1B::Concise(3pm) Perl Programmers Reference Guide B::Concise(3pm)
2
6 B::Concise - Walk Perl syntax tree, printing concise info about ops
7
9 perl -MO=Concise[,OPTIONS] foo.pl
10 use B::Concise qw(set_style add_callback);
12
14 This compiler backend prints the internal OPs of a Perl program's syn‐
15 tax tree in one of several space-efficient text formats suitable for
16 debugging the inner workings of perl or other compiler backends. It can
17 print OPs in the order they appear in the OP tree, in the order they
18 will execute, or in a text approximation to their tree structure, and
19 the format of the information displayed is customizable. Its function
20 is similar to that of perl's -Dx debugging flag or the B::Terse module,
21 but it is more sophisticated and flexible.
22
24 Here's an example of 2 outputs (aka 'renderings'), using the -exec and
25 -basic (i.e. default) formatting conventions on the same code snippet.
26 % perl -MO=Concise,-exec -e '$a = $b + 42'
28 1 <0> enter
29 2 <;> nextstate(main 1 -e:1) v
30 3 <#> gvsv[*b] s
31 4 <$> const[IV 42] s
32 * 5 <2> add[t3] sK/2
33 6 <#> gvsv[*a] s
34 7 <2> sassign vKS/2
35 8 <@> leave[1 ref] vKP/REFC
36 Each line corresponds to an opcode. The opcode marked with '*' is used
38 in a few examples below.
39 The 1st column is the op's sequence number, starting at 1, and is dis‐
41 played in base 36 by default. This rendering is in -exec (i.e. execu‐
42 tion) order.
43 The symbol between angle brackets indicates the op's type, for example;
45 <2> is a BINOP, <@> a LISTOP, and <#> is a PADOP, which is used in
46 threaded perls. (see "OP class abbreviations").
47 The opname, as in 'add[t1]', which may be followed by op-specific
49 information in parentheses or brackets (ex '[t1]').
50 The op-flags (ex 'sK/2') follow, and are described in ("OP flags abbre‐
52 viations").
53 % perl -MO=Concise -e '$a = $b + 42'
55 8 <@> leave[1 ref] vKP/REFC ->(end)
56 1 <0> enter ->2
57 2 <;> nextstate(main 1 -e:1) v ->3
58 7 <2> sassign vKS/2 ->8
59 * 5 <2> add[t1] sK/2 ->6
60 - <1> ex-rv2sv sK/1 ->4
61 3 <$> gvsv(*b) s ->4
62 4 <$> const(IV 42) s ->5
63 - <1> ex-rv2sv sKRM*/1 ->7
64 6 <$> gvsv(*a) s ->7
65 The default rendering is top-down, so they're not in execution order.
67 This form reflects the way the stack is used to parse and evaluate
68 expressions; the add operates on the two terms below it in the tree.
69 Nullops appear as "ex-opname", where opname is an op that has been
71 optimized away by perl. They're displayed with a sequence-number of
72 '-', because they are not executed (they don't appear in previous exam‐
73 ple), they're printed here because they reflect the parse.
74 The arrow points to the sequence number of the next op; they're not
76 displayed in -exec mode, for obvious reasons.
77 Note that because this rendering was done on a non-threaded perl, the
79 PADOPs in the previous examples are now SVOPs, and some (but not all)
80 of the square brackets have been replaced by round ones. This is a
81 subtle feature to provide some visual distinction between renderings on
82 threaded and un-threaded perls.
83
85 Arguments that don't start with a hyphen are taken to be the names of
86 subroutines to print the OPs of; if no such functions are specified,
87 the main body of the program (outside any subroutines, and not includ‐
88 ing use'd or require'd files) is rendered. Passing "BEGIN", "CHECK",
89 "INIT", or "END" will cause all of the corresponding special blocks to
90 be printed.
91 Options affect how things are rendered (ie printed). They're presented
93 here by their visual effect, 1st being strongest. They're grouped
94 according to how they interrelate; within each group the options are
95 mutually exclusive (unless otherwise stated).
96 Options for Opcode Ordering
98 These options control the 'vertical display' of opcodes. The display
100 'order' is also called 'mode' elsewhere in this document.
101 -basic
103 Print OPs in the order they appear in the OP tree (a preorder tra‐
104 versal, starting at the root). The indentation of each OP shows its
105 level in the tree, and the '->' at the end of the line indicates
106 the next opcode in execution order. This mode is the default, so
107 the flag is included simply for completeness.
108 -exec
110 Print OPs in the order they would normally execute (for the major‐
111 ity of constructs this is a postorder traversal of the tree, ending
112 at the root). In most cases the OP that usually follows a given OP
113 will appear directly below it; alternate paths are shown by inden‐
114 tation. In cases like loops when control jumps out of a linear
115 path, a 'goto' line is generated.
116 -tree
118 Print OPs in a text approximation of a tree, with the root of the
119 tree at the left and 'left-to-right' order of children transformed
120 into 'top-to-bottom'. Because this mode grows both to the right and
121 down, it isn't suitable for large programs (unless you have a very
122 wide terminal).
123 Options for Line-Style
125 These options select the line-style (or just style) used to render each
127 opcode, and dictates what info is actually printed into each line.
128 -concise
130 Use the author's favorite set of formatting conventions. This is
131 the default, of course.
132 -terse
134 Use formatting conventions that emulate the output of B::Terse. The
135 basic mode is almost indistinguishable from the real B::Terse, and
136 the exec mode looks very similar, but is in a more logical order
137 and lacks curly brackets. B::Terse doesn't have a tree mode, so the
138 tree mode is only vaguely reminiscent of B::Terse.
139 -linenoise
141 Use formatting conventions in which the name of each OP, rather
142 than being written out in full, is represented by a one- or two-
143 character abbreviation. This is mainly a joke.
144 -debug
146 Use formatting conventions reminiscent of B::Debug; these aren't
147 very concise at all.
148 -env
150 Use formatting conventions read from the environment variables
151 "B_CONCISE_FORMAT", "B_CONCISE_GOTO_FORMAT", and "B_CON‐
152 CISE_TREE_FORMAT".
153 Options for tree-specific formatting
155 -compact
157 Use a tree format in which the minimum amount of space is used for
158 the lines connecting nodes (one character in most cases). This
159 squeezes out a few precious columns of screen real estate.
160 -loose
162 Use a tree format that uses longer edges to separate OP nodes. This
163 format tends to look better than the compact one, especially in
164 ASCII, and is the default.
165 -vt Use tree connecting characters drawn from the VT100 line-drawing
167 set. This looks better if your terminal supports it.
168 -ascii
170 Draw the tree with standard ASCII characters like "+" and "⎪".
171 These don't look as clean as the VT100 characters, but they'll work
172 with almost any terminal (or the horizontal scrolling mode of
173 less(1)) and are suitable for text documentation or email. This is
174 the default.
175 These are pairwise exclusive, i.e. compact or loose, vt or ascii.
177 Options controlling sequence numbering
179 -basen
181 Print OP sequence numbers in base n. If n is greater than 10, the
182 digit for 11 will be 'a', and so on. If n is greater than 36, the
183 digit for 37 will be 'A', and so on until 62. Values greater than
184 62 are not currently supported. The default is 36.
185 -bigendian
187 Print sequence numbers with the most significant digit first. This
188 is the usual convention for Arabic numerals, and the default.
189 -littleendian
191 Print seqence numbers with the least significant digit first. This
192 is obviously mutually exclusive with bigendian.
193 Other options
195 These are pairwise exclusive.
197 -main
199 Include the main program in the output, even if subroutines were
200 also specified. This rendering is normally suppressed when a sub‐
201 routine name or reference is given.
202 -nomain
204 This restores the default behavior after you've changed it with
205 '-main' (it's not normally needed). If no subroutine name/ref is
206 given, main is rendered, regardless of this flag.
207 -nobanner
209 Renderings usually include a banner line identifying the function
210 name or stringified subref. This suppresses the printing of the
211 banner.
212 TBC: Remove the stringified coderef; while it provides a 'cookie'
214 for each function rendered, the cookies used should be 1,2,3.. not
215 a random hex-address. It also complicates string comparison of two
216 different trees.
217 -banner
219 restores default banner behavior.
220 -banneris => subref
222 TBC: a hookpoint (and an option to set it) for a user-supplied
223 function to produce a banner appropriate for users needs. It's not
224 ideal, because the rendering-state variables, which are a natural
225 candidate for use in concise.t, are unavailable to the user.
226 Option Stickiness
228 If you invoke Concise more than once in a program, you should know that
230 the options are 'sticky'. This means that the options you provide in
231 the first call will be remembered for the 2nd call, unless you re-spec‐
232 ify or change them.
233
235 The concise style uses symbols to convey maximum info with minimal
236 clutter (like hex addresses). With just a little practice, you can
237 start to see the flowers, not just the branches, in the trees.
238 OP class abbreviations
240 These symbols appear before the op-name, and indicate the B:: namespace
242 that represents the ops in your Perl code.
243 0 OP (aka BASEOP) An OP with no children
245 1 UNOP An OP with one child
246 2 BINOP An OP with two children
247 ⎪ LOGOP A control branch OP
248 @ LISTOP An OP that could have lots of children
249 / PMOP An OP with a regular expression
250 $ SVOP An OP with an SV
251 " PVOP An OP with a string
252 { LOOP An OP that holds pointers for a loop
253 ; COP An OP that marks the start of a statement
254 # PADOP An OP with a GV on the pad
255 OP flags abbreviations
257 OP flags are either public or private. The public flags alter the
259 behavior of each opcode in consistent ways, and are represented by 0 or
260 more single characters.
261 v OPf_WANT_VOID Want nothing (void context)
263 s OPf_WANT_SCALAR Want single value (scalar context)
264 l OPf_WANT_LIST Want list of any length (list context)
265 Want is unknown
266 K OPf_KIDS There is a firstborn child.
267 P OPf_PARENS This operator was parenthesized.
268 (Or block needs explicit scope entry.)
269 R OPf_REF Certified reference.
270 (Return container, not containee).
271 M OPf_MOD Will modify (lvalue).
272 S OPf_STACKED Some arg is arriving on the stack.
273 * OPf_SPECIAL Do something weird for this op (see op.h)
274 Private flags, if any are set for an opcode, are displayed after a '/'
276 8 <@> leave[1 ref] vKP/REFC ->(end)
278 7 <2> sassign vKS/2 ->8
279 They're opcode specific, and occur less often than the public ones, so
281 they're represented by short mnemonics instead of single-chars; see
282 op.h for gory details, or try this quick 2-liner:
283 $> perl -MB::Concise -de 1
285 DB<1> ⎪x \%B::Concise::priv
286
288 For each line-style ('concise', 'terse', 'linenoise', etc.) there are 3
289 format-specs which control how OPs are rendered.
290 The first is the 'default' format, which is used in both basic and exec
292 modes to print all opcodes. The 2nd, goto-format, is used in exec mode
293 when branches are encountered. They're not real opcodes, and are
294 inserted to look like a closing curly brace. The tree-format is tree
295 specific.
296 When a line is rendered, the correct format-spec is copied and scanned
298 for the following items; data is substituted in, and other manipula‐
299 tions like basic indenting are done, for each opcode rendered.
300 There are 3 kinds of items that may be populated; special patterns,
302 #vars, and literal text, which is copied verbatim. (Yes, it's a set of
303 s///g steps.)
304 Special Patterns
306 These items are the primitives used to perform indenting, and to select
308 text from amongst alternatives.
309 (x(exec_text;basic_text)x)
311 Generates exec_text in exec mode, or basic_text in basic mode.
312 (*(text)*)
314 Generates one copy of text for each indentation level.
315 (*(text1;text2)*)
317 Generates one fewer copies of text1 than the indentation level,
318 followed by one copy of text2 if the indentation level is more than
319 0.
320 (?(text1#varText2)?)
322 If the value of var is true (not empty or zero), generates the
323 value of var surrounded by text1 and Text2, otherwise nothing.
324 ~ Any number of tildes and surrounding whitespace will be collapsed
326 to a single space.
327 # Variables
329 These #vars represent opcode properties that you may want as part of
331 your rendering. The '#' is intended as a private sigil; a #var's value
332 is interpolated into the style-line, much like "read $this".
333 These vars take 3 forms:
335 #var
337 A property named 'var' is assumed to exist for the opcodes, and is
338 interpolated into the rendering.
339 #varN
341 Generates the value of var, left justified to fill N spaces. Note
342 that this means while you can have properties 'foo' and 'foo2', you
343 cannot render 'foo2', but you could with 'foo2a'. You would be
344 wise not to rely on this behavior going forward ;-)
345 #Var
347 This ucfirst form of #var generates a tag-value form of itself for
348 display; it converts '#Var' into a 'Var => #var' style, which is
349 then handled as described above. (Imp-note: #Vars cannot be used
350 for conditional-fills, because the => #var transform is done after
351 the check for #Var's value).
352 The following variables are 'defined' by B::Concise; when they are used
354 in a style, their respective values are plugged into the rendering of
355 each opcode.
356 Only some of these are used by the standard styles, the others are pro‐
358 vided for you to delve into optree mechanics, should you wish to add a
359 new style (see "add_style" below) that uses them. You can also add new
360 ones using "add_callback".
361 #addr
363 The address of the OP, in hexadecimal.
364 #arg
366 The OP-specific information of the OP (such as the SV for an SVOP,
367 the non-local exit pointers for a LOOP, etc.) enclosed in parenthe‐
368 ses.
369 #class
371 The B-determined class of the OP, in all caps.
372 #classsym
374 A single symbol abbreviating the class of the OP.
375 #coplabel
377 The label of the statement or block the OP is the start of, if any.
378 #exname
380 The name of the OP, or 'ex-foo' if the OP is a null that used to be
381 a foo.
382 #extarg
384 The target of the OP, or nothing for a nulled OP.
385 #firstaddr
387 The address of the OP's first child, in hexadecimal.
388 #flags
390 The OP's flags, abbreviated as a series of symbols.
391 #flagval
393 The numeric value of the OP's flags.
394 #hyphseq
396 The sequence number of the OP, or a hyphen if it doesn't have one.
397 #label
399 'NEXT', 'LAST', or 'REDO' if the OP is a target of one of those in
400 exec mode, or empty otherwise.
401 #lastaddr
403 The address of the OP's last child, in hexadecimal.
404 #name
406 The OP's name.
407 #NAME
409 The OP's name, in all caps.
410 #next
412 The sequence number of the OP's next OP.
413 #nextaddr
415 The address of the OP's next OP, in hexadecimal.
416 #noise
418 A one- or two-character abbreviation for the OP's name.
419 #private
421 The OP's private flags, rendered with abbreviated names if possi‐
422 ble.
423 #privval
425 The numeric value of the OP's private flags.
426 #seq
428 The sequence number of the OP. Note that this is a sequence number
429 generated by B::Concise.
430 #seqnum
432 5.8.x and earlier only. 5.9 and later do not provide this.
433 The real sequence number of the OP, as a regular number and not
435 adjusted to be relative to the start of the real program. (This
436 will generally be a fairly large number because all of B::Concise
437 is compiled before your program is).
438 #opt
440 Whether or not the op has been optimised by the peephole optimiser.
441 Only available in 5.9 and later.
443 #static
445 Whether or not the op is statically defined. This flag is used by
446 the B::C compiler backend and indicates that the op should not be
447 freed.
448 Only available in 5.9 and later.
450 #sibaddr
452 The address of the OP's next youngest sibling, in hexadecimal.
453 #svaddr
455 The address of the OP's SV, if it has an SV, in hexadecimal.
456 #svclass
458 The class of the OP's SV, if it has one, in all caps (e.g., 'IV').
459 #svval
461 The value of the OP's SV, if it has one, in a short human-readable
462 format.
463 #targ
465 The numeric value of the OP's targ.
466 #targarg
468 The name of the variable the OP's targ refers to, if any, otherwise
469 the letter t followed by the OP's targ in decimal.
470 #targarglife
472 Same as #targarg, but followed by the COP sequence numbers that
473 delimit the variable's lifetime (or 'end' for a variable in an open
474 scope) for a variable.
475 #typenum
477 The numeric value of the OP's type, in decimal.
478
480 The common (and original) usage of B::Concise was for command-line ren‐
481 derings of simple code, as given in EXAMPLE. But you can also use
482 B::Concise from your code, and call compile() directly, and repeatedly.
483 By doing so, you can avoid the compile-time only operation of O.pm, and
484 even use the debugger to step through B::Concise::compile() itself.
485 Once you're doing this, you may alter Concise output by adding new ren‐
487 dering styles, and by optionally adding callback routines which popu‐
488 late new variables, if such were referenced from those (just added)
489 styles.
490 Example: Altering Concise Renderings
492 use B::Concise qw(set_style add_callback);
494 add_style($yourStyleName => $defaultfmt, $gotofmt, $treefmt);
495 add_callback
496 ( sub {
497 my ($h, $op, $format, $level, $stylename) = @_;
498 $h->{variable} = some_func($op);
499 });
500 $walker = B::Concise::compile(@options,@subnames,@subrefs);
501 $walker->();
502 set_style()
504 set_style accepts 3 arguments, and updates the three format-specs com‐
506 prising a line-style (basic-exec, goto, tree). It has one minor draw‐
507 back though; it doesn't register the style under a new name. This can
508 become an issue if you render more than once and switch styles. Thus
509 you may prefer to use add_style() and/or set_style_standard() instead.
510 set_style_standard($name)
512 This restores one of the standard line-styles: "terse", "concise",
514 "linenoise", "debug", "env", into effect. It also accepts style names
515 previously defined with add_style().
516 add_style()
518 This subroutine accepts a new style name and three style arguments as
520 above, and creates, registers, and selects the newly named style. It
521 is an error to re-add a style; call set_style_standard() to switch
522 between several styles.
523 add_callback()
525 If your newly minted styles refer to any new #variables, you'll need to
527 define a callback subroutine that will populate (or modify) those vari‐
528 ables. They are then available for use in the style you've chosen.
529 The callbacks are called for each opcode visited by Concise, in the
531 same order as they are added. Each subroutine is passed five parame‐
532 ters.
533 1. A hashref, containing the variable names and values which are
535 populated into the report-line for the op
536 2. the op, as a B<B::OP> object
537 3. a reference to the format string
538 4. the formatting (indent) level
539 5. the selected stylename
540 To define your own variables, simply add them to the hash, or change
542 existing values if you need to. The level and format are passed in as
543 references to scalars, but it is unlikely that they will need to be
544 changed or even used.
545 Running B::Concise::compile()
547 compile accepts options as described above in "OPTIONS", and arguments,
549 which are either coderefs, or subroutine names.
550 It constructs and returns a $treewalker coderef, which when invoked,
552 traverses, or walks, and renders the optrees of the given arguments to
553 STDOUT. You can reuse this, and can change the rendering style used
554 each time; thereafter the coderef renders in the new style.
555 walk_output lets you change the print destination from STDOUT to
557 another open filehandle, or into a string passed as a ref (unless
558 you've built perl with -Uuseperlio).
559 my $walker = B::Concise::compile('-terse','aFuncName', \&aSubRef); # 1
561 walk_output(\my $buf);
562 $walker->(); # 1 renders -terse
563 set_style_standard('concise'); # 2
564 $walker->(); # 2 renders -concise
565 $walker->(@new); # 3 renders whatever
566 print "3 different renderings: terse, concise, and @new: $buf\n";
567 When $walker is called, it traverses the subroutines supplied when it
569 was created, and renders them using the current style. You can change
570 the style afterwards in several different ways:
571 1. call C<compile>, altering style or mode/order
573 2. call C<set_style_standard>
574 3. call $walker, passing @new options
575 Passing new options to the $walker is the easiest way to change amongst
577 any pre-defined styles (the ones you add are automatically recognized
578 as options), and is the only way to alter rendering order without call‐
579 ing compile again. Note however that rendering state is still shared
580 amongst multiple $walker objects, so they must still be used in a coor‐
581 dinated manner.
582 B::Concise::reset_sequence()
584 This function (not exported) lets you reset the sequence numbers (note
586 that they're numbered arbitrarily, their goal being to be human read‐
587 able). Its purpose is mostly to support testing, i.e. to compare the
588 concise output from two identical anonymous subroutines (but different
589 instances). Without the reset, B::Concise, seeing that they're sepa‐
590 rate optrees, generates different sequence numbers in the output.
591 Errors
593 Errors in rendering (non-existent function-name, non-existent coderef)
595 are written to the STDOUT, or wherever you've set it via walk_output().
596 Errors using the various *style* calls, and bad args to walk_output(),
598 result in die(). Use an eval if you wish to catch these errors and
599 continue processing.
600
602 Stephen McCamant, <smcc@CSUA.Berkeley.EDU>.
603perl v5.8.8 2001-09-21 B::Concise(3pm)