Commit | Line | Data |
c99ca59a |
1 | package B::Concise; |
62e36f8a |
2 | # Copyright (C) 2000-2002 Stephen McCamant. All rights reserved. |
c99ca59a |
3 | # This program is free software; you can redistribute and/or modify it |
4 | # under the same terms as Perl itself. |
5 | |
c99ca59a |
6 | use strict; |
78ad9108 |
7 | use warnings; |
8 | |
9 | use Exporter (); |
10 | |
11 | our $VERSION = "0.52"; |
12 | our @ISA = qw(Exporter); |
13 | our @EXPORT_OK = qw(set_style add_callback); |
14 | |
c99ca59a |
15 | use B qw(class ppname main_start main_root main_cv cstring svref_2object |
16 | SVf_IOK SVf_NOK SVf_POK OPf_KIDS); |
17 | |
18 | my %style = |
19 | ("terse" => |
c3caa09d |
20 | ["(?(#label =>\n)?)(*( )*)#class (#addr) #name (?([#targ])?) " |
21 | . "#svclass~(?((#svaddr))?)~#svval~(?(label \"#coplabel\")?)\n", |
c99ca59a |
22 | "(*( )*)goto #class (#addr)\n", |
23 | "#class pp_#name"], |
24 | "concise" => |
25 | ["#hyphseq2 (*( (x( ;)x))*)<#classsym> " |
26 | . "#exname#arg(?([#targarglife])?)~#flags(?(/#private)?)(x(;~->#next)x)\n", |
27 | " (*( )*) goto #seq\n", |
28 | "(?(<#seq>)?)#exname#arg(?([#targarglife])?)"], |
29 | "linenoise" => |
30 | ["(x(;(*( )*))x)#noise#arg(?([#targarg])?)(x( ;\n)x)", |
31 | "gt_#seq ", |
32 | "(?(#seq)?)#noise#arg(?([#targarg])?)"], |
33 | "debug" => |
34 | ["#class (#addr)\n\top_next\t\t#nextaddr\n\top_sibling\t#sibaddr\n\t" |
35 | . "op_ppaddr\tPL_ppaddr[OP_#NAME]\n\top_type\t\t#typenum\n\top_seq\t\t" |
36 | . "#seqnum\n\top_flags\t#flagval\n\top_private\t#privval\n" |
37 | . "(?(\top_first\t#firstaddr\n)?)(?(\top_last\t\t#lastaddr\n)?)" |
38 | . "(?(\top_sv\t\t#svaddr\n)?)", |
39 | " GOTO #addr\n", |
40 | "#addr"], |
41 | "env" => [$ENV{B_CONCISE_FORMAT}, $ENV{B_CONCISE_GOTO_FORMAT}, |
42 | $ENV{B_CONCISE_TREE_FORMAT}], |
43 | ); |
44 | |
45 | my($format, $gotofmt, $treefmt); |
46 | my $curcv; |
47 | my($seq_base, $cop_seq_base); |
78ad9108 |
48 | my @callbacks; |
49 | |
50 | sub set_style { |
51 | ($format, $gotofmt, $treefmt) = @_; |
52 | } |
53 | |
54 | sub add_callback { |
55 | push @callbacks, @_; |
56 | } |
c99ca59a |
57 | |
58 | sub concise_cv { |
59 | my ($order, $cvref) = @_; |
60 | my $cv = svref_2object($cvref); |
61 | $curcv = $cv; |
62 | if ($order eq "exec") { |
63 | walk_exec($cv->START); |
64 | } elsif ($order eq "basic") { |
65 | walk_topdown($cv->ROOT, sub { $_[0]->concise($_[1]) }, 0); |
66 | } else { |
67 | print tree($cv->ROOT, 0) |
68 | } |
69 | } |
70 | |
71 | my $start_sym = "\e(0"; # "\cN" sometimes also works |
72 | my $end_sym = "\e(B"; # "\cO" respectively |
73 | |
74 | my @tree_decorations = |
75 | ([" ", "--", "+-", "|-", "| ", "`-", "-", 1], |
76 | [" ", "-", "+", "+", "|", "`", "", 0], |
77 | [" ", map("$start_sym$_$end_sym", "qq", "wq", "tq", "x ", "mq", "q"), 1], |
78 | [" ", map("$start_sym$_$end_sym", "q", "w", "t", "x", "m"), "", 0], |
79 | ); |
80 | my $tree_style = 0; |
81 | |
82 | my $base = 36; |
83 | my $big_endian = 1; |
84 | |
85 | my $order = "basic"; |
86 | |
78ad9108 |
87 | set_style(@{$style{concise}}); |
88 | |
c99ca59a |
89 | sub compile { |
90 | my @options = grep(/^-/, @_); |
91 | my @args = grep(!/^-/, @_); |
92 | my $do_main = 0; |
c99ca59a |
93 | for my $o (@options) { |
94 | if ($o eq "-basic") { |
95 | $order = "basic"; |
96 | } elsif ($o eq "-exec") { |
97 | $order = "exec"; |
98 | } elsif ($o eq "-tree") { |
99 | $order = "tree"; |
100 | } elsif ($o eq "-compact") { |
101 | $tree_style |= 1; |
102 | } elsif ($o eq "-loose") { |
103 | $tree_style &= ~1; |
104 | } elsif ($o eq "-vt") { |
105 | $tree_style |= 2; |
106 | } elsif ($o eq "-ascii") { |
107 | $tree_style &= ~2; |
108 | } elsif ($o eq "-main") { |
109 | $do_main = 1; |
110 | } elsif ($o =~ /^-base(\d+)$/) { |
111 | $base = $1; |
112 | } elsif ($o eq "-bigendian") { |
113 | $big_endian = 1; |
114 | } elsif ($o eq "-littleendian") { |
115 | $big_endian = 0; |
116 | } elsif (exists $style{substr($o, 1)}) { |
78ad9108 |
117 | set_style(@{$style{substr($o, 1)}}); |
c99ca59a |
118 | } else { |
119 | warn "Option $o unrecognized"; |
120 | } |
121 | } |
122 | if (@args) { |
123 | return sub { |
124 | for my $objname (@args) { |
125 | $objname = "main::" . $objname unless $objname =~ /::/; |
126 | eval "concise_cv(\$order, \\&$objname)"; |
127 | die "concise_cv($order, \\&$objname) failed: $@" if $@; |
128 | } |
129 | } |
130 | } |
131 | if (!@args or $do_main) { |
132 | if ($order eq "exec") { |
133 | return sub { return if class(main_start) eq "NULL"; |
134 | $curcv = main_cv; |
135 | walk_exec(main_start) } |
136 | } elsif ($order eq "tree") { |
137 | return sub { return if class(main_root) eq "NULL"; |
138 | $curcv = main_cv; |
139 | print tree(main_root, 0) } |
140 | } elsif ($order eq "basic") { |
141 | return sub { return if class(main_root) eq "NULL"; |
142 | $curcv = main_cv; |
143 | walk_topdown(main_root, |
144 | sub { $_[0]->concise($_[1]) }, 0); } |
145 | } |
146 | } |
147 | } |
148 | |
149 | my %labels; |
150 | my $lastnext; |
151 | |
152 | my %opclass = ('OP' => "0", 'UNOP' => "1", 'BINOP' => "2", 'LOGOP' => "|", |
153 | 'LISTOP' => "@", 'PMOP' => "/", 'SVOP' => "\$", 'GVOP' => "*", |
051f02e9 |
154 | 'PVOP' => '"', 'LOOP' => "{", 'COP' => ";", 'PADOP' => "#"); |
c99ca59a |
155 | |
35fc55f1 |
156 | no warnings 'qw'; # "Possible attempt to put comments..." |
157 | my @linenoise = |
158 | qw'# () sc ( @? 1 $* gv *{ m$ m@ m% m? p/ *$ $ $# & a& pt \\ s\\ rf bl |
c99ca59a |
159 | ` *? <> ?? ?/ r/ c/ // qr s/ /c y/ = @= C sC Cp sp df un BM po +1 +I |
160 | -1 -I 1+ I+ 1- I- ** * i* / i/ %$ i% x + i+ - i- . " << >> < i< |
161 | > i> <= i, >= i. == i= != i! <? i? s< s> s, s. s= s! s? b& b^ b| -0 -i |
162 | ! ~ a2 si cs rd sr e^ lg sq in %x %o ab le ss ve ix ri sf FL od ch cy |
163 | uf lf uc lc qm @ [f [ @[ eh vl ky dl ex % ${ @{ uk pk st jn ) )[ a@ |
164 | a% sl +] -] [- [+ so rv GS GW MS MW .. f. .f && || ^^ ?: &= |= -> s{ s} |
165 | v} ca wa di rs ;; ; ;d }{ { } {} f{ it {l l} rt }l }n }r dm }g }e ^o |
166 | ^c ^| ^# um bm t~ u~ ~d DB db ^s se ^g ^r {w }w pf pr ^O ^K ^R ^W ^d ^v |
167 | ^e ^t ^k t. fc ic fl .s .p .b .c .l .a .h g1 s1 g2 s2 ?. l? -R -W -X -r |
168 | -w -x -e -o -O -z -s -M -A -C -S -c -b -f -d -p -l -u -g -k -t -T -B cd |
169 | co cr u. cm ut r. l@ s@ r@ mD uD oD rD tD sD wD cD f$ w$ p$ sh e$ k$ g3 |
170 | g4 s4 g5 s5 T@ C@ L@ G@ A@ S@ Hg Hc Hr Hw Mg Mc Ms Mr Sg Sc So rq do {e |
171 | e} {t t} g6 G6 6e g7 G7 7e g8 G8 8e g9 G9 9e 6s 7s 8s 9s 6E 7E 8E 9E Pn |
35fc55f1 |
172 | Pu GP SP EP Gn Gg GG SG EG g0 c$ lk t$ ;s n>'; |
c99ca59a |
173 | |
174 | my $chars = "0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ"; |
175 | |
176 | sub op_flags { |
177 | my($x) = @_; |
178 | my(@v); |
179 | push @v, "v" if ($x & 3) == 1; |
180 | push @v, "s" if ($x & 3) == 2; |
181 | push @v, "l" if ($x & 3) == 3; |
182 | push @v, "K" if $x & 4; |
183 | push @v, "P" if $x & 8; |
184 | push @v, "R" if $x & 16; |
185 | push @v, "M" if $x & 32; |
186 | push @v, "S" if $x & 64; |
187 | push @v, "*" if $x & 128; |
188 | return join("", @v); |
189 | } |
190 | |
191 | sub base_n { |
192 | my $x = shift; |
193 | return "-" . base_n(-$x) if $x < 0; |
194 | my $str = ""; |
195 | do { $str .= substr($chars, $x % $base, 1) } while $x = int($x / $base); |
196 | $str = reverse $str if $big_endian; |
197 | return $str; |
198 | } |
199 | |
200 | sub seq { return $_[0]->seq ? base_n($_[0]->seq - $seq_base) : "-" } |
201 | |
202 | sub walk_topdown { |
203 | my($op, $sub, $level) = @_; |
204 | $sub->($op, $level); |
205 | if ($op->flags & OPf_KIDS) { |
206 | for (my $kid = $op->first; $$kid; $kid = $kid->sibling) { |
207 | walk_topdown($kid, $sub, $level + 1); |
208 | } |
209 | } |
b2a3cfdd |
210 | if (class($op) eq "PMOP" and $ {$op->pmreplroot} |
211 | and $op->pmreplroot->isa("B::OP")) { |
c99ca59a |
212 | walk_topdown($op->pmreplroot, $sub, $level + 1); |
213 | } |
214 | } |
215 | |
216 | sub walklines { |
217 | my($ar, $level) = @_; |
218 | for my $l (@$ar) { |
219 | if (ref($l) eq "ARRAY") { |
220 | walklines($l, $level + 1); |
221 | } else { |
222 | $l->concise($level); |
223 | } |
224 | } |
225 | } |
226 | |
227 | sub walk_exec { |
228 | my($top, $level) = @_; |
229 | my %opsseen; |
230 | my @lines; |
231 | my @todo = ([$top, \@lines]); |
232 | while (@todo and my($op, $targ) = @{shift @todo}) { |
233 | for (; $$op; $op = $op->next) { |
234 | last if $opsseen{$$op}++; |
235 | push @$targ, $op; |
236 | my $name = $op->name; |
62e36f8a |
237 | if (class($op) eq "LOGOP") { |
c99ca59a |
238 | my $ar = []; |
239 | push @$targ, $ar; |
240 | push @todo, [$op->other, $ar]; |
241 | } elsif ($name eq "subst" and $ {$op->pmreplstart}) { |
242 | my $ar = []; |
243 | push @$targ, $ar; |
244 | push @todo, [$op->pmreplstart, $ar]; |
245 | } elsif ($name =~ /^enter(loop|iter)$/) { |
246 | $labels{$op->nextop->seq} = "NEXT"; |
247 | $labels{$op->lastop->seq} = "LAST"; |
248 | $labels{$op->redoop->seq} = "REDO"; |
249 | } |
250 | } |
251 | } |
252 | walklines(\@lines, 0); |
253 | } |
254 | |
255 | sub fmt_line { |
256 | my($hr, $fmt, $level) = @_; |
257 | my $text = $fmt; |
258 | $text =~ s/\(\?\(([^\#]*?)\#(\w+)([^\#]*?)\)\?\)/ |
259 | $hr->{$2} ? $1.$hr->{$2}.$3 : ""/eg; |
260 | $text =~ s/\(x\((.*?);(.*?)\)x\)/$order eq "exec" ? $1 : $2/egs; |
261 | $text =~ s/\(\*\(([^;]*?)\)\*\)/$1 x $level/egs; |
262 | $text =~ s/\(\*\((.*?);(.*?)\)\*\)/$1 x ($level - 1) . $2 x ($level>0)/egs; |
263 | $text =~ s/#([a-zA-Z]+)(\d+)/sprintf("%-$2s", $hr->{$1})/eg; |
264 | $text =~ s/#([a-zA-Z]+)/$hr->{$1}/eg; |
265 | $text =~ s/[ \t]*~+[ \t]*/ /g; |
266 | return $text; |
267 | } |
268 | |
269 | my %priv; |
270 | $priv{$_}{128} = "LVINTRO" |
271 | for ("pos", "substr", "vec", "threadsv", "gvsv", "rv2sv", "rv2hv", "rv2gv", |
272 | "rv2av", "rv2arylen", "aelem", "helem", "aslice", "hslice", "padsv", |
273 | "padav", "padhv"); |
274 | $priv{$_}{64} = "REFC" for ("leave", "leavesub", "leavesublv", "leavewrite"); |
275 | $priv{"aassign"}{64} = "COMMON"; |
c99ca59a |
276 | $priv{"sassign"}{64} = "BKWARD"; |
277 | $priv{$_}{64} = "RTIME" for ("match", "subst", "substcont"); |
278 | @{$priv{"trans"}}{1,2,4,8,16,64} = ("<UTF", ">UTF", "IDENT", "SQUASH", "DEL", |
279 | "COMPL", "GROWS"); |
280 | $priv{"repeat"}{64} = "DOLIST"; |
281 | $priv{"leaveloop"}{64} = "CONT"; |
282 | @{$priv{$_}}{32,64,96} = ("DREFAV", "DREFHV", "DREFSV") |
283 | for ("entersub", map("rv2${_}v", "a", "s", "h", "g"), "aelem", "helem"); |
284 | $priv{"entersub"}{16} = "DBG"; |
285 | $priv{"entersub"}{32} = "TARG"; |
286 | @{$priv{$_}}{4,8,128} = ("INARGS","AMPER","NO()") for ("entersub", "rv2cv"); |
287 | $priv{"gv"}{32} = "EARLYCV"; |
288 | $priv{"aelem"}{16} = $priv{"helem"}{16} = "LVDEFER"; |
289 | $priv{$_}{16} = "OURINTR" for ("gvsv", "rv2sv", "rv2av", "rv2hv", "r2gv"); |
290 | $priv{$_}{16} = "TARGMY" |
291 | for (map(($_,"s$_"),"chop", "chomp"), |
292 | map(($_,"i_$_"), "postinc", "postdec", "multiply", "divide", "modulo", |
293 | "add", "subtract", "negate"), "pow", "concat", "stringify", |
294 | "left_shift", "right_shift", "bit_and", "bit_xor", "bit_or", |
295 | "complement", "atan2", "sin", "cos", "rand", "exp", "log", "sqrt", |
296 | "int", "hex", "oct", "abs", "length", "index", "rindex", "sprintf", |
297 | "ord", "chr", "crypt", "quotemeta", "join", "push", "unshift", "flock", |
298 | "chdir", "chown", "chroot", "unlink", "chmod", "utime", "rename", |
299 | "link", "symlink", "mkdir", "rmdir", "wait", "waitpid", "system", |
300 | "exec", "kill", "getppid", "getpgrp", "setpgrp", "getpriority", |
301 | "setpriority", "time", "sleep"); |
7a9b44b9 |
302 | @{$priv{"const"}}{8,16,32,64,128} = ("STRICT","ENTERED", '$[', "BARE", "WARN"); |
c99ca59a |
303 | $priv{"flip"}{64} = $priv{"flop"}{64} = "LINENUM"; |
304 | $priv{"list"}{64} = "GUESSED"; |
305 | $priv{"delete"}{64} = "SLICE"; |
306 | $priv{"exists"}{64} = "SUB"; |
307 | $priv{$_}{64} = "LOCALE" |
308 | for ("sort", "prtf", "sprintf", "slt", "sle", "seq", "sne", "sgt", "sge", |
309 | "scmp", "lc", "uc", "lcfirst", "ucfirst"); |
310 | @{$priv{"sort"}}{1,2,4} = ("NUM", "INT", "REV"); |
311 | $priv{"threadsv"}{64} = "SVREFd"; |
312 | $priv{$_}{16} = "INBIN" for ("open", "backtick"); |
313 | $priv{$_}{32} = "INCR" for ("open", "backtick"); |
314 | $priv{$_}{64} = "OUTBIN" for ("open", "backtick"); |
315 | $priv{$_}{128} = "OUTCR" for ("open", "backtick"); |
316 | $priv{"exit"}{128} = "VMS"; |
317 | |
318 | sub private_flags { |
319 | my($name, $x) = @_; |
320 | my @s; |
321 | for my $flag (128, 96, 64, 32, 16, 8, 4, 2, 1) { |
322 | if ($priv{$name}{$flag} and $x & $flag and $x >= $flag) { |
323 | $x -= $flag; |
324 | push @s, $priv{$name}{$flag}; |
325 | } |
326 | } |
327 | push @s, $x if $x; |
328 | return join(",", @s); |
329 | } |
330 | |
331 | sub concise_op { |
332 | my ($op, $level, $format) = @_; |
333 | my %h; |
334 | $h{exname} = $h{name} = $op->name; |
335 | $h{NAME} = uc $h{name}; |
336 | $h{class} = class($op); |
337 | $h{extarg} = $h{targ} = $op->targ; |
338 | $h{extarg} = "" unless $h{extarg}; |
339 | if ($h{name} eq "null" and $h{targ}) { |
340 | $h{exname} = "ex-" . substr(ppname($h{targ}), 3); |
341 | $h{extarg} = ""; |
342 | } elsif ($h{targ}) { |
343 | my $padname = (($curcv->PADLIST->ARRAY)[0]->ARRAY)[$h{targ}]; |
344 | if (defined $padname and class($padname) ne "SPECIAL") { |
0b40bd6d |
345 | $h{targarg} = $padname->PVX; |
c99ca59a |
346 | my $intro = $padname->NVX - $cop_seq_base; |
347 | my $finish = int($padname->IVX) - $cop_seq_base; |
348 | $finish = "end" if $finish == 999999999 - $cop_seq_base; |
349 | $h{targarglife} = "$h{targarg}:$intro,$finish"; |
350 | } else { |
351 | $h{targarglife} = $h{targarg} = "t" . $h{targ}; |
352 | } |
353 | } |
354 | $h{arg} = ""; |
355 | $h{svclass} = $h{svaddr} = $h{svval} = ""; |
356 | if ($h{class} eq "PMOP") { |
357 | my $precomp = $op->precomp; |
7a9b44b9 |
358 | if (defined $precomp) { |
359 | # Escape literal control sequences |
360 | for ($precomp) { |
361 | s/\t/\\t/g; s/\n/\\n/g; s/\r/\\r/g; |
362 | # How can we do the below portably? |
363 | #s/([\0-\037\177-\377])/"\\".sprintf("%03o", ord($1))/eg; |
364 | } |
365 | $precomp = "/$precomp/"; |
366 | } |
367 | else { $precomp = ""; } |
b2a3cfdd |
368 | my $pmreplroot = $op->pmreplroot; |
34a48b4b |
369 | my $pmreplstart; |
370 | if ($$pmreplroot && $pmreplroot->isa("B::GV")) { |
b2a3cfdd |
371 | # with C<@stash_array = split(/pat/, str);>, |
372 | # *stash_array is stored in pmreplroot. |
373 | $h{arg} = "($precomp => \@" . $pmreplroot->NAME . ")"; |
374 | } elsif ($ {$op->pmreplstart}) { |
c99ca59a |
375 | undef $lastnext; |
376 | $pmreplstart = "replstart->" . seq($op->pmreplstart); |
377 | $h{arg} = "(" . join(" ", $precomp, $pmreplstart) . ")"; |
378 | } else { |
379 | $h{arg} = "($precomp)"; |
380 | } |
381 | } elsif ($h{class} eq "PVOP" and $h{name} ne "trans") { |
382 | $h{arg} = '("' . $op->pv . '")'; |
383 | $h{svval} = '"' . $op->pv . '"'; |
384 | } elsif ($h{class} eq "COP") { |
385 | my $label = $op->label; |
c3caa09d |
386 | $h{coplabel} = $label; |
c99ca59a |
387 | $label = $label ? "$label: " : ""; |
388 | my $loc = $op->file; |
389 | $loc =~ s[.*/][]; |
390 | $loc .= ":" . $op->line; |
391 | my($stash, $cseq) = ($op->stash->NAME, $op->cop_seq - $cop_seq_base); |
392 | my $arybase = $op->arybase; |
393 | $arybase = $arybase ? ' $[=' . $arybase : ""; |
394 | $h{arg} = "($label$stash $cseq $loc$arybase)"; |
395 | } elsif ($h{class} eq "LOOP") { |
396 | $h{arg} = "(next->" . seq($op->nextop) . " last->" . seq($op->lastop) |
397 | . " redo->" . seq($op->redoop) . ")"; |
398 | } elsif ($h{class} eq "LOGOP") { |
399 | undef $lastnext; |
400 | $h{arg} = "(other->" . seq($op->other) . ")"; |
401 | } elsif ($h{class} eq "SVOP") { |
402 | my $sv = $op->sv; |
403 | $h{svclass} = class($sv); |
404 | $h{svaddr} = sprintf("%#x", $$sv); |
405 | if ($h{svclass} eq "GV") { |
406 | my $gv = $sv; |
407 | my $stash = $gv->STASH->NAME; |
408 | if ($stash eq "main") { |
409 | $stash = ""; |
410 | } else { |
411 | $stash = $stash . "::"; |
412 | } |
002b978b |
413 | $h{arg} = "(*$stash" . $gv->SAFENAME . ")"; |
414 | $h{svval} = "*$stash" . $gv->SAFENAME; |
c99ca59a |
415 | } else { |
416 | while (class($sv) eq "RV") { |
417 | $h{svval} .= "\\"; |
418 | $sv = $sv->RV; |
419 | } |
420 | if (class($sv) eq "SPECIAL") { |
421 | $h{svval} = ["Null", "sv_undef", "sv_yes", "sv_no"]->[$$sv]; |
422 | } elsif ($sv->FLAGS & SVf_NOK) { |
423 | $h{svval} = $sv->NV; |
424 | } elsif ($sv->FLAGS & SVf_IOK) { |
425 | $h{svval} = $sv->IV; |
426 | } elsif ($sv->FLAGS & SVf_POK) { |
427 | $h{svval} = cstring($sv->PV); |
428 | } |
429 | $h{arg} = "($h{svclass} $h{svval})"; |
430 | } |
431 | } |
432 | $h{seq} = $h{hyphseq} = seq($op); |
433 | $h{seq} = "" if $h{seq} eq "-"; |
434 | $h{seqnum} = $op->seq; |
435 | $h{next} = $op->next; |
436 | $h{next} = (class($h{next}) eq "NULL") ? "(end)" : seq($h{next}); |
437 | $h{nextaddr} = sprintf("%#x", $ {$op->next}); |
438 | $h{sibaddr} = sprintf("%#x", $ {$op->sibling}); |
439 | $h{firstaddr} = sprintf("%#x", $ {$op->first}) if $op->can("first"); |
440 | $h{lastaddr} = sprintf("%#x", $ {$op->last}) if $op->can("last"); |
441 | |
442 | $h{classsym} = $opclass{$h{class}}; |
443 | $h{flagval} = $op->flags; |
444 | $h{flags} = op_flags($op->flags); |
445 | $h{privval} = $op->private; |
446 | $h{private} = private_flags($h{name}, $op->private); |
447 | $h{addr} = sprintf("%#x", $$op); |
448 | $h{label} = $labels{$op->seq}; |
449 | $h{typenum} = $op->type; |
450 | $h{noise} = $linenoise[$op->type]; |
78ad9108 |
451 | $_->(\%h, $op, \$format, \$level) for @callbacks; |
c99ca59a |
452 | return fmt_line(\%h, $format, $level); |
453 | } |
454 | |
455 | sub B::OP::concise { |
456 | my($op, $level) = @_; |
457 | if ($order eq "exec" and $lastnext and $$lastnext != $$op) { |
458 | my $h = {"seq" => seq($lastnext), "class" => class($lastnext), |
459 | "addr" => sprintf("%#x", $$lastnext)}; |
460 | print fmt_line($h, $gotofmt, $level+1); |
461 | } |
462 | $lastnext = $op->next; |
463 | print concise_op($op, $level, $format); |
464 | } |
465 | |
466 | sub tree { |
467 | my $op = shift; |
468 | my $level = shift; |
469 | my $style = $tree_decorations[$tree_style]; |
470 | my($space, $single, $kids, $kid, $nokid, $last, $lead, $size) = @$style; |
471 | my $name = concise_op($op, $level, $treefmt); |
472 | if (not $op->flags & OPf_KIDS) { |
473 | return $name . "\n"; |
474 | } |
475 | my @lines; |
476 | for (my $kid = $op->first; $$kid; $kid = $kid->sibling) { |
477 | push @lines, tree($kid, $level+1); |
478 | } |
479 | my $i; |
480 | for ($i = $#lines; substr($lines[$i], 0, 1) eq " "; $i--) { |
481 | $lines[$i] = $space . $lines[$i]; |
482 | } |
483 | if ($i > 0) { |
484 | $lines[$i] = $last . $lines[$i]; |
485 | while ($i-- > 1) { |
486 | if (substr($lines[$i], 0, 1) eq " ") { |
487 | $lines[$i] = $nokid . $lines[$i]; |
488 | } else { |
489 | $lines[$i] = $kid . $lines[$i]; |
490 | } |
491 | } |
492 | $lines[$i] = $kids . $lines[$i]; |
493 | } else { |
494 | $lines[0] = $single . $lines[0]; |
495 | } |
496 | return("$name$lead" . shift @lines, |
497 | map(" " x (length($name)+$size) . $_, @lines)); |
498 | } |
499 | |
213a1a26 |
500 | # *** Warning: fragile kludge ahead *** |
501 | # Because the B::* modules run in the same interpreter as the code |
502 | # they're compiling, their presence tends to distort the view we have |
503 | # of the code we're looking at. In particular, perl gives sequence |
504 | # numbers to both OPs in general and COPs in particular. If the |
505 | # program we're looking at were run on its own, these numbers would |
506 | # start at 1. Because all of B::Concise and all the modules it uses |
507 | # are compiled first, though, by the time we get to the user's program |
508 | # the sequence numbers are alreay at pretty high numbers, which would |
509 | # be distracting if you're trying to tell OPs apart. Therefore we'd |
510 | # like to subtract an offset from all the sequence numbers we display, |
511 | # to restore the simpler view of the world. The trick is to know what |
512 | # that offset will be, when we're still compiling B::Concise! If we |
513 | # hardcoded a value, it would have to change every time B::Concise or |
514 | # other modules we use do. To help a little, what we do here is |
515 | # compile a little code at the end of the module, and compute the base |
516 | # sequence number for the user's program as being a small offset |
517 | # later, so all we have to worry about are changes in the offset. |
518 | |
519 | # When you say "perl -MO=Concise -e '$a'", the output should look like: |
520 | |
521 | # 4 <@> leave[t1] vKP/REFC ->(end) |
522 | # 1 <0> enter ->2 |
523 | #^ smallest OP sequence number should be 1 |
524 | # 2 <;> nextstate(main 1 -e:1) v ->3 |
525 | # ^ smallest COP sequence number should be 1 |
526 | # - <1> ex-rv2sv vK/1 ->4 |
527 | # 3 <$> gvsv(*a) s ->4 |
528 | |
529 | # If either of the marked numbers there aren't 1, it means you need to |
530 | # update the corresponding magic number in the next two lines. |
e69a2255 |
531 | # Remember, these need to stay the last things in the module. |
532 | |
533 | # Why these are different for MacOS? Does it matter? |
534 | my $cop_seq_mnum = $^O eq 'MacOS' ? 12 : 11; |
e9f19e3c |
535 | my $seq_mnum = $^O eq 'MacOS' ? 102 : 86; |
e69a2255 |
536 | $cop_seq_base = svref_2object(eval 'sub{0;}')->START->cop_seq + $cop_seq_mnum; |
537 | $seq_base = svref_2object(eval 'sub{}')->START->seq + $seq_mnum; |
c99ca59a |
538 | |
539 | 1; |
540 | |
541 | __END__ |
542 | |
543 | =head1 NAME |
544 | |
545 | B::Concise - Walk Perl syntax tree, printing concise info about ops |
546 | |
547 | =head1 SYNOPSIS |
548 | |
549 | perl -MO=Concise[,OPTIONS] foo.pl |
550 | |
78ad9108 |
551 | use B::Concise qw(set_style add_callback); |
552 | |
c99ca59a |
553 | =head1 DESCRIPTION |
554 | |
555 | This compiler backend prints the internal OPs of a Perl program's syntax |
556 | tree in one of several space-efficient text formats suitable for debugging |
557 | the inner workings of perl or other compiler backends. It can print OPs in |
558 | the order they appear in the OP tree, in the order they will execute, or |
559 | in a text approximation to their tree structure, and the format of the |
560 | information displyed is customizable. Its function is similar to that of |
561 | perl's B<-Dx> debugging flag or the B<B::Terse> module, but it is more |
562 | sophisticated and flexible. |
563 | |
f8a679e6 |
564 | =head1 EXAMPLE |
565 | |
566 | Here's is a short example of output, using the default formatting |
567 | conventions : |
568 | |
569 | % perl -MO=Concise -e '$a = $b + 42' |
570 | 8 <@> leave[t1] vKP/REFC ->(end) |
571 | 1 <0> enter ->2 |
572 | 2 <;> nextstate(main 1 -e:1) v ->3 |
573 | 7 <2> sassign vKS/2 ->8 |
574 | 5 <2> add[t1] sK/2 ->6 |
575 | - <1> ex-rv2sv sK/1 ->4 |
576 | 3 <$> gvsv(*b) s ->4 |
577 | 4 <$> const(IV 42) s ->5 |
578 | - <1> ex-rv2sv sKRM*/1 ->7 |
579 | 6 <$> gvsv(*a) s ->7 |
580 | |
581 | Each line corresponds to an operator. Null ops appear as C<ex-opname>, |
582 | where I<opname> is the op that has been optimized away by perl. |
583 | |
584 | The number on the first row indicates the op's sequence number. It's |
585 | given in base 36 by default. |
586 | |
587 | The symbol between angle brackets indicates the op's type : for example, |
588 | <2> is a BINOP, <@> a LISTOP, etc. (see L</"OP class abbreviations">). |
589 | |
590 | The opname may be followed by op-specific information in parentheses |
591 | (e.g. C<gvsv(*b)>), and by targ information in brackets (e.g. |
592 | C<leave[t1]>). |
593 | |
594 | Next come the op flags. The common flags are listed below |
595 | (L</"OP flags abbreviations">). The private flags follow, separated |
596 | by a slash. For example, C<vKP/REFC> means that the leave op has |
597 | public flags OPf_WANT_VOID, OPf_KIDS, and OPf_PARENS, and the private |
598 | flag OPpREFCOUNTED. |
599 | |
600 | Finally an arrow points to the sequence number of the next op. |
601 | |
c99ca59a |
602 | =head1 OPTIONS |
603 | |
604 | Arguments that don't start with a hyphen are taken to be the names of |
605 | subroutines to print the OPs of; if no such functions are specified, the |
606 | main body of the program (outside any subroutines, and not including use'd |
607 | or require'd files) is printed. |
608 | |
609 | =over 4 |
610 | |
611 | =item B<-basic> |
612 | |
613 | Print OPs in the order they appear in the OP tree (a preorder |
614 | traversal, starting at the root). The indentation of each OP shows its |
615 | level in the tree. This mode is the default, so the flag is included |
616 | simply for completeness. |
617 | |
618 | =item B<-exec> |
619 | |
620 | Print OPs in the order they would normally execute (for the majority |
621 | of constructs this is a postorder traversal of the tree, ending at the |
622 | root). In most cases the OP that usually follows a given OP will |
623 | appear directly below it; alternate paths are shown by indentation. In |
624 | cases like loops when control jumps out of a linear path, a 'goto' |
625 | line is generated. |
626 | |
627 | =item B<-tree> |
628 | |
629 | Print OPs in a text approximation of a tree, with the root of the tree |
630 | at the left and 'left-to-right' order of children transformed into |
631 | 'top-to-bottom'. Because this mode grows both to the right and down, |
632 | it isn't suitable for large programs (unless you have a very wide |
633 | terminal). |
634 | |
635 | =item B<-compact> |
636 | |
637 | Use a tree format in which the minimum amount of space is used for the |
638 | lines connecting nodes (one character in most cases). This squeezes out |
639 | a few precious columns of screen real estate. |
640 | |
641 | =item B<-loose> |
642 | |
643 | Use a tree format that uses longer edges to separate OP nodes. This format |
644 | tends to look better than the compact one, especially in ASCII, and is |
645 | the default. |
646 | |
647 | =item B<-vt> |
648 | |
649 | Use tree connecting characters drawn from the VT100 line-drawing set. |
650 | This looks better if your terminal supports it. |
651 | |
652 | =item B<-ascii> |
653 | |
654 | Draw the tree with standard ASCII characters like C<+> and C<|>. These don't |
655 | look as clean as the VT100 characters, but they'll work with almost any |
656 | terminal (or the horizontal scrolling mode of less(1)) and are suitable |
657 | for text documentation or email. This is the default. |
658 | |
659 | =item B<-main> |
660 | |
661 | Include the main program in the output, even if subroutines were also |
662 | specified. |
663 | |
664 | =item B<-base>I<n> |
665 | |
666 | Print OP sequence numbers in base I<n>. If I<n> is greater than 10, the |
667 | digit for 11 will be 'a', and so on. If I<n> is greater than 36, the digit |
668 | for 37 will be 'A', and so on until 62. Values greater than 62 are not |
669 | currently supported. The default is 36. |
670 | |
671 | =item B<-bigendian> |
672 | |
673 | Print sequence numbers with the most significant digit first. This is the |
674 | usual convention for Arabic numerals, and the default. |
675 | |
676 | =item B<-littleendian> |
677 | |
678 | Print seqence numbers with the least significant digit first. |
679 | |
680 | =item B<-concise> |
681 | |
682 | Use the author's favorite set of formatting conventions. This is the |
683 | default, of course. |
684 | |
685 | =item B<-terse> |
686 | |
687 | Use formatting conventions that emulate the ouput of B<B::Terse>. The |
688 | basic mode is almost indistinguishable from the real B<B::Terse>, and the |
689 | exec mode looks very similar, but is in a more logical order and lacks |
690 | curly brackets. B<B::Terse> doesn't have a tree mode, so the tree mode |
691 | is only vaguely reminiscient of B<B::Terse>. |
692 | |
693 | =item B<-linenoise> |
694 | |
695 | Use formatting conventions in which the name of each OP, rather than being |
696 | written out in full, is represented by a one- or two-character abbreviation. |
697 | This is mainly a joke. |
698 | |
699 | =item B<-debug> |
700 | |
701 | Use formatting conventions reminiscient of B<B::Debug>; these aren't |
702 | very concise at all. |
703 | |
704 | =item B<-env> |
705 | |
706 | Use formatting conventions read from the environment variables |
707 | C<B_CONCISE_FORMAT>, C<B_CONCISE_GOTO_FORMAT>, and C<B_CONCISE_TREE_FORMAT>. |
708 | |
709 | =back |
710 | |
711 | =head1 FORMATTING SPECIFICATIONS |
712 | |
713 | For each general style ('concise', 'terse', 'linenoise', etc.) there are |
714 | three specifications: one of how OPs should appear in the basic or exec |
715 | modes, one of how 'goto' lines should appear (these occur in the exec |
716 | mode only), and one of how nodes should appear in tree mode. Each has the |
717 | same format, described below. Any text that doesn't match a special |
718 | pattern is copied verbatim. |
719 | |
720 | =over 4 |
721 | |
722 | =item B<(x(>I<exec_text>B<;>I<basic_text>B<)x)> |
723 | |
724 | Generates I<exec_text> in exec mode, or I<basic_text> in basic mode. |
725 | |
726 | =item B<(*(>I<text>B<)*)> |
727 | |
728 | Generates one copy of I<text> for each indentation level. |
729 | |
730 | =item B<(*(>I<text1>B<;>I<text2>B<)*)> |
731 | |
732 | Generates one fewer copies of I<text1> than the indentation level, followed |
733 | by one copy of I<text2> if the indentation level is more than 0. |
734 | |
735 | =item B<(?(>I<text1>B<#>I<var>I<Text2>B<)?)> |
736 | |
737 | If the value of I<var> is true (not empty or zero), generates the |
738 | value of I<var> surrounded by I<text1> and I<Text2>, otherwise |
739 | nothing. |
740 | |
741 | =item B<#>I<var> |
742 | |
743 | Generates the value of the variable I<var>. |
744 | |
745 | =item B<#>I<var>I<N> |
746 | |
747 | Generates the value of I<var>, left jutified to fill I<N> spaces. |
748 | |
749 | =item B<~> |
750 | |
751 | Any number of tildes and surrounding whitespace will be collapsed to |
752 | a single space. |
753 | |
754 | =back |
755 | |
756 | The following variables are recognized: |
757 | |
758 | =over 4 |
759 | |
760 | =item B<#addr> |
761 | |
762 | The address of the OP, in hexidecimal. |
763 | |
764 | =item B<#arg> |
765 | |
766 | The OP-specific information of the OP (such as the SV for an SVOP, the |
767 | non-local exit pointers for a LOOP, etc.) enclosed in paretheses. |
768 | |
769 | =item B<#class> |
770 | |
771 | The B-determined class of the OP, in all caps. |
772 | |
f8a679e6 |
773 | =item B<#classsym> |
c99ca59a |
774 | |
775 | A single symbol abbreviating the class of the OP. |
776 | |
c3caa09d |
777 | =item B<#coplabel> |
778 | |
779 | The label of the statement or block the OP is the start of, if any. |
780 | |
c99ca59a |
781 | =item B<#exname> |
782 | |
783 | The name of the OP, or 'ex-foo' if the OP is a null that used to be a foo. |
784 | |
785 | =item B<#extarg> |
786 | |
787 | The target of the OP, or nothing for a nulled OP. |
788 | |
789 | =item B<#firstaddr> |
790 | |
791 | The address of the OP's first child, in hexidecimal. |
792 | |
793 | =item B<#flags> |
794 | |
795 | The OP's flags, abbreviated as a series of symbols. |
796 | |
797 | =item B<#flagval> |
798 | |
799 | The numeric value of the OP's flags. |
800 | |
f8a679e6 |
801 | =item B<#hyphseq> |
c99ca59a |
802 | |
803 | The sequence number of the OP, or a hyphen if it doesn't have one. |
804 | |
805 | =item B<#label> |
806 | |
807 | 'NEXT', 'LAST', or 'REDO' if the OP is a target of one of those in exec |
808 | mode, or empty otherwise. |
809 | |
810 | =item B<#lastaddr> |
811 | |
812 | The address of the OP's last child, in hexidecimal. |
813 | |
814 | =item B<#name> |
815 | |
816 | The OP's name. |
817 | |
818 | =item B<#NAME> |
819 | |
820 | The OP's name, in all caps. |
821 | |
822 | =item B<#next> |
823 | |
824 | The sequence number of the OP's next OP. |
825 | |
826 | =item B<#nextaddr> |
827 | |
828 | The address of the OP's next OP, in hexidecimal. |
829 | |
830 | =item B<#noise> |
831 | |
832 | The two-character abbreviation for the OP's name. |
833 | |
834 | =item B<#private> |
835 | |
836 | The OP's private flags, rendered with abbreviated names if possible. |
837 | |
838 | =item B<#privval> |
839 | |
840 | The numeric value of the OP's private flags. |
841 | |
842 | =item B<#seq> |
843 | |
844 | The sequence number of the OP. |
845 | |
846 | =item B<#seqnum> |
847 | |
848 | The real sequence number of the OP, as a regular number and not adjusted |
849 | to be relative to the start of the real program. (This will generally be |
850 | a fairly large number because all of B<B::Concise> is compiled before |
851 | your program is). |
852 | |
853 | =item B<#sibaddr> |
854 | |
855 | The address of the OP's next youngest sibling, in hexidecimal. |
856 | |
857 | =item B<#svaddr> |
858 | |
859 | The address of the OP's SV, if it has an SV, in hexidecimal. |
860 | |
861 | =item B<#svclass> |
862 | |
863 | The class of the OP's SV, if it has one, in all caps (e.g., 'IV'). |
864 | |
865 | =item B<#svval> |
866 | |
867 | The value of the OP's SV, if it has one, in a short human-readable format. |
868 | |
869 | =item B<#targ> |
870 | |
871 | The numeric value of the OP's targ. |
872 | |
873 | =item B<#targarg> |
874 | |
875 | The name of the variable the OP's targ refers to, if any, otherwise the |
876 | letter t followed by the OP's targ in decimal. |
877 | |
878 | =item B<#targarglife> |
879 | |
880 | Same as B<#targarg>, but followed by the COP sequence numbers that delimit |
881 | the variable's lifetime (or 'end' for a variable in an open scope) for a |
882 | variable. |
883 | |
884 | =item B<#typenum> |
885 | |
886 | The numeric value of the OP's type, in decimal. |
887 | |
888 | =back |
889 | |
890 | =head1 ABBREVIATIONS |
891 | |
892 | =head2 OP flags abbreviations |
893 | |
894 | v OPf_WANT_VOID Want nothing (void context) |
895 | s OPf_WANT_SCALAR Want single value (scalar context) |
896 | l OPf_WANT_LIST Want list of any length (list context) |
897 | K OPf_KIDS There is a firstborn child. |
898 | P OPf_PARENS This operator was parenthesized. |
899 | (Or block needs explicit scope entry.) |
900 | R OPf_REF Certified reference. |
901 | (Return container, not containee). |
902 | M OPf_MOD Will modify (lvalue). |
903 | S OPf_STACKED Some arg is arriving on the stack. |
904 | * OPf_SPECIAL Do something weird for this op (see op.h) |
905 | |
906 | =head2 OP class abbreviations |
907 | |
908 | 0 OP (aka BASEOP) An OP with no children |
909 | 1 UNOP An OP with one child |
910 | 2 BINOP An OP with two children |
911 | | LOGOP A control branch OP |
912 | @ LISTOP An OP that could have lots of children |
913 | / PMOP An OP with a regular expression |
914 | $ SVOP An OP with an SV |
915 | " PVOP An OP with a string |
916 | { LOOP An OP that holds pointers for a loop |
917 | ; COP An OP that marks the start of a statement |
051f02e9 |
918 | # PADOP An OP with a GV on the pad |
c99ca59a |
919 | |
78ad9108 |
920 | =head1 Using B::Concise outside of the O framework |
921 | |
922 | It is possible to extend B<B::Concise> by using it outside of the B<O> |
923 | framework and providing new styles and new variables. |
924 | |
925 | use B::Concise qw(set_style add_callback); |
926 | set_style($format, $gotofmt, $treefmt); |
927 | add_callback |
928 | ( |
929 | sub |
930 | { |
931 | my ($h, $op, $level, $format) = @_; |
932 | $h->{variable} = some_func($op); |
933 | } |
934 | ); |
935 | B::Concise::compile(@options)->(); |
936 | |
937 | You can specify a style by calling the B<set_style> subroutine. If you |
938 | have a new variable in your style, or you want to change the value of an |
939 | existing variable, you will need to add a callback to specify the value |
940 | for that variable. |
941 | |
942 | This is done by calling B<add_callback> passing references to any |
943 | callback subroutines. The subroutines are called in the same order as |
944 | they are added. Each subroutine is passed four parameters. These are a |
945 | reference to a hash, the keys of which are the names of the variables |
946 | and the values of which are their values, the op, the level and the |
947 | format. |
948 | |
949 | To define your own variables, simply add them to the hash, or change |
950 | existing values if you need to. The level and format are passed in as |
951 | references to scalars, but it is unlikely that they will need to be |
952 | changed or even used. |
953 | |
954 | To see the output, call the subroutine returned by B<compile> in the |
955 | same way that B<O> does. |
956 | |
c99ca59a |
957 | =head1 AUTHOR |
958 | |
959 | Stephen McCamant, C<smcc@CSUA.Berkeley.EDU> |
960 | |
961 | =cut |