Perl 5 version 32.0 documentation

perldebguts

NAME

perldebguts - Guts of Perl debugging

DESCRIPTION

This is not perldebug, which tells you how to use the debugger. This manpage describes low-level details concerning the debugger's internals, which range from difficult to impossible to understand for anyone who isn't incredibly intimate with Perl's guts. Caveat lector.

Debugger Internals

Perl has special debugging hooks at compile-time and run-time used to create debugging environments. These hooks are not to be confused with the perl -Dxxx command described in perlrun, which is usable only if a special Perl is built per the instructions in the INSTALL podpage in the Perl source tree.

For example, whenever you call Perl's built-in caller function from the package DB , the arguments that the corresponding stack frame was called with are copied to the @DB::args array. These mechanisms are enabled by calling Perl with the -d switch. Specifically, the following additional features are enabled (cf. $^P in perlvar):

  • Perl inserts the contents of $ENV{PERL5DB} (or BEGIN {require 'perl5db.pl'} if not present) before the first line of your program.

  • Each array @{"_<$filename"} holds the lines of $filename for a file compiled by Perl. The same is also true for evaled strings that contain subroutines, or which are currently being executed. The $filename for evaled strings looks like (eval 34) .

    Values in this array are magical in numeric context: they compare equal to zero only if the line is not breakable.

  • Each hash %{"_<$filename"} contains breakpoints and actions keyed by line number. Individual entries (as opposed to the whole hash) are settable. Perl only cares about Boolean true here, although the values used by perl5db.pl have the form "$break_condition\0$action" .

    The same holds for evaluated strings that contain subroutines, or which are currently being executed. The $filename for evaled strings looks like (eval 34) .

  • Each scalar ${"_<$filename"} contains "_<$filename" . This is also the case for evaluated strings that contain subroutines, or which are currently being executed. The $filename for evaled strings looks like (eval 34) .

  • After each required file is compiled, but before it is executed, DB::postponed(*{"_<$filename"}) is called if the subroutine DB::postponed exists. Here, the $filename is the expanded name of the required file, as found in the values of %INC.

  • After each subroutine subname is compiled, the existence of $DB::postponed{subname} is checked. If this key exists, DB::postponed(subname) is called if the DB::postponed subroutine also exists.

  • A hash %DB::sub is maintained, whose keys are subroutine names and whose values have the form filename:startline-endline . filename has the form (eval 34) for subroutines defined inside evals.

  • When the execution of your program reaches a point that can hold a breakpoint, the DB::DB() subroutine is called if any of the variables $DB::trace , $DB::single , or $DB::signal is true. These variables are not localizable. This feature is disabled when executing inside DB::DB() , including functions called from it unless $^D & (1<<30) is true.

  • When execution of the program reaches a subroutine call, a call to &DB::sub (args) is made instead, with $DB::sub set to identify the called subroutine. (This doesn't happen if the calling subroutine was compiled in the DB package.) $DB::sub normally holds the name of the called subroutine, if it has a name by which it can be looked up. Failing that, $DB::sub will hold a reference to the called subroutine. Either way, the &DB::sub subroutine can use $DB::sub as a reference by which to call the called subroutine, which it will normally want to do.

    If the call is to an lvalue subroutine, and &DB::lsub is defined &DB::lsub (args) is called instead, otherwise falling back to &DB::sub (args).

  • When execution of the program uses goto to enter a non-XS subroutine and the 0x80 bit is set in $^P , a call to &DB::goto is made, with $DB::sub set to identify the subroutine being entered. The call to &DB::goto does not replace the goto; the requested subroutine will still be entered once &DB::goto has returned. $DB::sub normally holds the name of the subroutine being entered, if it has one. Failing that, $DB::sub will hold a reference to the subroutine being entered. Unlike when &DB::sub is called, it is not guaranteed that $DB::sub can be used as a reference to operate on the subroutine being entered.

Note that if &DB::sub needs external data for it to work, no subroutine call is possible without it. As an example, the standard debugger's &DB::sub depends on the $DB::deep variable (it defines how many levels of recursion deep into the debugger you can go before a mandatory break). If $DB::deep is not defined, subroutine calls are not possible, even though &DB::sub exists.

Writing Your Own Debugger

Environment Variables

The PERL5DB environment variable can be used to define a debugger. For example, the minimal "working" debugger (it actually doesn't do anything) consists of one line:

  1.   sub DB::DB {}

It can easily be defined like this:

  1.   $ PERL5DB="sub DB::DB {}" perl -d your-script

Another brief debugger, slightly more useful, can be created with only the line:

  1.   sub DB::DB {print ++$i; scalar <STDIN>}

This debugger prints a number which increments for each statement encountered and waits for you to hit a newline before continuing to the next statement.

The following debugger is actually useful:

  1.   {
  2.     package DB;
  3.     sub DB  {}
  4.     sub sub {print ++$i, " $sub\n"; &$sub}
  5.   }

It prints the sequence number of each subroutine call and the name of the called subroutine. Note that &DB::sub is being compiled into the package DB through the use of the package directive.

When it starts, the debugger reads your rc file (./.perldb or ~/.perldb under Unix), which can set important options. (A subroutine (&afterinit ) can be defined here as well; it is executed after the debugger completes its own initialization.)

After the rc file is read, the debugger reads the PERLDB_OPTS environment variable and uses it to set debugger options. The contents of this variable are treated as if they were the argument of an o ... debugger command (q.v. in Configurable Options in perldebug).

Debugger Internal Variables

In addition to the file and subroutine-related variables mentioned above, the debugger also maintains various magical internal variables.

  • @DB::dbline is an alias for @{"::_<current_file"} , which holds the lines of the currently-selected file (compiled by Perl), either explicitly chosen with the debugger's f command, or implicitly by flow of execution.

    Values in this array are magical in numeric context: they compare equal to zero only if the line is not breakable.

  • %DB::dbline is an alias for %{"::_<current_file"} , which contains breakpoints and actions keyed by line number in the currently-selected file, either explicitly chosen with the debugger's f command, or implicitly by flow of execution.

    As previously noted, individual entries (as opposed to the whole hash) are settable. Perl only cares about Boolean true here, although the values used by perl5db.pl have the form "$break_condition\0$action" .

Debugger Customization Functions

Some functions are provided to simplify customization.

  • See Configurable Options in perldebug for a description of options parsed by DB::parse_options(string) .

  • DB::dump_trace(skip[,count]) skips the specified number of frames and returns a list containing information about the calling frames (all of them, if count is missing). Each entry is reference to a hash with keys context (either ., $ , or @ ), sub (subroutine name, or info about eval), args (undef or a reference to an array), file , and line .

  • DB::print_trace(FH, skip[, count[, short]]) prints formatted info about caller frames. The last two functions may be convenient as arguments to <, << commands.

Note that any variables and functions that are not documented in this manpages (or in perldebug) are considered for internal use only, and as such are subject to change without notice.

Frame Listing Output Examples

The frame option can be used to control the output of frame information. For example, contrast this expression trace:

  1.  $ perl -de 42
  2.  Stack dump during die enabled outside of evals.
  4.  Loading DB routines from perl5db.pl patch level 0.94
  5.  Emacs support available.
  7.  Enter h or 'h h' for help.
  9.  main::(-e:1):   0
  10.    DB<1> sub foo { 14 }
  12.    DB<2> sub bar { 3 }
  14.    DB<3> t print foo() * bar()
  15.  main::((eval 172):3):   print foo() + bar();
  16.  main::foo((eval 168):2):
  17.  main::bar((eval 170):2):
  18.  42

with this one, once the o ption frame=2 has been set:

  1.    DB<4> o f=2
  2.                 frame = '2'
  3.    DB<5> t print foo() * bar()
  4.  3:      foo() * bar()
  5.  entering main::foo
  6.   2:     sub foo { 14 };
  7.  exited main::foo
  8.  entering main::bar
  9.   2:     sub bar { 3 };
  10.  exited main::bar
  11.  42

By way of demonstration, we present below a laborious listing resulting from setting your PERLDB_OPTS environment variable to the value f=n N , and running perl -d -V from the command line. Examples using various values of n are shown to give you a feel for the difference between settings. Long though it may be, this is not a complete listing, but only excerpts.

    1.  entering main::BEGIN
    2.   entering Config::BEGIN
    3.    Package lib/Exporter.pm.
    4.    Package lib/Carp.pm.
    5.   Package lib/Config.pm.
    6.   entering Config::TIEHASH
    7.   entering Exporter::import
    8.    entering Exporter::export
    9.  entering Config::myconfig
    10.   entering Config::FETCH
    11.   entering Config::FETCH
    12.   entering Config::FETCH
    13.   entering Config::FETCH
    1.  entering main::BEGIN
    2.   entering Config::BEGIN
    3.    Package lib/Exporter.pm.
    4.    Package lib/Carp.pm.
    5.   exited Config::BEGIN
    6.   Package lib/Config.pm.
    7.   entering Config::TIEHASH
    8.   exited Config::TIEHASH
    9.   entering Exporter::import
    10.    entering Exporter::export
    11.    exited Exporter::export
    12.   exited Exporter::import
    13.  exited main::BEGIN
    14.  entering Config::myconfig
    15.   entering Config::FETCH
    16.   exited Config::FETCH
    17.   entering Config::FETCH
    18.   exited Config::FETCH
    19.   entering Config::FETCH
    1.  in  $=main::BEGIN() from /dev/null:0
    2.   in  $=Config::BEGIN() from lib/Config.pm:2
    3.    Package lib/Exporter.pm.
    4.    Package lib/Carp.pm.
    5.   Package lib/Config.pm.
    6.   in  $=Config::TIEHASH('Config') from lib/Config.pm:644
    7.   in  $=Exporter::import('Config', 'myconfig', 'config_vars') from /dev/null:0
    8.    in  $=Exporter::export('Config', 'main', 'myconfig', 'config_vars') from li
    9.  in  @=Config::myconfig() from /dev/null:0
    10.   in  $=Config::FETCH(ref(Config), 'package') from lib/Config.pm:574
    11.   in  $=Config::FETCH(ref(Config), 'baserev') from lib/Config.pm:574
    12.   in  $=Config::FETCH(ref(Config), 'PERL_VERSION') from lib/Config.pm:574
    13.   in  $=Config::FETCH(ref(Config), 'PERL_SUBVERSION') from lib/Config.pm:574
    14.   in  $=Config::FETCH(ref(Config), 'osname') from lib/Config.pm:574
    15.   in  $=Config::FETCH(ref(Config), 'osvers') from lib/Config.pm:574
    1.  in  $=main::BEGIN() from /dev/null:0
    2.   in  $=Config::BEGIN() from lib/Config.pm:2
    3.    Package lib/Exporter.pm.
    4.    Package lib/Carp.pm.
    5.   out $=Config::BEGIN() from lib/Config.pm:0
    6.   Package lib/Config.pm.
    7.   in  $=Config::TIEHASH('Config') from lib/Config.pm:644
    8.   out $=Config::TIEHASH('Config') from lib/Config.pm:644
    9.   in  $=Exporter::import('Config', 'myconfig', 'config_vars') from /dev/null:0
    10.    in  $=Exporter::export('Config', 'main', 'myconfig', 'config_vars') from lib/
    11.    out $=Exporter::export('Config', 'main', 'myconfig', 'config_vars') from lib/
    12.   out $=Exporter::import('Config', 'myconfig', 'config_vars') from /dev/null:0
    13.  out $=main::BEGIN() from /dev/null:0
    14.  in  @=Config::myconfig() from /dev/null:0
    15.   in  $=Config::FETCH(ref(Config), 'package') from lib/Config.pm:574
    16.   out $=Config::FETCH(ref(Config), 'package') from lib/Config.pm:574
    17.   in  $=Config::FETCH(ref(Config), 'baserev') from lib/Config.pm:574
    18.   out $=Config::FETCH(ref(Config), 'baserev') from lib/Config.pm:574
    19.   in  $=Config::FETCH(ref(Config), 'PERL_VERSION') from lib/Config.pm:574
    20.   out $=Config::FETCH(ref(Config), 'PERL_VERSION') from lib/Config.pm:574
    21.   in  $=Config::FETCH(ref(Config), 'PERL_SUBVERSION') from lib/Config.pm:574
    1.  in  $=main::BEGIN() from /dev/null:0
    2.   in  $=Config::BEGIN() from lib/Config.pm:2
    3.    Package lib/Exporter.pm.
    4.    Package lib/Carp.pm.
    5.   out $=Config::BEGIN() from lib/Config.pm:0
    6.   Package lib/Config.pm.
    7.   in  $=Config::TIEHASH('Config') from lib/Config.pm:644
    8.   out $=Config::TIEHASH('Config') from lib/Config.pm:644
    9.   in  $=Exporter::import('Config', 'myconfig', 'config_vars') from /dev/null:0
    10.    in  $=Exporter::export('Config', 'main', 'myconfig', 'config_vars') from lib/E
    11.    out $=Exporter::export('Config', 'main', 'myconfig', 'config_vars') from lib/E
    12.   out $=Exporter::import('Config', 'myconfig', 'config_vars') from /dev/null:0
    13.  out $=main::BEGIN() from /dev/null:0
    14.  in  @=Config::myconfig() from /dev/null:0
    15.   in  $=Config::FETCH('Config=HASH(0x1aa444)', 'package') from lib/Config.pm:574
    16.   out $=Config::FETCH('Config=HASH(0x1aa444)', 'package') from lib/Config.pm:574
    17.   in  $=Config::FETCH('Config=HASH(0x1aa444)', 'baserev') from lib/Config.pm:574
    18.   out $=Config::FETCH('Config=HASH(0x1aa444)', 'baserev') from lib/Config.pm:574
    1.  in  $=CODE(0x15eca4)() from /dev/null:0
    2.   in  $=CODE(0x182528)() from lib/Config.pm:2
    3.    Package lib/Exporter.pm.
    4.   out $=CODE(0x182528)() from lib/Config.pm:0
    5.   scalar context return from CODE(0x182528): undef
    6.   Package lib/Config.pm.
    7.   in  $=Config::TIEHASH('Config') from lib/Config.pm:628
    8.   out $=Config::TIEHASH('Config') from lib/Config.pm:628
    9.   scalar context return from Config::TIEHASH:   empty hash
    10.   in  $=Exporter::import('Config', 'myconfig', 'config_vars') from /dev/null:0
    11.    in  $=Exporter::export('Config', 'main', 'myconfig', 'config_vars') from lib/Exporter.pm:171
    12.    out $=Exporter::export('Config', 'main', 'myconfig', 'config_vars') from lib/Exporter.pm:171
    13.    scalar context return from Exporter::export: ''
    14.   out $=Exporter::import('Config', 'myconfig', 'config_vars') from /dev/null:0
    15.   scalar context return from Exporter::import: ''

    • In all cases shown above, the line indentation shows the call tree. If bit 2 of frame is set, a line is printed on exit from a subroutine as well. If bit 4 is set, the arguments are printed along with the caller info. If bit 8 is set, the arguments are printed even if they are tied or references. If bit 16 is set, the return value is printed, too.

    When a package is compiled, a line like this

    1.     Package lib/Carp.pm.

    is printed with proper indentation.

    Debugging Regular Expressions

    There are two ways to enable debugging output for regular expressions.

    If your perl is compiled with -DDEBUGGING , you may use the -Dr flag on the command line, and -Drv for more verbose information.

    Otherwise, one can use re 'debug' , which has effects at both compile time and run time. Since Perl 5.9.5, this pragma is lexically scoped.

    Compile-time Output

    The debugging output at compile time looks like this:

    1.   Compiling REx '[bc]d(ef*g)+h[ij]k$'
    2.   size 45 Got 364 bytes for offset annotations.
    3.   first at 1
    4.   rarest char g at 0
    5.   rarest char d at 0
    6.      1: ANYOF[bc](12)
    7.     12: EXACT <d>(14)
    8.     14: CURLYX[0] {1,32767}(28)
    9.     16:   OPEN1(18)
    10.     18:     EXACT <e>(20)
    11.     20:     STAR(23)
    12.     21:       EXACT <f>(0)
    13.     23:     EXACT <g>(25)
    14.     25:   CLOSE1(27)
    15.     27:   WHILEM[1/1](0)
    16.     28: NOTHING(29)
    17.     29: EXACT <h>(31)
    18.     31: ANYOF[ij](42)
    19.     42: EXACT <k>(44)
    20.     44: EOL(45)
    21.     45: END(0)
    22.   anchored 'de' at 1 floating 'gh' at 3..2147483647 (checking floating) 
    23.         stclass 'ANYOF[bc]' minlen 7 
    24.   Offsets: [45]
    25.   	1[4] 0[0] 0[0] 0[0] 0[0] 0[0] 0[0] 0[0] 0[0] 0[0] 0[0] 5[1]
    26.   	0[0] 12[1] 0[0] 6[1] 0[0] 7[1] 0[0] 9[1] 8[1] 0[0] 10[1] 0[0]
    27.   	11[1] 0[0] 12[0] 12[0] 13[1] 0[0] 14[4] 0[0] 0[0] 0[0] 0[0]
    28.   	0[0] 0[0] 0[0] 0[0] 0[0] 0[0] 18[1] 0[0] 19[1] 20[0]  
    29.   Omitting $` $& $' support.

    The first line shows the pre-compiled form of the regex. The second shows the size of the compiled form (in arbitrary units, usually 4-byte words) and the total number of bytes allocated for the offset/length table, usually 4+size *8. The next line shows the label id of the first node that does a match.

    The

    1.   anchored 'de' at 1 floating 'gh' at 3..2147483647 (checking floating) 
    2.         stclass 'ANYOF[bc]' minlen 7

    line (split into two lines above) contains optimizer information. In the example shown, the optimizer found that the match should contain a substring de at offset 1, plus substring gh at some offset between 3 and infinity. Moreover, when checking for these substrings (to abandon impossible matches quickly), Perl will check for the substring gh before checking for the substring de . The optimizer may also use the knowledge that the match starts (at the first id) with a character class, and no string shorter than 7 characters can possibly match.

    The fields of interest which may appear in this line are

    • anchored STRING at POS
    • floating STRING at POS1..POS2

      See above.

    • matching floating/anchored

      Which substring to check first.

    • minlen

      The minimal length of the match.

    • stclass TYPE

      Type of first matching node.

    • noscan

      Don't scan for the found substrings.

    • isall

      Means that the optimizer information is all that the regular expression contains, and thus one does not need to enter the regex engine at all.

    • GPOS

      Set if the pattern contains \G .

    • plus

      Set if the pattern starts with a repeated char (as in x+y).

    • implicit

      Set if the pattern starts with .*.

    • with eval

      Set if the pattern contain eval-groups, such as (?{ code }) and (??{ code }) .

    • anchored(TYPE)

      If the pattern may match only at a handful of places, with TYPE being SBOL , MBOL , or GPOS . See the table below.

    If a substring is known to match at end-of-line only, it may be followed by $ , as in floating 'k'$.

    The optimizer-specific information is used to avoid entering (a slow) regex engine on strings that will not definitely match. If the isall flag is set, a call to the regex engine may be avoided even when the optimizer found an appropriate place for the match.

    Above the optimizer section is the list of nodes of the compiled form of the regex. Each line has format

    id: TYPE OPTIONAL-INFO (next-id)

    Types of Nodes

    Here are the current possible types, with short descriptions:

    1.  # TYPE arg-description [regnode-struct-suffix] [longjump-len] DESCRIPTION
    3.  # Exit points
    5.  END              no         End of program.
    6.  SUCCEED          no         Return from a subroutine, basically.
    8.  # Line Start Anchors:
    9.  SBOL             no         Match "" at beginning of line: /^/, /\A/
    10.  MBOL             no         Same, assuming multiline: /^/m
    12.  # Line End Anchors:
    13.  SEOL             no         Match "" at end of line: /$/
    14.  MEOL             no         Same, assuming multiline: /$/m
    15.  EOS              no         Match "" at end of string: /\z/
    17.  # Match Start Anchors:
    18.  GPOS             no         Matches where last m//g left off.
    20.  # Word Boundary Opcodes:
    21.  BOUND            no         Like BOUNDA for non-utf8, otherwise like
    22.                              BOUNDU
    23.  BOUNDL           no         Like BOUND/BOUNDU, but \w and \W are
    24.                              defined by current locale
    25.  BOUNDU           no         Match "" at any boundary of a given type
    26.                              using /u rules.
    27.  BOUNDA           no         Match "" at any boundary between \w\W or
    28.                              \W\w, where \w is [_a-zA-Z0-9]
    29.  NBOUND           no         Like NBOUNDA for non-utf8, otherwise like
    30.                              BOUNDU
    31.  NBOUNDL          no         Like NBOUND/NBOUNDU, but \w and \W are
    32.                              defined by current locale
    33.  NBOUNDU          no         Match "" at any non-boundary of a given
    34.                              type using using /u rules.
    35.  NBOUNDA          no         Match "" betweeen any \w\w or \W\W, where
    36.                              \w is [_a-zA-Z0-9]
    38.  # [Special] alternatives:
    39.  REG_ANY          no         Match any one character (except newline).
    40.  SANY             no         Match any one character.
    41.  ANYOF            sv         Match character in (or not in) this class,
    42.                   charclass  single char match only
    43.  ANYOFD           sv         Like ANYOF, but /d is in effect
    44.                   charclass
    45.  ANYOFL           sv         Like ANYOF, but /l is in effect
    46.                   charclass
    47.  ANYOFPOSIXL      sv         Like ANYOFL, but matches [[:posix:]]
    48.                   charclass_ classes
    49.                   posixl
    51.  ANYOFH           sv 1       Like ANYOF, but only has "High" matches,
    52.                              none in the bitmap; the flags field
    53.                              contains the lowest matchable UTF-8 start
    54.                              byte
    55.  ANYOFHb          sv 1       Like ANYOFH, but all matches share the same
    56.                              UTF-8 start byte, given in the flags field
    57.  ANYOFHr          sv 1       Like ANYOFH, but the flags field contains
    58.                              packed bounds for all matchable UTF-8 start
    59.                              bytes.
    60.  ANYOFHs          sv 1       Like ANYOFHb, but has a string field that
    61.                              gives the leading matchable UTF-8 bytes;
    62.                              flags field is len
    63.  ANYOFR           packed 1   Matches any character in the range given by
    64.                              its packed args: upper 12 bits is the max
    65.                              delta from the base lower 20; the flags
    66.                              field contains the lowest matchable UTF-8
    67.                              start byte
    68.  ANYOFRb          packed 1   Like ANYOFR, but all matches share the same
    69.                              UTF-8 start byte, given in the flags field
    71.  ANYOFM           byte 1     Like ANYOF, but matches an invariant byte
    72.                              as determined by the mask and arg
    73.  NANYOFM          byte 1     complement of ANYOFM
    75.  # POSIX Character Classes:
    76.  POSIXD           none       Some [[:class:]] under /d; the FLAGS field
    77.                              gives which one
    78.  POSIXL           none       Some [[:class:]] under /l; the FLAGS field
    79.                              gives which one
    80.  POSIXU           none       Some [[:class:]] under /u; the FLAGS field
    81.                              gives which one
    82.  POSIXA           none       Some [[:class:]] under /a; the FLAGS field
    83.                              gives which one
    84.  NPOSIXD          none       complement of POSIXD, [[:^class:]]
    85.  NPOSIXL          none       complement of POSIXL, [[:^class:]]
    86.  NPOSIXU          none       complement of POSIXU, [[:^class:]]
    87.  NPOSIXA          none       complement of POSIXA, [[:^class:]]
    89.  CLUMP            no         Match any extended grapheme cluster
    90.                              sequence
    92.  # Alternation
    94.  # BRANCH        The set of branches constituting a single choice are
    95.  #               hooked together with their "next" pointers, since
    96.  #               precedence prevents anything being concatenated to
    97.  #               any individual branch.  The "next" pointer of the last
    98.  #               BRANCH in a choice points to the thing following the
    99.  #               whole choice.  This is also where the final "next"
    100.  #               pointer of each individual branch points; each branch
    101.  #               starts with the operand node of a BRANCH node.
    102.  #
    103.  BRANCH           node       Match this alternative, or the next...
    105.  # Literals
    107.  EXACT            str        Match this string (flags field is the
    108.                              length).
    110.  # In a long string node, the U32 argument is the length, and is
    111.  # immediately followed by the string.
    112.  LEXACT           len:str 1  Match this long string (preceded by length;
    113.                              flags unused).
    114.  EXACTL           str        Like EXACT, but /l is in effect (used so
    115.                              locale-related warnings can be checked for)
    116.  EXACTF           str        Like EXACT, but match using /id rules;
    117.                              (string not UTF-8, ASCII folded; non-ASCII
    118.                              not)
    119.  EXACTFL          str        Like EXACT, but match using /il rules;
    120.                              (string not likely to be folded)
    121.  EXACTFU          str        Like EXACT, but match using /iu rules;
    122.                              (string folded)
    124.  EXACTFAA         str        Like EXACT, but match using /iaa rules;
    125.                              (string folded except in non-UTF8 patterns:
    126.                              MICRO, SHARP S; folded length <= unfolded)
    128.  EXACTFUP         str        Like EXACT, but match using /iu rules;
    129.                              (string not UTF-8, folded except MICRO,
    130.                              SHARP S: hence Problematic)
    132.  EXACTFLU8        str        Like EXACTFU, but use /il, UTF-8, (string
    133.                              is folded, and everything in it is above
    134.                              255
    135.  EXACTFAA_NO_TRIE str        Like EXACT, but match using /iaa rules
    136.                              (string not UTF-8, not guaranteed to be
    137.                              folded, not currently trie-able)
    139.  EXACT_REQ8       str        Like EXACT, but only UTF-8 encoded targets
    140.                              can match
    141.  LEXACT_REQ8      len:str 1  Like LEXACT, but only UTF-8 encoded targets
    142.                              can match
    143.  EXACTFU_REQ8     str        Like EXACTFU, but only UTF-8 encoded
    144.                              targets can match
    146.  EXACTFU_S_EDGE   str        /di rules, but nothing in it precludes /ui,
    147.                              except begins and/or ends with [Ss];
    148.                              (string not UTF-8; compile-time only)
    150.  # Do nothing types
    152.  NOTHING          no         Match empty string.
    153.  # A variant of above which delimits a group, thus stops optimizations
    154.  TAIL             no         Match empty string. Can jump here from
    155.                              outside.
    157.  # Loops
    159.  # STAR,PLUS    '?', and complex '*' and '+', are implemented as
    160.  #               circular BRANCH structures.  Simple cases
    161.  #               (one character per match) are implemented with STAR
    162.  #               and PLUS for speed and to minimize recursive plunges.
    163.  #
    164.  STAR             node       Match this (simple) thing 0 or more times.
    165.  PLUS             node       Match this (simple) thing 1 or more times.
    167.  CURLY            sv 2       Match this simple thing {n,m} times.
    168.  CURLYN           no 2       Capture next-after-this simple thing
    169.  CURLYM           no 2       Capture this medium-complex thing {n,m}
    170.                              times.
    171.  CURLYX           sv 2       Match this complex thing {n,m} times.
    173.  # This terminator creates a loop structure for CURLYX
    174.  WHILEM           no         Do curly processing and see if rest
    175.                              matches.
    177.  # Buffer related
    179.  # OPEN,CLOSE,GROUPP     ...are numbered at compile time.
    180.  OPEN             num 1      Mark this point in input as start of #n.
    181.  CLOSE            num 1      Close corresponding OPEN of #n.
    182.  SROPEN           none       Same as OPEN, but for script run
    183.  SRCLOSE          none       Close preceding SROPEN
    185.  REF              num 1      Match some already matched string
    186.  REFF             num 1      Match already matched string, using /di
    187.                              rules.
    188.  REFFL            num 1      Match already matched string, using /li
    189.                              rules.
    190.  REFFU            num 1      Match already matched string, usng /ui.
    191.  REFFA            num 1      Match already matched string, using /aai
    192.                              rules.
    194.  # Named references.  Code in regcomp.c assumes that these all are after
    195.  # the numbered references
    196.  REFN             no-sv 1    Match some already matched string
    197.  REFFN            no-sv 1    Match already matched string, using /di
    198.                              rules.
    199.  REFFLN           no-sv 1    Match already matched string, using /li
    200.                              rules.
    201.  REFFUN           num 1      Match already matched string, using /ui
    202.                              rules.
    203.  REFFAN           num 1      Match already matched string, using /aai
    204.                              rules.
    206.  # Support for long RE
    207.  LONGJMP          off 1 1    Jump far away.
    208.  BRANCHJ          off 1 1    BRANCH with long offset.
    210.  # Special Case Regops
    211.  IFMATCH          off 1 1    Succeeds if the following matches; non-zero
    212.                              flags "f", next_off "o" means lookbehind
    213.                              assertion starting "f..(f-o)" characters
    214.                              before current
    215.  UNLESSM          off 1 1    Fails if the following matches; non-zero
    216.                              flags "f", next_off "o" means lookbehind
    217.                              assertion starting "f..(f-o)" characters
    218.                              before current
    219.  SUSPEND          off 1 1    "Independent" sub-RE.
    220.  IFTHEN           off 1 1    Switch, should be preceded by switcher.
    221.  GROUPP           num 1      Whether the group matched.
    223.  # The heavy worker
    225.  EVAL             evl/flags  Execute some Perl code.
    226.                   2L
    228.  # Modifiers
    230.  MINMOD           no         Next operator is not greedy.
    231.  LOGICAL          no         Next opcode should set the flag only.
    233.  # This is not used yet
    234.  RENUM            off 1 1    Group with independently numbered parens.
    236.  # Trie Related
    238.  # Behave the same as A|LIST|OF|WORDS would. The '..C' variants
    239.  # have inline charclass data (ascii only), the 'C' store it in the
    240.  # structure.
    242.  TRIE             trie 1     Match many EXACT(F[ALU]?)? at once.
    243.                              flags==type
    244.  TRIEC            trie       Same as TRIE, but with embedded charclass
    245.                   charclass  data
    247.  AHOCORASICK      trie 1     Aho Corasick stclass. flags==type
    248.  AHOCORASICKC     trie       Same as AHOCORASICK, but with embedded
    249.                   charclass  charclass data
    251.  # Regex Subroutines
    252.  GOSUB            num/ofs 2L recurse to paren arg1 at (signed) ofs arg2
    254.  # Special conditionals
    255.  GROUPPN          no-sv 1    Whether the group matched.
    256.  INSUBP           num 1      Whether we are in a specific recurse.
    257.  DEFINEP          none 1     Never execute directly.
    259.  # Backtracking Verbs
    260.  ENDLIKE          none       Used only for the type field of verbs
    261.  OPFAIL           no-sv 1    Same as (?!), but with verb arg
    262.  ACCEPT           no-sv/num  Accepts the current matched string, with
    263.                   2L         verbar
    265.  # Verbs With Arguments
    266.  VERB             no-sv 1    Used only for the type field of verbs
    267.  PRUNE            no-sv 1    Pattern fails at this startpoint if no-
    268.                              backtracking through this
    269.  MARKPOINT        no-sv 1    Push the current location for rollback by
    270.                              cut.
    271.  SKIP             no-sv 1    On failure skip forward (to the mark)
    272.                              before retrying
    273.  COMMIT           no-sv 1    Pattern fails outright if backtracking
    274.                              through this
    275.  CUTGROUP         no-sv 1    On failure go to the next alternation in
    276.                              the group
    278.  # Control what to keep in $&.
    279.  KEEPS            no         $& begins here.
    281.  # New charclass like patterns
    282.  LNBREAK          none       generic newline pattern
    284.  # SPECIAL  REGOPS
    286.  # This is not really a node, but an optimized away piece of a "long"
    287.  # node.  To simplify debugging output, we mark it as if it were a node
    288.  OPTIMIZED        off        Placeholder for dump.
    290.  # Special opcode with the property that no opcode in a compiled program
    291.  # will ever be of this type. Thus it can be used as a flag value that
    292.  # no other opcode has been seen. END is used similarly, in that an END
    293.  # node cant be optimized. So END implies "unoptimizable" and PSEUDO
    294.  # mean "not seen anything to optimize yet".
    295.  PSEUDO           off        Pseudo opcode for internal use.
    297.  REGEX_SET        depth p    Regex set, temporary node used in pre-
    298.                              optimization compilation

    Following the optimizer information is a dump of the offset/length table, here split across several lines:

    1.   Offsets: [45]
    2.   	1[4] 0[0] 0[0] 0[0] 0[0] 0[0] 0[0] 0[0] 0[0] 0[0] 0[0] 5[1]
    3.   	0[0] 12[1] 0[0] 6[1] 0[0] 7[1] 0[0] 9[1] 8[1] 0[0] 10[1] 0[0]
    4.   	11[1] 0[0] 12[0] 12[0] 13[1] 0[0] 14[4] 0[0] 0[0] 0[0] 0[0]
    5.   	0[0] 0[0] 0[0] 0[0] 0[0] 0[0] 18[1] 0[0] 19[1] 20[0]

    The first line here indicates that the offset/length table contains 45 entries. Each entry is a pair of integers, denoted by offset[length] . Entries are numbered starting with 1, so entry #1 here is 1[4] and entry #12 is 5[1] . 1[4] indicates that the node labeled 1: (the 1: ANYOF[bc]) begins at character position 1 in the pre-compiled form of the regex, and has a length of 4 characters. 5[1] in position 12 indicates that the node labeled 12: (the 12: EXACT <d>) begins at character position 5 in the pre-compiled form of the regex, and has a length of 1 character. 12[1] in position 14 indicates that the node labeled 14: (the 14: CURLYX[0] {1,32767}) begins at character position 12 in the pre-compiled form of the regex, and has a length of 1 character---that is, it corresponds to the + symbol in the precompiled regex.

    0[0] items indicate that there is no corresponding node.

    Run-time Output

    First of all, when doing a match, one may get no run-time output even if debugging is enabled. This means that the regex engine was never entered and that all of the job was therefore done by the optimizer.

    If the regex engine was entered, the output may look like this:

    1.   Matching '[bc]d(ef*g)+h[ij]k$' against 'abcdefg__gh__'
    2.     Setting an EVAL scope, savestack=3
    3.      2 <ab> <cdefg__gh_>    |  1: ANYOF
    4.      3 <abc> <defg__gh_>    | 11: EXACT <d>
    5.      4 <abcd> <efg__gh_>    | 13: CURLYX {1,32767}
    6.      4 <abcd> <efg__gh_>    | 26:   WHILEM
    7. 				0 out of 1..32767  cc=effff31c
    8.      4 <abcd> <efg__gh_>    | 15:     OPEN1
    9.      4 <abcd> <efg__gh_>    | 17:     EXACT <e>
    10.      5 <abcde> <fg__gh_>    | 19:     STAR
    11. 			     EXACT <f> can match 1 times out of 32767...
    12.     Setting an EVAL scope, savestack=3
    13.      6 <bcdef> <g__gh__>    | 22:       EXACT <g>
    14.      7 <bcdefg> <__gh__>    | 24:       CLOSE1
    15.      7 <bcdefg> <__gh__>    | 26:       WHILEM
    16. 				    1 out of 1..32767  cc=effff31c
    17.     Setting an EVAL scope, savestack=12
    18.      7 <bcdefg> <__gh__>    | 15:         OPEN1
    19.      7 <bcdefg> <__gh__>    | 17:         EXACT <e>
    20.        restoring \1 to 4(4)..7
    21. 				    failed, try continuation...
    22.      7 <bcdefg> <__gh__>    | 27:         NOTHING
    23.      7 <bcdefg> <__gh__>    | 28:         EXACT <h>
    24. 				    failed...
    25. 				failed...

    The most significant information in the output is about the particular node of the compiled regex that is currently being tested against the target string. The format of these lines is

    STRING-OFFSET <PRE-STRING> <POST-STRING> |ID: TYPE

    The TYPE info is indented with respect to the backtracking level. Other incidental information appears interspersed within.

    Debugging Perl Memory Usage

    Perl is a profligate wastrel when it comes to memory use. There is a saying that to estimate memory usage of Perl, assume a reasonable algorithm for memory allocation, multiply that estimate by 10, and while you still may miss the mark, at least you won't be quite so astonished. This is not absolutely true, but may provide a good grasp of what happens.

    Assume that an integer cannot take less than 20 bytes of memory, a float cannot take less than 24 bytes, a string cannot take less than 32 bytes (all these examples assume 32-bit architectures, the result are quite a bit worse on 64-bit architectures). If a variable is accessed in two of three different ways (which require an integer, a float, or a string), the memory footprint may increase yet another 20 bytes. A sloppy malloc(3) implementation can inflate these numbers dramatically.

    On the opposite end of the scale, a declaration like

    1.   sub foo;

    may take up to 500 bytes of memory, depending on which release of Perl you're running.

    Anecdotal estimates of source-to-compiled code bloat suggest an eightfold increase. This means that the compiled form of reasonable (normally commented, properly indented etc.) code will take about eight times more space in memory than the code took on disk.

    The -DL command-line switch is obsolete since circa Perl 5.6.0 (it was available only if Perl was built with -DDEBUGGING ). The switch was used to track Perl's memory allocations and possible memory leaks. These days the use of malloc debugging tools like Purify or valgrind is suggested instead. See also PERL_MEM_LOG in perlhacktips.

    One way to find out how much memory is being used by Perl data structures is to install the Devel::Size module from CPAN: it gives you the minimum number of bytes required to store a particular data structure. Please be mindful of the difference between the size() and total_size().

    If Perl has been compiled using Perl's malloc you can analyze Perl memory usage by setting $ENV{PERL_DEBUG_MSTATS}.

    Using $ENV{PERL_DEBUG_MSTATS}

    If your perl is using Perl's malloc() and was compiled with the necessary switches (this is the default), then it will print memory usage statistics after compiling your code when $ENV{PERL_DEBUG_MSTATS} > 1 , and before termination of the program when $ENV{PERL_DEBUG_MSTATS} >= 1 . The report format is similar to the following example:

    1.  $ PERL_DEBUG_MSTATS=2 perl -e "require Carp"
    2.  Memory allocation statistics after compilation: (buckets 4(4)..8188(8192)
    3.     14216 free:   130   117    28     7     9   0   2     2   1 0 0
    4. 		437    61    36     0     5
    5.     60924 used:   125   137   161    55     7   8   6    16   2 0 1
    6. 		 74   109   304    84    20
    7.  Total sbrk(): 77824/21:119. Odd ends: pad+heads+chain+tail: 0+636+0+2048.
    8.  Memory allocation statistics after execution:   (buckets 4(4)..8188(8192)
    9.     30888 free:   245    78    85    13     6   2   1     3   2 0 1
    10. 		315   162    39    42    11
    11.    175816 used:   265   176  1112   111    26  22  11    27   2 1 1
    12. 		196   178  1066   798    39
    13.  Total sbrk(): 215040/47:145. Odd ends: pad+heads+chain+tail: 0+2192+0+6144.

    It is possible to ask for such a statistic at arbitrary points in your execution using the mstat() function out of the standard Devel::Peek module.

    Here is some explanation of that format:

    • buckets SMALLEST(APPROX)..GREATEST(APPROX)

      Perl's malloc() uses bucketed allocations. Every request is rounded up to the closest bucket size available, and a bucket is taken from the pool of buckets of that size.

      The line above describes the limits of buckets currently in use. Each bucket has two sizes: memory footprint and the maximal size of user data that can fit into this bucket. Suppose in the above example that the smallest bucket were size 4. The biggest bucket would have usable size 8188, and the memory footprint would be 8192.

      In a Perl built for debugging, some buckets may have negative usable size. This means that these buckets cannot (and will not) be used. For larger buckets, the memory footprint may be one page greater than a power of 2. If so, the corresponding power of two is printed in the APPROX field above.

    • Free/Used

      The 1 or 2 rows of numbers following that correspond to the number of buckets of each size between SMALLEST and GREATEST . In the first row, the sizes (memory footprints) of buckets are powers of two--or possibly one page greater. In the second row, if present, the memory footprints of the buckets are between the memory footprints of two buckets "above".

      For example, suppose under the previous example, the memory footprints were

      1.    free:    8     16    32    64    128  256 512 1024 2048 4096 8192
      2. 	   4     12    24    48    80

      With a non-DEBUGGING perl, the buckets starting from 128 have a 4-byte overhead, and thus an 8192-long bucket may take up to 8188-byte allocations.

    • Total sbrk(): SBRKed/SBRKs:CONTINUOUS

      The first two fields give the total amount of memory perl sbrk(2)ed (ess-broken? :-) and number of sbrk(2)s used. The third number is what perl thinks about continuity of returned chunks. So long as this number is positive, malloc() will assume that it is probable that sbrk(2) will provide continuous memory.

      Memory allocated by external libraries is not counted.

    • pad: 0

      The amount of sbrk(2)ed memory needed to keep buckets aligned.

    • heads: 2192

      Although memory overhead of bigger buckets is kept inside the bucket, for smaller buckets, it is kept in separate areas. This field gives the total size of these areas.

    • chain: 0

      malloc() may want to subdivide a bigger bucket into smaller buckets. If only a part of the deceased bucket is left unsubdivided, the rest is kept as an element of a linked list. This field gives the total size of these chunks.

    • tail: 6144

      To minimize the number of sbrk(2)s, malloc() asks for more memory. This field gives the size of the yet unused part, which is sbrk(2)ed, but never touched.

    SEE ALSO

    perldebug, perlguts, perlrun re, and Devel::DProf.