop_space_harden', # Whether to turn ". " into ".[nbsp] ";
85 'nix_X_codes', # whether to ignore X<...> codes
86 'merge_text', # whether to avoid breaking a single piece of
87 # text up into several events
89 'preserve_whitespace', # whether to try to keep whitespace as-is
91 'content_seen', # whether we've seen any real Pod content
92 'errors_seen', # TODO: document. whether we've seen any errors (fatal or not)
94 'codes_in_verbatim', # for PseudoPod extensions
96 'code_handler', # coderef to call when a code (non-pod) line is seen
97 'cut_handler', # coderef to call when a =cut line is seen
99 # $code_handler->($line, $self->{'line_count'}, $self) if $code_handler;
100 # $cut_handler->($line, $self->{'line_count'}, $self) if $cut_handler;
104 #@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
106 sub any_errata_seen { # good for using as an exit() value...
107 return shift->{'errors_seen'} || 0;
110 #@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
111 # Pull in some functions that, for some reason, I expect to see here too:
113 *pretty = \&Pod::Simple::BlackBox::pretty;
114 *stringify_lol = \&Pod::Simple::BlackBox::stringify_lol;
117 #@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
120 my $class = ref($_[0]) || $_[0];
121 if($class eq __PACKAGE__) {
122 return "$class $VERSION";
124 my $v = $class->VERSION;
125 return "$class $v (" . __PACKAGE__ . " $VERSION)";
129 #@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
131 #sub curr_open { # read-only list accessor
132 # return @{ $_[0]{'curr_open'} || return() };
134 #sub _curr_open_listref { $_[0]{'curr_open'} ||= [] }
138 # Works by faking out output_fh. Simplifies our code.
141 return $this->{'output_string'} unless @_; # GET.
143 require Pod::Simple::TiedOutFH;
144 my $x = (defined($_[0]) and ref($_[0])) ? $_[0] : \( $_[0] );
145 $$x = '' unless defined $$x;
146 DEBUG > 4 and print "# Output string set to $x ($$x)\n";
147 $this->{'output_fh'} = Pod::Simple::TiedOutFH->handle_on($_[0]);
149 $this->{'output_string'} = $_[0];
150 #${ ${ $this->{'output_fh'} } };
153 sub abandon_output_string { $_[0]->abandon_output_fh; delete $_[0]{'output_string'} }
154 sub abandon_output_fh { $_[0]->output_fh(undef) }
155 # These don't delete the string or close the FH -- they just delete our
156 # references to it/them.
157 # TODO: document these
159 #@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
162 # takes no parameters
163 my $class = ref($_[0]) || $_[0];
164 #Carp::croak(__PACKAGE__ . " is a virtual base class -- see perldoc "
167 'accept_codes' => { map( ($_=>$_), @Known_formatting_codes ) },
168 'accept_directives' => { %Known_directives },
169 'accept_targets' => {},
175 # TODO: an option for whether to interpolate E<...>'s, or just resolve to codes.
177 #@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
179 sub _handle_element_start { # OVERRIDE IN DERIVED CLASS
180 my($self, $element_name, $attr_hash_r) = @_;
184 sub _handle_element_end { # OVERRIDE IN DERIVED CLASS
185 my($self, $element_name) = @_;
189 sub _handle_text { # OVERRIDE IN DERIVED CLASS
190 my($self, $text) = @_;
194 #@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
196 # And now directives (not targets)
198 sub accept_directive_as_verbatim { shift->_accept_directives('Verbatim', @_) }
199 sub accept_directive_as_data { shift->_accept_directives('Data', @_) }
200 sub accept_directive_as_processed { shift->_accept_directives('Plain', @_) }
202 sub _accept_directives {
203 my($this, $type) = splice @_,0,2;
205 next unless defined $d and length $d;
206 Carp::croak "\"$d\" isn't a valid directive name"
207 unless $d =~ m/^[a-zA-Z][a-zA-Z0-9]*$/s;
208 Carp::croak "\"$d\" is already a reserved Pod directive name"
209 if exists $Known_directives{$d};
210 $this->{'accept_directives'}{$d} = $type;
211 DEBUG > 2 and print "Learning to accept \"=$d\" as directive of type $type\n";
213 DEBUG > 6 and print "$this\'s accept_directives : ",
214 pretty($this->{'accept_directives'}), "\n";
216 return sort keys %{ $this->{'accept_directives'} } if wantarray;
220 #--------------------------------------------------------------------------
221 # TODO: document these:
223 sub unaccept_directive { shift->unaccept_directives(@_) };
225 sub unaccept_directives {
228 next unless defined $d and length $d;
229 Carp::croak "\"$d\" isn't a valid directive name"
230 unless $d =~ m/^[a-zA-Z][a-zA-Z0-9]*$/s;
231 Carp::croak "But you must accept \"$d\" directives -- it's a builtin!"
232 if exists $Known_directives{$d};
233 delete $this->{'accept_directives'}{$d};
234 DEBUG > 2 and print "OK, won't accept \"=$d\" as directive.\n";
236 return sort keys %{ $this->{'accept_directives'} } if wantarray;
240 #@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
242 # And now targets (not directives)
244 sub accept_target { shift->accept_targets(@_) } # alias
245 sub accept_target_as_text { shift->accept_targets_as_text(@_) } # alias
248 sub accept_targets { shift->_accept_targets('1', @_) }
250 sub accept_targets_as_text { shift->_accept_targets('force_resolve', @_) }
251 # forces them to be processed, even when there's no ":".
253 sub _accept_targets {
254 my($this, $type) = splice @_,0,2;
256 next unless defined $t and length $t;
257 # TODO: enforce some limitations on what a target name can be?
258 $this->{'accept_targets'}{$t} = $type;
259 DEBUG > 2 and print "Learning to accept \"$t\" as target of type $type\n";
261 return sort keys %{ $this->{'accept_targets'} } if wantarray;
265 #--------------------------------------------------------------------------
266 sub unaccept_target { shift->unaccept_targets(@_) }
268 sub unaccept_targets {
271 next unless defined $t and length $t;
272 # TODO: enforce some limitations on what a target name can be?
273 delete $this->{'accept_targets'}{$t};
274 DEBUG > 2 and print "OK, won't accept \"$t\" as target.\n";
276 return sort keys %{ $this->{'accept_targets'} } if wantarray;
280 #@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
282 # And now codes (not targets or directives)
284 sub accept_code { shift->accept_codes(@_) } # alias
286 sub accept_codes { # Add some codes
289 foreach my $new_code (@_) {
290 next unless defined $new_code and length $new_code;
292 # A good-enough check that it's good as an XML Name symbol:
293 Carp::croak "\"$new_code\" isn't a valid element name"
295 m/[\x00-\x2C\x2F\x39\x3B-\x40\x5B-\x5E\x60\x7B-\x7F]/
296 # Characters under 0x80 that aren't legal in an XML Name.
297 or $new_code =~ m/^[-\.0-9]/s
298 or $new_code =~ m/:[-\.0-9]/s;
299 # The legal under-0x80 Name characters that
300 # an XML Name still can't start with.
303 $this->{'accept_codes'}{$new_code} = $new_code;
305 # Yes, map to itself -- just so that when we
306 # see "=extend W [whatever] thatelementname", we say that W maps
307 # to whatever $this->{accept_codes}{thatelementname} is,
308 # i.e., "thatelementname". Then when we go re-mapping,
309 # a "W" in the treelet turns into "thatelementname". We only
311 # If we say we accept "W", then a "W" in the treelet simply turns
318 #--------------------------------------------------------------------------
319 sub unaccept_code { shift->unaccept_codes(@_) }
321 sub unaccept_codes { # remove some codes
324 foreach my $new_code (@_) {
325 next unless defined $new_code and length $new_code;
327 # A good-enough check that it's good as an XML Name symbol:
328 Carp::croak "\"$new_code\" isn't a valid element name"
330 m/[\x00-\x2C\x2F\x39\x3B-\x40\x5B-\x5E\x60\x7B-\x7F]/
331 # Characters under 0x80 that aren't legal in an XML Name.
332 or $new_code =~ m/^[-\.0-9]/s
333 or $new_code =~ m/:[-\.0-9]/s;
334 # The legal under-0x80 Name characters that
335 # an XML Name still can't start with.
338 Carp::croak "But you must accept \"$new_code\" codes -- it's a builtin!"
339 if grep $new_code eq $_, @Known_formatting_codes;
341 delete $this->{'accept_codes'}{$new_code};
343 DEBUG > 2 and print "OK, won't accept the code $new_code<...>.\n";
350 #@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
351 #@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
353 sub parse_string_document {
356 foreach my $line_group (@_) {
357 next unless defined $line_group and length $line_group;
358 pos($line_group) = 0;
360 m/([^\n\r]*)((?:\r?\n)?)/g
363 $self->parse_lines($1)
364 if length($1) or length($2)
365 or pos($line_group) != length($line_group);
366 # I.e., unless it's a zero-length "empty line" at the very
367 # end of "foo\nbar\n" (i.e., between the \n and the EOS).
370 $self->parse_lines(undef); # to signal EOF
374 sub _init_fh_source {
375 my($self, $source) = @_;
377 #DEBUG > 1 and print "Declaring $source as :raw for starters\n";
378 #$self->_apply_binmode($source, ':raw');
379 #binmode($source, ":raw");
384 #:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.
388 my($self, $source) = (@_);
390 if(!defined $source) {
391 Carp::croak("Can't use empty-string as a source for parse_file");
392 } elsif(ref(\$source) eq 'GLOB') {
393 $self->{'source_filename'} = '' . ($source);
394 } elsif(ref $source) {
395 $self->{'source_filename'} = '' . ($source);
396 } elsif(!length $source) {
397 Carp::croak("Can't use empty-string as a source for parse_file");
401 open(PODSOURCE, "<$source") || Carp::croak("Can't open $source: $!");
402 $self->{'source_filename'} = $source;
403 $source = *PODSOURCE{IO};
405 $self->_init_fh_source($source);
407 # By here, $source is a FH.
409 $self->{'source_fh'} = $source;
412 until( $self->{'source_dead'} ) {
414 for($i = MANY_LINES; $i--;) { # read those many lines at a time
416 push @lines, scalar(<$source>); # readline
417 last unless defined $lines[-1];
418 # but pass thru the undef, which will set source_dead to true
420 $self->parse_lines(@lines);
422 delete($self->{'source_fh'}); # so it can be GC'd
426 #:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.
428 sub parse_from_file {
429 # An emulation of Pod::Parser's interface, for the sake of Perldoc.
430 # Basically just a wrapper around parse_file.
432 my($self, $source, $to) = @_;
433 $self = $self->new unless ref($self); # so we tolerate being a class method
435 if(!defined $source) { $source = *STDIN{IO}
436 } elsif(ref(\$source) eq 'GLOB') { # stet
437 } elsif(ref($source) ) { # stet
438 } elsif(!length $source
439 or $source eq '-' or $source =~ m/^<&(STDIN|0)$/i
441 $source = *STDIN{IO};
444 if(!defined $to) { $self->output_fh( *STDOUT{IO} );
445 } elsif(ref(\$to) eq 'GLOB') { $self->output_fh( $to );
446 } elsif(ref($to)) { $self->output_fh( $to );
448 or $to eq '-' or $to =~ m/^>&?(?:STDOUT|1)$/i
450 $self->output_fh( *STDOUT{IO} );
453 my $out_fh = Symbol::gensym();
454 DEBUG and print "Write-opening to $to\n";
455 open($out_fh, ">$to") or Carp::croak "Can't write-open $to: $!";
457 if $self->can('write_with_binmode') and $self->write_with_binmode;
458 $self->output_fh($out_fh);
461 return $self->parse_file($source);
464 #-----------------------------------------------------------------------------
467 #my($self,$line,$complaint) = @_;
468 my $self = shift(@_);
469 ++$self->{'errors_seen'};
470 if($self->{'no_whining'}) {
471 DEBUG > 9 and print "Discarding complaint (at line $_[0]) $_[1]\n because no_whining is on.\n";
474 return $self->_complain_warn(@_) if $self->{'complain_stderr'};
475 return $self->_complain_errata(@_);
478 sub scream { # like whine, but not suppressable
479 #my($self,$line,$complaint) = @_;
480 my $self = shift(@_);
481 ++$self->{'errors_seen'};
482 return $self->_complain_warn(@_) if $self->{'complain_stderr'};
483 return $self->_complain_errata(@_);
487 my($self,$line,$complaint) = @_;
488 return printf STDERR "%s around line %s: %s\n",
489 $self->{'source_filename'} || 'Pod input', $line, $complaint;
492 sub _complain_errata {
493 my($self,$line,$complaint) = @_;
494 if( $self->{'no_errata_section'} ) {
495 DEBUG > 9 and print "Discarding erratum (at line $line) $complaint\n because no_errata_section is on.\n";
497 DEBUG > 9 and print "Queuing erratum (at line $line) $complaint\n";
498 push @{$self->{'errata'}{$line}}, $complaint
499 # for a report to be generated later!
504 #@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
506 sub _get_initial_item_type {
507 # A hack-wrapper here for when you have like "=over\n\n=item 456\n\n"
508 my($self, $para) = @_;
509 return $para->[1]{'~type'} if $para->[1]{'~type'};
511 return $para->[1]{'~type'} = 'text'
512 if join("\n", @{$para}[2 .. $#$para]) =~ m/^\s*(\d+)\.?\s*$/s and $1 ne '1';
513 # Else fall thru to the general case:
514 return $self->_get_item_type($para);
519 sub _get_item_type { # mutates the item!!
520 my($self, $para) = @_;
521 return $para->[1]{'~type'} if $para->[1]{'~type'};
524 # Otherwise we haven't yet been to this node. Maybe alter it...
526 my $content = join "\n", @{$para}[2 .. $#$para];
528 if($content =~ m/^\s*\*\s*$/s or $content =~ m/^\s*$/s) {
529 # Like: "=item *", "=item * ", "=item"
530 splice @$para, 2; # so it ends up just being ['=item', { attrhash } ]
531 $para->[1]{'~orig_content'} = $content;
532 return $para->[1]{'~type'} = 'bullet';
534 } elsif($content =~ m/^\s*\*\s+(.+)/s) { # tolerance
536 # Like: "=item * Foo bar baz";
537 $para->[1]{'~orig_content'} = $content;
538 $para->[1]{'~_freaky_para_hack'} = $1;
539 DEBUG > 2 and print " Tolerating $$para[2] as =item *\\n\\n$1\n";
540 splice @$para, 2; # so it ends up just being ['=item', { attrhash } ]
541 return $para->[1]{'~type'} = 'bullet';
543 } elsif($content =~ m/^\s*(\d+)\.?\s*$/s) {
544 # Like: "=item 1.", "=item 123412"
546 $para->[1]{'~orig_content'} = $content;
547 $para->[1]{'number'} = $1; # Yes, stores the number there!
549 splice @$para, 2; # so it ends up just being ['=item', { attrhash } ]
550 return $para->[1]{'~type'} = 'number';
553 # It's anything else.
554 return $para->[1]{'~type'} = 'text';
559 #-----------------------------------------------------------------------------
562 my $self = shift; # and ($para, $start_line)
566 } if(ref $_[0] and ref $_[0][0] and $_[0][0][0] eq '~Top') {
567 # Hack so we can pass in fake-o pre-cooked paragraphs:
568 # just have the first line be a reference to a ['~Top', {}, ...]
569 # We use this feechure in gen_errata and stuff.
571 DEBUG and print "Applying precooked treelet hack to $_[0][0]\n";
573 splice @$treelet, 0, 2; # lop the top off
576 $treelet = $self->_treelet_from_formatting_codes(@_);
579 if( $self->_remap_sequences($treelet) ) {
580 $self->_treat_Zs($treelet); # Might as well nix these first
581 $self->_treat_Ls($treelet); # L has to precede E and S
582 $self->_treat_Es($treelet);
583 $self->_treat_Ss($treelet); # S has to come after E
585 $self->_wrap_up($treelet); # Nix X's and merge texties
588 DEBUG and print "Formatless treelet gets fast-tracked.\n";
592 splice @$treelet, 0, 2; # lop the top off
597 #:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.
600 my($self, @stack) = @_;
601 my $nixx = $self->{'nix_X_codes'};
602 my $merge = $self->{'merge_text' };
603 return unless $nixx or $merge;
605 DEBUG > 2 and print "\nStarting _wrap_up traversal.\n",
606 $merge ? (" Merge mode on\n") : (),
607 $nixx ? (" Nix-X mode on\n") : (),
612 while($treelet = shift @stack) {
613 DEBUG > 3 and print " Considering children of this $treelet->[0] node...\n";
614 for($i = 2; $i < @$treelet; ++$i) { # iterate over children
615 DEBUG > 3 and print " Considering child at $i ", pretty($treelet->[$i]), "\n";
616 if($nixx and ref $treelet->[$i] and $treelet->[$i][0] eq 'X') {
617 DEBUG > 3 and print " Nixing X node at $i\n";
618 splice(@$treelet, $i, 1); # just nix this node (and its descendants)
619 # no need to back-update the counter just yet
622 } elsif($merge and $i != 2 and # non-initial
623 !ref $treelet->[$i] and !ref $treelet->[$i - 1]
625 DEBUG > 3 and print " Merging ", $i-1,
626 ":[$treelet->[$i-1]] and $i\:[$treelet->[$i]]\n";
627 $treelet->[$i-1] .= ( splice(@$treelet, $i, 1) )[0];
628 DEBUG > 4 and print " Now: ", $i-1, ":[$treelet->[$i-1]]\n";
631 # since we just pulled the possibly last node out from under
632 # ourselves, we can't just redo()
634 } elsif( ref $treelet->[$i] ) {
635 DEBUG > 4 and print " Enqueuing ", pretty($treelet->[$i]), " for traversal.\n";
636 push @stack, $treelet->[$i];
638 if($treelet->[$i][0] eq 'L') {
640 foreach my $attrname ('section', 'to') {
641 if(defined($thing = $treelet->[$i][1]{$attrname}) and ref $thing) {
642 unshift @stack, $thing;
643 DEBUG > 4 and print " +Enqueuing ",
644 pretty( $treelet->[$i][1]{$attrname} ),
645 " as an attribute value to tweak.\n";
652 DEBUG > 2 and print "End of _wrap_up traversal.\n\n";
657 #:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.
659 sub _remap_sequences {
660 my($self,@stack) = @_;
662 if(@stack == 1 and @{ $stack[0] } == 3 and !ref $stack[0][2]) {
663 # VERY common case: abort it.
664 DEBUG and print "Skipping _remap_sequences: formatless treelet.\n";
668 my $map = ($self->{'accept_codes'} || die "NO accept_codes in $self?!?");
670 my $start_line = $stack[0][1]{'start_line'};
672 "\nAbout to start _remap_sequences on treelet from line %s.\n",
675 DEBUG > 3 and print " Map: ",
676 join('; ', map "$_=" . (
677 ref($map->{$_}) ? join(",", @{$map->{$_}}) : $map->{$_}
680 ("B~C~E~F~I~L~S~X~Z" eq join '~', sort keys %$map)
681 ? " (all normal)\n" : "\n"
684 # A recursive algorithm implemented iteratively! Whee!
686 my($is, $was, $i, $treelet); # scratch
687 while($treelet = shift @stack) {
688 DEBUG > 3 and print " Considering children of this $treelet->[0] node...\n";
689 for($i = 2; $i < @$treelet; ++$i) { # iterate over children
690 next unless ref $treelet->[$i]; # text nodes are uninteresting
692 DEBUG > 4 and print " Noting child $i : $treelet->[$i][0]<...>\n";
694 $is = $treelet->[$i][0] = $map->{ $was = $treelet->[$i][0] };
697 print " Code $was<> is UNKNOWN!\n";
698 } elsif($is eq $was) {
699 DEBUG > 4 and print " Code $was<> stays the same.\n";
701 print " Code $was<> maps to ",
703 ? ( "tags ", map("$_<", @$is), '...', map('>', @$is), "\n" )
709 $self->whine($start_line, "Deleting unknown formatting code $was<>");
710 $is = $treelet->[$i][0] = '1'; # But saving the children!
711 # I could also insert a leading "$was<" and tailing ">" as
712 # children of this node, but something about that seems icky.
716 DEBUG > 4 and print " Renaming $was node to $dynasty[-1]\n";
717 $treelet->[$i][0] = pop @dynasty;
721 " Grafting a new %s node between %s and %s\n",
722 $dynasty[-1], $treelet->[0], $treelet->[$i][0],
726 splice @$treelet, $i, 1, [pop(@dynasty), {}, $treelet->[$i]];
727 # relace node with a new parent
729 } elsif($is eq '0') {
730 splice(@$treelet, $i, 1); # just nix this node (and its descendants)
731 --$i; # back-update the counter
732 } elsif($is eq '1') {
733 splice(@$treelet, $i, 1 # replace this node with its children!
734 => splice @{ $treelet->[$i] },2
735 # (not catching its first two (non-child) items)
737 --$i; # back up for new stuff
739 # otherwise it's unremarkable
740 unshift @stack, $treelet->[$i]; # just recurse
745 DEBUG > 2 and print "End of _remap_sequences traversal.\n\n";
747 if(@_ == 2 and @{ $_[1] } == 3 and !ref $_[1][2]) {
748 DEBUG and print "Noting that the treelet is now formatless.\n";
754 # . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
758 # "Go to an extreme, move back to a more comfortable place"
759 # -- /Oblique Strategies/, Brian Eno and Peter Schmidt
761 my($self, $para) = @_;
762 my $content = join ' ', splice @$para, 2;
763 $content =~ s/^\s+//s;
764 $content =~ s/\s+$//s;
766 DEBUG > 2 and print "Ogling extensor: =extend $content\n";
772 (\S+) # 2 : fallback(s)
773 (?:\s+(\S+))? # 3 : element name(s)
779 my $fallbacks_one = $2;
781 $elements_one = defined($3) ? $3 : $1;
783 DEBUG > 2 and print "Extensor has good syntax.\n";
785 unless($new_letter =~ m/^[A-Z]$/s or $new_letter) {
786 DEBUG > 2 and print " $new_letter isn't a valid thing to entend.\n";
788 $para->[1]{'start_line'},
789 "You can extend only formatting codes A-Z, not like \"$new_letter\""
794 if(grep $new_letter eq $_, @Known_formatting_codes) {
795 DEBUG > 2 and print " $new_letter isn't a good thing to extend, because known.\n";
797 $para->[1]{'start_line'},
798 "You can't extend an established code like \"$new_letter\""
801 #TODO: or allow if last bit is same?
806 unless($fallbacks_one =~ m/^[A-Z](,[A-Z])*$/s # like "B", "M,I", etc.
807 or $fallbacks_one eq '0' or $fallbacks_one eq '1'
810 $para->[1]{'start_line'},
811 "Format for second =extend parameter must be like"
812 . " M or 1 or 0 or M,N or M,N,O but you have it like "
818 unless($elements_one =~ m/^[^ ,]+(,[^ ,]+)*$/s) { # like "B", "M,I", etc.
820 $para->[1]{'start_line'},
821 "Format for third =extend parameter: like foo or bar,Baz,qu:ux but not like "
827 my @fallbacks = split ',', $fallbacks_one, -1;
828 my @elements = split ',', $elements_one, -1;
830 foreach my $f (@fallbacks) {
831 next if exists $Known_formatting_codes{$f} or $f eq '0' or $f eq '1';
832 DEBUG > 2 and print " Can't fall back on unknown code $f\n";
834 $para->[1]{'start_line'},
835 "Can't use unknown formatting code '$f' as a fallback for '$new_letter'"
840 DEBUG > 3 and printf "Extensor: Fallbacks <%s> Elements <%s>.\n",
841 @fallbacks, @elements;
844 foreach my $e (@elements) {
845 if(exists $self->{'accept_codes'}{$e}) {
846 DEBUG > 1 and print " Mapping '$new_letter' to known extension '$e'\n";
847 $canonical_form = $e;
848 last; # first acceptable elementname wins!
850 DEBUG > 1 and print " Can't map '$new_letter' to unknown extension '$e'\n";
855 if( defined $canonical_form ) {
856 # We found a good N => elementname mapping
857 $self->{'accept_codes'}{$new_letter} = $canonical_form;
859 "Extensor maps $new_letter => known element $canonical_form.\n";
861 # We have to use the fallback(s), which might be '0', or '1'.
862 $self->{'accept_codes'}{$new_letter}
863 = (@fallbacks == 1) ? $fallbacks[0] : \@fallbacks;
865 "Extensor maps $new_letter => fallbacks @fallbacks.\n";
869 DEBUG > 2 and print "Extensor has bad syntax.\n";
871 $para->[1]{'start_line'},
872 "Unknown =extend syntax: $content"
879 #:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.
881 sub _treat_Zs { # Nix Z<...>'s
882 my($self,@stack) = @_;
885 my $start_line = $stack[0][1]{'start_line'};
887 # A recursive algorithm implemented iteratively! Whee!
889 while($treelet = shift @stack) {
890 for($i = 2; $i < @$treelet; ++$i) { # iterate over children
891 next unless ref $treelet->[$i]; # text nodes are uninteresting
892 unless($treelet->[$i][0] eq 'Z') {
893 unshift @stack, $treelet->[$i]; # recurse
897 DEBUG > 1 and print "Nixing Z node @{$treelet->[$i]}\n";
899 # bitch UNLESS it's empty
900 unless( @{$treelet->[$i]} == 2
901 or (@{$treelet->[$i]} == 3 and $treelet->[$i][2] eq '')
903 $self->whine( $start_line, "A non-empty Z<>" );
904 } # but kill it anyway
906 splice(@$treelet, $i, 1); # thereby just nix this node.
915 # . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
917 # Quoting perlpodspec:
919 # In parsing an L<...> code, Pod parsers must distinguish at least four
922 ############# Not used. Expressed via the element children plus
923 ############# the value of the "content-implicit" flag.
925 # The link-text. If there is none, this must be undef. (E.g., in "L<Perl
926 # Functions|perlfunc>", the link-text is "Perl Functions". In
927 # "L<Time::HiRes>" and even "L<|Time::HiRes>", there is no link text. Note
928 # that link text may contain formatting.)
931 ############# The element children
933 # The possibly inferred link-text -- i.e., if there was no real link text,
934 # then this is the text that we'll infer in its place. (E.g., for
935 # "L<Getopt::Std>", the inferred link text is "Getopt::Std".)
938 ############# The "to" attribute (which might be text, or a treelet)
940 # The name or URL, or undef if none. (E.g., in "L<Perl
941 # Functions|perlfunc>", the name -- also sometimes called the page -- is
942 # "perlfunc". In "L</CAVEATS>", the name is undef.)
945 ############# The "section" attribute (which might be next, or a treelet)
947 # The section (AKA "item" in older perlpods), or undef if none. E.g., in
948 # Getopt::Std/DESCRIPTION, "DESCRIPTION" is the section. (Note that this
949 # is not the same as a manpage section like the "5" in "man 5 crontab".
950 # "Section Foo" in the Pod sense means the part of the text that's
951 # introduced by the heading or item whose text is "Foo".)
953 # Pod parsers may also note additional attributes including:
956 ############# The "type" attribute.
958 # A flag for whether item 3 (if present) is a URL (like
959 # "http://lists.perl.org" is), in which case there should be no section
960 # attribute; a Pod name (like "perldoc" and "Getopt::Std" are); or
961 # possibly a man page name (like "crontab(5)" is).
964 ############# Not implemented, I guess.
966 # The raw original L<...> content, before text is split on "|", "/", etc,
967 # and before E<...> codes are expanded.
970 # For L<...> codes without a "name|" part, only E<...> and Z<> codes may
971 # occur -- no other formatting codes. That is, authors should not use
974 # Note, however, that formatting codes and Z<>'s can occur in any and all
975 # parts of an L<...> (i.e., in name, section, text, and url).
977 sub _treat_Ls { # Process our dear dear friends, the L<...> sequences
980 # L<name/"sec"> or L<name/sec>
981 # L</"sec"> or L</sec> or L<"sec">
983 # L<text|name/"sec"> or L<text|name/sec>
984 # L<text|/"sec"> or L<text|/sec> or L<text|"sec">
988 my($self,@stack) = @_;
991 my $start_line = $stack[0][1]{'start_line'};
993 # A recursive algorithm implemented iteratively! Whee!
995 while($treelet = shift @stack) {
996 for(my $i = 2; $i < @$treelet; ++$i) {
997 # iterate over children of current tree node
998 next unless ref $treelet->[$i]; # text nodes are uninteresting
999 unless($treelet->[$i][0] eq 'L') {
1000 unshift @stack, $treelet->[$i]; # recurse
1005 # By here, $treelet->[$i] is definitely an L node
1006 my $ell = $treelet->[$i];
1007 DEBUG > 1 and print "Ogling L node $ell\n";
1009 # bitch if it's empty
1011 or (@{$ell} == 3 and $ell->[2] eq '')
1013 $self->whine( $start_line, "An empty L<>" );
1014 $treelet->[$i] = 'L<>'; # just make it a text node
1020 # there are a number of possible cases:
1021 # 1) text node containing url: http://foo.com
1022 # -> [ 'http://foo.com' ]
1023 # 2) text node containing url and text: foo|http://foo.com
1024 # -> [ 'foo|http://foo.com' ]
1025 # 3) text node containing url start: mailto:xE<at>foo.com
1026 # -> [ 'mailto:x', [ E ... ], 'foo.com' ]
1027 # 4) text node containing url start and text: foo|mailto:xE<at>foo.com
1028 # -> [ 'foo|mailto:x', [ E ... ], 'foo.com' ]
1029 # 5) other nodes containing text and url start: OE<39>Malley|http://foo.com
1030 # -> [ 'O', [ E ... ], 'Malley', '|http://foo.com' ]
1033 # anything before the url is part of the text.
1034 # anything after it is part of the url.
1035 # the url text node itself may contain parts of both.
1037 if (my ($url_index, $text_part, $url_part) =
1038 # grep is no good here; we want to bail out immediately so that we can
1039 # use $1, $2, etc. without having to do the match twice.
1042 next if ref $ell->[$_];
1043 next unless $ell->[$_] =~ m/^(?:([^|]*)\|)?(\w+:[^:\s]\S*)$/s;
1044 return ($_, $1, $2);
1049 $ell->[1]{'type'} = 'url';
1051 my @text = @{$ell}[2..$url_index-1];
1052 push @text, $text_part if defined $text_part;
1054 my @url = @{$ell}[$url_index+1..$#$ell];
1055 unshift @url, $url_part;
1058 $ell->[1]{'content-implicit'} = 'yes';
1062 $ell->[1]{to} = Pod::Simple::LinkSection->new(
1068 splice @$ell, 2, $#$ell, @text;
1073 # Catch some very simple and/or common cases
1074 if(@{$ell} == 3 and ! ref $ell->[2]) {
1076 if($it =~ m/^[-a-zA-Z0-9]+\([-a-zA-Z0-9]+\)$/s) { # man sections
1077 # Hopefully neither too broad nor too restrictive a RE
1078 DEBUG > 1 and print "Catching \"$it\" as manpage link.\n";
1079 $ell->[1]{'type'} = 'man';
1080 # This's the only place where man links can get made.
1081 $ell->[1]{'content-implicit'} = 'yes';
1083 Pod::Simple::LinkSection->new( $it ); # treelet!
1087 if($it =~ m/^[^\/\|,\$\%\@\ \"\<\>\:\#\&\*\{\}\[\]\(\)]+(\:\:[^\/\|,\$\%\@\ \"\<\>\:\#\&\*\{\}\[\]\(\)]+)*$/s) {
1088 # Extremely forgiving idea of what constitutes a bare
1089 # modulename link like L<Foo::Bar> or even L<Thing::1.0::Docs::Tralala>
1090 DEBUG > 1 and print "Catching \"$it\" as ho-hum L<Modulename> link.\n";
1091 $ell->[1]{'type'} = 'pod';
1092 $ell->[1]{'content-implicit'} = 'yes';
1094 Pod::Simple::LinkSection->new( $it ); # treelet!
1102 # ...Uhoh, here's the real L<...> parsing stuff...
1103 # "With the ill behavior, with the ill behavior, with the ill behavior..."
1105 DEBUG > 1 and print "Running a real parse on this non-trivial L\n";
1108 my $link_text; # set to an arrayref if found
1109 my @ell_content = @$ell;
1110 splice @ell_content,0,2; # Knock off the 'L' and {} bits
1112 DEBUG > 3 and print " Ell content to start: ",
1113 pretty(@ell_content), "\n";
1116 # Look for the "|" -- only in CHILDREN (not all underlings!)
1117 # Like L<I like the strictness|strict>
1119 print " Peering at L content for a '|' ...\n";
1120 for(my $j = 0; $j < @ell_content; ++$j) {
1121 next if ref $ell_content[$j];
1123 print " Peering at L-content text bit \"$ell_content[$j]\" for a '|'.\n";
1125 if($ell_content[$j] =~ m/^([^\|]*)\|(.*)$/s) {
1126 my @link_text = ($1); # might be 0-length
1127 $ell_content[$j] = $2; # might be 0-length
1130 print " FOUND a '|' in it. Splitting into [$1] + [$2]\n";
1132 unshift @link_text, splice @ell_content, 0, $j;
1133 # leaving only things at J and after
1134 @ell_content = grep ref($_)||length($_), @ell_content ;
1135 $link_text = [grep ref($_)||length($_), @link_text ];
1136 DEBUG > 3 and printf
1137 " So link text is %s\n and remaining ell content is %s\n",
1138 pretty($link_text), pretty(@ell_content);
1144 # Now look for the "/" -- only in CHILDREN (not all underlings!)
1145 # And afterward, anything left in @ell_content will be the raw name
1146 # Like L<Foo::Bar/Object Methods>
1147 my $section_name; # set to arrayref if found
1148 DEBUG > 3 and print " Peering at L-content for a '/' ...\n";
1149 for(my $j = 0; $j < @ell_content; ++$j) {
1150 next if ref $ell_content[$j];
1152 print " Peering at L-content text bit \"$ell_content[$j]\" for a '/'.\n";
1154 if($ell_content[$j] =~ m/^([^\/]*)\/(.*)$/s) {
1155 my @section_name = ($2); # might be 0-length
1156 $ell_content[$j] = $1; # might be 0-length
1159 print " FOUND a '/' in it.",
1160 " Splitting to page [...$1] + section [$2...]\n";
1162 push @section_name, splice @ell_content, 1+$j;
1163 # leaving only things before and including J
1165 @ell_content = grep ref($_)||length($_), @ell_content ;
1166 @section_name = grep ref($_)||length($_), @section_name ;
1168 # Turn L<.../"foo"> into L<.../foo>
1170 and !ref($section_name[0]) and !ref($section_name[-1])
1171 and $section_name[ 0] =~ m/^\"/s
1172 and $section_name[-1] =~ m/\"$/s
1173 and !( # catch weird degenerate case of L<"> !
1174 @section_name == 1 and $section_name[0] eq '"'
1177 $section_name[ 0] =~ s/^\"//s;
1178 $section_name[-1] =~ s/\"$//s;
1180 print " Quotes removed: ", pretty(@section_name), "\n";
1183 print " No need to remove quotes in ", pretty(@section_name), "\n";
1186 $section_name = \@section_name;
1191 # Turn L<"Foo Bar"> into L</Foo Bar>
1192 if(!$section_name and @ell_content
1193 and !ref($ell_content[0]) and !ref($ell_content[-1])
1194 and $ell_content[ 0] =~ m/^\"/s
1195 and $ell_content[-1] =~ m/\"$/s
1196 and !( # catch weird degenerate case of L<"> !
1197 @ell_content == 1 and $ell_content[0] eq '"'
1200 $section_name = [splice @ell_content];
1201 $section_name->[ 0] =~ s/^\"//s;
1202 $section_name->[-1] =~ s/\"$//s;
1205 # Turn L<Foo Bar> into L</Foo Bar>.
1206 if(!$section_name and !$link_text and @ell_content
1207 and grep !ref($_) && m/ /s, @ell_content
1209 $section_name = [splice @ell_content];
1210 # That's support for the now-deprecated syntax.
1211 # (Maybe generate a warning eventually?)
1212 # Note that it deliberately won't work on L<...|Foo Bar>
1216 # Now make up the link_text
1217 # L<Foo> -> L<Foo|Foo>
1218 # L</Bar> -> L<"Bar"|Bar>
1219 # L<Foo/Bar> -> L<"Bar" in Foo/Foo>
1220 unless($link_text) {
1221 $ell->[1]{'content-implicit'} = 'yes';
1223 push @$link_text, '"', @$section_name, '"' if $section_name;
1226 $link_text->[-1] .= ' in ' if $section_name;
1227 push @$link_text, @ell_content;
1232 # And the E resolver will have to deal with all our treeletty things:
1234 if(@ell_content == 1 and !ref($ell_content[0])
1235 and $ell_content[0] =~ m/^[-a-zA-Z0-9]+\([-a-zA-Z0-9]+\)$/s
1237 $ell->[1]{'type'} = 'man';
1238 DEBUG > 3 and print "Considering this ($ell_content[0]) a man link.\n";
1240 $ell->[1]{'type'} = 'pod';
1241 DEBUG > 3 and print "Considering this a pod link (not man or url).\n";
1244 if( defined $section_name ) {
1245 $ell->[1]{'section'} = Pod::Simple::LinkSection->new(
1246 ['', {}, @$section_name]
1248 DEBUG > 3 and print "L-section content: ", pretty($ell->[1]{'section'}), "\n";
1251 if( @ell_content ) {
1252 $ell->[1]{'to'} = Pod::Simple::LinkSection->new(
1253 ['', {}, @ell_content]
1255 DEBUG > 3 and print "L-to content: ", pretty($ell->[1]{'to'}), "\n";
1258 # And update children to be the link-text:
1259 @$ell = (@$ell[0,1], defined($link_text) ? splice(@$link_text) : '');
1261 DEBUG > 2 and print "End of L-parsing for this node $treelet->[$i]\n";
1263 unshift @stack, $treelet->[$i]; # might as well recurse
1270 # . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1273 my($self,@stack) = @_;
1275 my($i, $treelet, $content, $replacer, $charnum);
1276 my $start_line = $stack[0][1]{'start_line'};
1278 # A recursive algorithm implemented iteratively! Whee!
1281 # Has frightening side effects on L nodes' attributes.
1285 while($treelet = shift @stack) {
1286 for(my $i = 2; $i < @$treelet; ++$i) { # iterate over children
1287 next unless ref $treelet->[$i]; # text nodes are uninteresting
1288 if($treelet->[$i][0] eq 'L') {
1289 # SPECIAL STUFF for semi-processed L<>'s
1292 foreach my $attrname ('section', 'to') {
1293 if(defined($thing = $treelet->[$i][1]{$attrname}) and ref $thing) {
1294 unshift @stack, $thing;
1295 DEBUG > 2 and print " Enqueuing ",
1296 pretty( $treelet->[$i][1]{$attrname} ),
1297 " as an attribute value to tweak.\n";
1301 unshift @stack, $treelet->[$i]; # recurse
1303 } elsif($treelet->[$i][0] ne 'E') {
1304 unshift @stack, $treelet->[$i]; # recurse
1308 DEBUG > 1 and print "Ogling E node ", pretty($treelet->[$i]), "\n";
1310 # bitch if it's empty
1311 if( @{$treelet->[$i]} == 2
1312 or (@{$treelet->[$i]} == 3 and $treelet->[$i][2] eq '')
1314 $self->whine( $start_line, "An empty E<>" );
1315 $treelet->[$i] = 'E<>'; # splice in a literal
1319 # bitch if content is weird
1320 unless(@{$treelet->[$i]} == 3 and !ref($content = $treelet->[$i][2])) {
1321 $self->whine( $start_line, "An E<...> surrounding strange content" );
1322 $replacer = $treelet->[$i]; # scratch
1323 splice(@$treelet, $i, 1, # fake out a literal
1325 splice(@$replacer,2), # promote its content
1328 # Don't need to do --$i, as the 'E<' we just added isn't interesting.
1332 DEBUG > 1 and print "Ogling E<$content>\n";
1334 $charnum = Pod::Escapes::e2charnum($content);
1335 DEBUG > 1 and print " Considering E<$content> with char ",
1336 defined($charnum) ? $charnum : "undef", ".\n";
1338 if(!defined( $charnum )) {
1339 DEBUG > 1 and print "I don't know how to deal with E<$content>.\n";
1340 $self->whine( $start_line, "Unknown E content in E<$content>" );
1341 $replacer = "E<$content>"; # better than nothing
1342 } elsif($charnum >= 255 and !UNICODE) {
1343 $replacer = ASCII ? "\xA4" : "?";
1344 DEBUG > 1 and print "This Perl version can't handle ",
1345 "E<$content> (chr $charnum), so replacing with $replacer\n";
1347 $replacer = Pod::Escapes::e2char($content);
1348 DEBUG > 1 and print " Replacing E<$content> with $replacer\n";
1351 splice(@$treelet, $i, 1, $replacer); # no need to back up $i, tho
1359 # . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1362 my($self,$treelet) = @_;
1364 _change_S_to_nbsp($treelet,0) if $self->{'nbsp_for_S'};
1366 # TODO: or a change_nbsp_to_S
1367 # Normalizing nbsp's to S is harder: for each text node, make S content
1368 # out of anything matching m/([^ \xA0]*(?:\xA0+[^ \xA0]*)+)/
1375 sub _change_S_to_nbsp { # a recursive function
1376 # Sanely assumes that the top node in the excursion won't be an S node.
1377 my($treelet, $in_s) = @_;
1379 my $is_s = ('S' eq $treelet->[0]);
1380 $in_s ||= $is_s; # So in_s is on either by this being an S element,
1381 # or by an ancestor being an S element.
1383 for(my $i = 2; $i < @$treelet; ++$i) {
1384 if(ref $treelet->[$i]) {
1385 if( _change_S_to_nbsp( $treelet->[$i], $in_s ) ) {
1386 my $to_pull_up = $treelet->[$i];
1387 splice @$to_pull_up,0,2; # ...leaving just its content
1388 splice @$treelet, $i, 1, @$to_pull_up; # Pull up content
1389 $i += @$to_pull_up - 1; # Make $i skip the pulled-up stuff
1392 $treelet->[$i] =~ s/\s/\xA0/g if ASCII and $in_s;
1393 # (If not in ASCIIland, we can't assume that \xA0 == nbsp.)
1395 # Note that if you apply nbsp_for_S to text, and so turn
1396 # "foo S<bar baz> quux" into "foo bar faz quux", you
1397 # end up with something that fails to say "and don't hyphenate
1398 # any part of 'bar baz'". However, hyphenation is such a vexing
1399 # problem anyway, that most Pod renderers just don't render it
1400 # at all. But if you do want to implement hyphenation, I guess
1401 # that you'd better have nbsp_for_S off.
1408 #-----------------------------------------------------------------------------
1410 sub _accessorize { # A simple-minded method-maker
1412 foreach my $attrname (@_) {
1413 next if $attrname =~ m/::/; # a hack
1414 *{caller() . '::' . $attrname} = sub {
1416 $Carp::CarpLevel = 1, Carp::croak(
1417 "Accessor usage: \$obj->$attrname() or \$obj->$attrname(\$new_value)"
1418 ) unless (@_ == 1 or @_ == 2) and ref $_[0];
1419 (@_ == 1) ? $_[0]->{$attrname}
1420 : ($_[0]->{$attrname} = $_[1]);
1423 # Ya know, they say accessories make the ensemble!
1427 # . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1428 # . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1429 #=============================================================================
1432 my($class, $source) = @_;
1433 my $new = $class->new;
1434 $new->output_fh(*STDOUT{IO});
1436 if(ref($source || '') eq 'SCALAR') {
1437 $new->parse_string_document( $$source );
1438 } elsif(ref($source)) { # it's a file handle
1439 $new->parse_file($source);
1440 } else { # it's a filename
1441 $new->parse_file($source);
1448 #-----------------------------------------------------------------------------
1451 # For use in testing: Class->_out($source)
1452 # returns the transformation of $source
1454 my $class = shift(@_);
1456 my $mutor = shift(@_) if @_ and ref($_[0] || '') eq 'CODE';
1458 DEBUG and print "\n\n", '#' x 76,
1459 "\nAbout to parse source: {{\n$_[0]\n}}\n\n";
1462 my $parser = ref $class && $class->isa(__PACKAGE__) ? $class : $class->new;
1463 $parser->hide_line_numbers(1);
1466 $parser->output_string( \$out );
1467 DEBUG and print " _out to ", \$out, "\n";
1469 $mutor->($parser) if $mutor;
1471 $parser->parse_string_document( $_[0] );
1472 # use Data::Dumper; print Dumper($parser), "\n";
1478 # For use in testing: Class->_duo($source1, $source2)
1479 # returns the parse trees of $source1 and $source2.
1480 # Good in things like: &ok( Class->duo(... , ...) );
1482 my $class = shift(@_);
1484 Carp::croak "But $class->_duo is useful only in list context!"
1487 my $mutor = shift(@_) if @_ and ref($_[0] || '') eq 'CODE';
1489 Carp::croak "But $class->_duo takes two parameters, not: @_"
1495 my $parser = $class->new;
1498 $parser->output_string( \( $out[-1] ) );
1500 DEBUG and print " _duo out to ", $parser->output_string(),
1501 " = $parser->{'output_string'}\n";
1503 $parser->hide_line_numbers(1);
1504 $mutor->($parser) if $mutor;
1505 $parser->parse_string_document( shift( @_ ) );
1506 # use Data::Dumper; print Dumper($parser), "\n";
1514 #-----------------------------------------------------------------------------
1519 A start_formatting_code and end_formatting_code methods, which in the
1520 base class call start_L, end_L, start_C, end_C, etc., if they are
1523 have the POD FORMATTING ERRORS section note the localtime, and the
1524 version of Pod::Simple.
1526 option to delete all E<shy>s?
1527 option to scream if under-0x20 literals are found in the input, or
1528 under-E<32> E codes are found in the tree. And ditto \x7f-\x9f
1530 Option to turn highbit characters into their compromised form? (applies
1533 TODO: BOM/encoding things.
1535 TODO: ascii-compat things in the XML classes?