perldebguts - Guts of Perl debugging

This is not the perldebug(1) manpage, 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.

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 CWcaller function from the package CWDB, the arguments that the corresponding stack frame was called with are copied to the CW@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 CW$ENV{PERL5DB} (or CWBEGIN {require 'perl5db.pl'} if not present) before the first line of your program.

•

Each array CW@{"_<$filename"} holds the lines of CW$filename for a file compiled by Perl. The same is also true for CWevaled strings that contain subroutines, or which are currently being executed. The CW$filename for CWevaled strings looks like CW(eval 34). Code assertions in regexes look like CW(re_eval 19). Values in this array are magical in numeric context: they compare equal to zero only if the line is not breakable.

•

Each hash CW%{"_<$filename"} contains breakpoints and actions keyed by line number. Individual entries (as opposed to the whole hash) are settable. Perl only cares about Boolean true here, although the values used by perl5db.pl have the form CW"$break_condition\0$action". The same holds for evaluated strings that contain subroutines, or which are currently being executed. The CW$filename for CWevaled strings looks like CW(eval 34) or CW(re_eval 19).

•

Each scalar CW${"_<$filename"} contains CW"_<$filename". This is also the case for evaluated strings that contain subroutines, or which are currently being executed. The CW$filename for CWevaled strings looks like CW(eval 34) or CW(re_eval 19).

•

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

•

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

•

A hash CW%DB::sub is maintained, whose keys are subroutine names and whose values have the form CWfilename:startline-endline. CWfilename has the form CW(eval 34) for subroutines defined inside CWevals, or CW(re_eval 19) for those within regex code assertions.

•

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

•

When execution of the program reaches a subroutine call, a call to CW&DB::sub(args) is made instead, with CW$DB::sub holding the name of the called subroutine. (This doesn't happen if the subroutine was compiled in the CWDB package.)

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

Environment Variables

The CWPERL5DB 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:

  sub DB::DB {}

It can easily be defined like this:

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

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

  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:

  {
    package DB;
    sub DB  {}
    sub sub {print ++$i, " $sub\n"; &$sub}
  }

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

When it starts, the debugger reads your rc file (./.perldb or ~/.perldb under Unix), which can set important options. (A subroutine (CW&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 CWo ... debugger command (q.v. in 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.

•

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

•

CW%DB::dbline, is an alias for CW%{"::_<current_file"}, which contains breakpoints and actions keyed by line number in the currently-selected file, either explicitly chosen with the debugger's CWf 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 CW"$break_condition\0$action".

Debugger customization functions

Some functions are provided to simplify customization.

•

See Options in perldebug for description of options parsed by CWDB::parse_options(string) parses debugger options; see Options in pperldebug for a description of options recognized.

•

CWDB::dump_trace(skip[,count]) skips the specified number of frames and returns a list containing information about the calling frames (all of them, if CWcount is missing). Each entry is reference to a hash with keys CWcontext (either CW., CW$, or CW@), CWsub (subroutine name, or info about CWeval), CWargs (CWundef or a reference to an array), CWfile, and CWline.

•

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

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

 $ perl -de 42
 Stack dump during die enabled outside of evals.

 Loading DB routines from perl5db.pl patch level 0.94
 Emacs support available.

 Enter h or `h h' for help.

 main::(-e:1):   0
   DB<1> sub foo { 14 }

   DB<2> sub bar { 3 }

   DB<3> t print foo() * bar()
 main::((eval 172):3):   print foo() + bar();
 main::foo((eval 168):2):
 main::bar((eval 170):2):
 42

with this one, once the CWoption CWframe=2 has been set:

   DB<4> o f=2
                frame = '2'
   DB<5> t print foo() * bar()
 3:      foo() * bar()
 entering main::foo
  2:     sub foo { 14 };
 exited main::foo
 entering main::bar
  2:     sub bar { 3 };
 exited main::bar
 42

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

1

  entering main::BEGIN
   entering Config::BEGIN
    Package lib/Exporter.pm.
    Package lib/Carp.pm.
   Package lib/Config.pm.
   entering Config::TIEHASH
   entering Exporter::import
    entering Exporter::export
  entering Config::myconfig
   entering Config::FETCH
   entering Config::FETCH
   entering Config::FETCH
   entering Config::FETCH

2

  entering main::BEGIN
   entering Config::BEGIN
    Package lib/Exporter.pm.
    Package lib/Carp.pm.
   exited Config::BEGIN
   Package lib/Config.pm.
   entering Config::TIEHASH
   exited Config::TIEHASH
   entering Exporter::import
    entering Exporter::export
    exited Exporter::export
   exited Exporter::import
  exited main::BEGIN
  entering Config::myconfig
   entering Config::FETCH
   exited Config::FETCH
   entering Config::FETCH
   exited Config::FETCH
   entering Config::FETCH

4

  in  $=main::BEGIN() from /dev/null:0
   in  $=Config::BEGIN() from lib/Config.pm:2
    Package lib/Exporter.pm.
    Package lib/Carp.pm.
   Package lib/Config.pm.
   in  $=Config::TIEHASH('Config') from lib/Config.pm:644
   in  $=Exporter::import('Config', 'myconfig', 'config_vars') from /dev/null:0
    in  $=Exporter::export('Config', 'main', 'myconfig', 'config_vars') from li
  in  @=Config::myconfig() from /dev/null:0
   in  $=Config::FETCH(ref(Config), 'package') from lib/Config.pm:574
   in  $=Config::FETCH(ref(Config), 'baserev') from lib/Config.pm:574
   in  $=Config::FETCH(ref(Config), 'PERL_VERSION') from lib/Config.pm:574
   in  $=Config::FETCH(ref(Config), 'PERL_SUBVERSION') from lib/Config.pm:574
   in  $=Config::FETCH(ref(Config), 'osname') from lib/Config.pm:574
   in  $=Config::FETCH(ref(Config), 'osvers') from lib/Config.pm:574

6

  in  $=main::BEGIN() from /dev/null:0
   in  $=Config::BEGIN() from lib/Config.pm:2
    Package lib/Exporter.pm.
    Package lib/Carp.pm.
   out $=Config::BEGIN() from lib/Config.pm:0
   Package lib/Config.pm.
   in  $=Config::TIEHASH('Config') from lib/Config.pm:644
   out $=Config::TIEHASH('Config') from lib/Config.pm:644
   in  $=Exporter::import('Config', 'myconfig', 'config_vars') from /dev/null:0
    in  $=Exporter::export('Config', 'main', 'myconfig', 'config_vars') from lib/
    out $=Exporter::export('Config', 'main', 'myconfig', 'config_vars') from lib/
   out $=Exporter::import('Config', 'myconfig', 'config_vars') from /dev/null:0
  out $=main::BEGIN() from /dev/null:0
  in  @=Config::myconfig() from /dev/null:0
   in  $=Config::FETCH(ref(Config), 'package') from lib/Config.pm:574
   out $=Config::FETCH(ref(Config), 'package') from lib/Config.pm:574
   in  $=Config::FETCH(ref(Config), 'baserev') from lib/Config.pm:574
   out $=Config::FETCH(ref(Config), 'baserev') from lib/Config.pm:574
   in  $=Config::FETCH(ref(Config), 'PERL_VERSION') from lib/Config.pm:574
   out $=Config::FETCH(ref(Config), 'PERL_VERSION') from lib/Config.pm:574
   in  $=Config::FETCH(ref(Config), 'PERL_SUBVERSION') from lib/Config.pm:574

14

  in  $=main::BEGIN() from /dev/null:0
   in  $=Config::BEGIN() from lib/Config.pm:2
    Package lib/Exporter.pm.
    Package lib/Carp.pm.
   out $=Config::BEGIN() from lib/Config.pm:0
   Package lib/Config.pm.
   in  $=Config::TIEHASH('Config') from lib/Config.pm:644
   out $=Config::TIEHASH('Config') from lib/Config.pm:644
   in  $=Exporter::import('Config', 'myconfig', 'config_vars') from /dev/null:0
    in  $=Exporter::export('Config', 'main', 'myconfig', 'config_vars') from lib/E
    out $=Exporter::export('Config', 'main', 'myconfig', 'config_vars') from lib/E
   out $=Exporter::import('Config', 'myconfig', 'config_vars') from /dev/null:0
  out $=main::BEGIN() from /dev/null:0
  in  @=Config::myconfig() from /dev/null:0
   in  $=Config::FETCH('Config=HASH(0)', 'package') from lib/Config.pm:574
   out $=Config::FETCH('Config=HASH(0)', 'package') from lib/Config.pm:574
   in  $=Config::FETCH('Config=HASH(0)', 'baserev') from lib/Config.pm:574
   out $=Config::FETCH('Config=HASH(0)', 'baserev') from lib/Config.pm:574

30

  in  $=CODE(0)() from /dev/null:0
   in  $=CODE(0)() from lib/Config.pm:2
    Package lib/Exporter.pm.
   out $=CODE(0)() from lib/Config.pm:0
   scalar context return from CODE(0): undef
   Package lib/Config.pm.
   in  $=Config::TIEHASH('Config') from lib/Config.pm:628
   out $=Config::TIEHASH('Config') from lib/Config.pm:628
   scalar context return from Config::TIEHASH:   empty hash
   in  $=Exporter::import('Config', 'myconfig', 'config_vars') from /dev/null:0
    in  $=Exporter::export('Config', 'main', 'myconfig', 'config_vars') from lib/Exporter.pm:171
    out $=Exporter::export('Config', 'main', 'myconfig', 'config_vars') from lib/Exporter.pm:171
    scalar context return from Exporter::export: ''
   out $=Exporter::import('Config', 'myconfig', 'config_vars') from /dev/null:0
   scalar context return from Exporter::import: ''

In all cases shown above, the line indentation shows the call tree. If bit 2 of CWframe 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

    Package lib/Carp.pm.

is printed with proper indentation.

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

If your perl is compiled with CW-DDEBUGGING, you may use the -Dr flag on the command line.

Otherwise, one can CWuse re 'debug', which has effects at compile time and run time. It is not lexically scoped.

The debugging output at compile time looks like this:

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

The

  anchored `de' at 1 floating `gh' at 3..2147483647 (checking floating) 
        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 CWde at offset 1, plus substring CWgh at some offset between 3 and infinity. Moreover, when checking for these substrings (to abandon impossible matches quickly), Perl will check for the substring CWgh before checking for the substring CWde. The optimizer may also use the knowledge that the match starts (at the CWfirst 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 See above. Which substring to check first. The minimal length of the match. Type of first matching node. Don't scan for the found substrings. 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. Set if the pattern contains CW\G. Set if the pattern starts with a repeated char (as in CWx+y). Set if the pattern starts with CW.*. Set if the pattern contain eval-groups, such as CW(?{ code }) and CW(??{ code }). If the pattern may match only at a handful of places, (with CWTYPE being CWBOL, CWMBOL, or CWGPOS. See the table below.

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

The optimizer-specific information is used to avoid entering (a slow) regex engine on strings that will not definitely match. If the CWisall 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

CW id: TYPE OPTIONAL-INFO (next-id)

Here are the possible types, with short descriptions:

    # TYPE arg-description [num-args] [longjump-len] DESCRIPTION

    # Exit points
    END         no      End of program.
    SUCCEED     no      Return from a subroutine, basically.

    # Anchors:
    BOL         no      Match "" at beginning of line.
    MBOL        no      Same, assuming multiline.
    SBOL        no      Same, assuming singleline.
    EOS         no      Match "" at end of string.
    EOL         no      Match "" at end of line.
    MEOL        no      Same, assuming multiline.
    SEOL        no      Same, assuming singleline.
    BOUND       no      Match "" at any word boundary
    BOUNDL      no      Match "" at any word boundary
    NBOUND      no      Match "" at any word non-boundary
    NBOUNDL     no      Match "" at any word non-boundary
    GPOS        no      Matches where last m//g left off.

    # [Special] alternatives
    ANY         no      Match any one character (except newline).
    SANY        no      Match any one character.
    ANYOF       sv      Match character in (or not in) this class.
    ALNUM       no      Match any alphanumeric character
    ALNUML      no      Match any alphanumeric char in locale
    NALNUM      no      Match any non-alphanumeric character
    NALNUML     no      Match any non-alphanumeric char in locale
    SPACE       no      Match any whitespace character
    SPACEL      no      Match any whitespace char in locale
    NSPACE      no      Match any non-whitespace character
    NSPACEL     no      Match any non-whitespace char in locale
    DIGIT       no      Match any numeric character
    NDIGIT      no      Match any non-numeric character

    # BRANCH    The set of branches constituting a single choice are hooked
    #           together with their "next" pointers, since precedence prevents
    #           anything being concatenated to any individual branch.  The
    #           "next" pointer of the last BRANCH in a choice points to the
    #           thing following the whole choice.  This is also where the
    #           final "next" pointer of each individual branch points; each
    #           branch starts with the operand node of a BRANCH node.
    #
    BRANCH      node    Match this alternative, or the next...

    # BACK      Normal "next" pointers all implicitly point forward; BACK
    #           exists to make loop structures possible.
    # not used
    BACK        no      Match "", "next" ptr points backward.

    # Literals
    EXACT       sv      Match this string (preceded by length).
    EXACTF      sv      Match this string, folded (prec. by length).
    EXACTFL     sv      Match this string, folded in locale (w/len).

    # Do nothing
    NOTHING     no      Match empty string.
    # A variant of above which delimits a group, thus stops optimizations
    TAIL        no      Match empty string. Can jump here from outside.

    # STAR,PLUS '?', and complex '*' and '+', are implemented as circular
    #           BRANCH structures using BACK.  Simple cases (one character
    #           per match) are implemented with STAR and PLUS for speed
    #           and to minimize recursive plunges.
    #
    STAR        node    Match this (simple) thing 0 or more times.
    PLUS        node    Match this (simple) thing 1 or more times.

    CURLY       sv 2    Match this simple thing {n,m} times.
    CURLYN      no 2    Match next-after-this simple thing 
    #                   {n,m} times, set parens.
    CURLYM      no 2    Match this medium-complex thing {n,m} times.
    CURLYX      sv 2    Match this complex thing {n,m} times.

    # This terminator creates a loop structure for CURLYX
    WHILEM      no      Do curly processing and see if rest matches.

    # OPEN,CLOSE,GROUPP ...are numbered at compile time.
    OPEN        num 1   Mark this point in input as start of #n.
    CLOSE       num 1   Analogous to OPEN.

    REF         num 1   Match some already matched string
    REFF        num 1   Match already matched string, folded
    REFFL       num 1   Match already matched string, folded in loc.

    # grouping assertions
    IFMATCH     off 1 2 Succeeds if the following matches.
    UNLESSM     off 1 2 Fails if the following matches.
    SUSPEND     off 1 1 "Independent" sub-regex.
    IFTHEN      off 1 1 Switch, should be preceded by switcher .
    GROUPP      num 1   Whether the group matched.

    # Support for long regex
    LONGJMP     off 1 1 Jump far away.
    BRANCHJ     off 1 1 BRANCH with long offset.

    # The heavy worker
    EVAL        evl 1   Execute some Perl code.

    # Modifiers
    MINMOD      no      Next operator is not greedy.
    LOGICAL     no      Next opcode should set the flag only.

    # This is not used yet
    RENUM       off 1 1 Group with independently numbered parens.

    # This is not really a node, but an optimized away piece of a "long" node.
    # To simplify debugging output, we mark it as if it were a node
    OPTIMIZED   off     Placeholder for dump.

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

  Offsets: [45]
        1[4] 0[0] 0[0] 0[0] 0[0] 0[0] 0[0] 0[0] 0[0] 0[0] 0[0] 5[1]
        0[0] 12[1] 0[0] 6[1] 0[0] 7[1] 0[0] 9[1] 8[1] 0[0] 10[1] 0[0]
        11[1] 0[0] 12[0] 12[0] 13[1] 0[0] 14[4] 0[0] 0[0] 0[0] 0[0]
        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 CWoffset[length]. Entries are numbered starting with 1, so entry #1 here is CW1[4] and entry #12 is CW5[1]. CW1[4] indicates that the node labeled CW1: (the CW1: ANYOF[bc]) begins at character position 1 in the pre-compiled form of the regex, and has a length of 4 characters. CW5[1] in position 12 indicates that the node labeled CW12: (the CW12: EXACT <d>) begins at character position 5 in the pre-compiled form of the regex, and has a length of 1 character. CW12[1] in position 14 indicates that the node labeled CW14: (the CW14: 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 CW+ symbol in the precompiled regex.

CW0[0] items indicate that there is no corresponding node.

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:

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

CW 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.

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

  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 CW-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.

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

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

Free/Used

The 1 or 2 rows of numbers following that correspond to the number of buckets of each size between CWSMALLEST and CWGREATEST. In the first row, the sizes (memory footprints) of buckets are powers of twoor 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

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

With non-CWDEBUGGING perl, the buckets starting from CW128 have a 4-byte overhead, and thus an 8192-long bucket may take up to 8188-byte allocations. 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. The amount of sbrk(2)ed memory needed to keep buckets aligned. 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. 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. 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.

(Note that -DL is obsolete since circa 5.6.0, and even before that Perl needed to be compiled with -DDEBUGGING.)

Below we show how to analyse memory usage by

  do 'lib/auto/POSIX/autosplit.ix';

The file in question contains a header and 146 lines similar to

  sub getcwd;

WARNING: The discussion below supposes 32-bit architecture. In newer releases of Perl, memory usage of the constructs discussed here is greatly improved, but the story discussed below is a real-life story. This story is mercilessly terse, and assumes rather more than cursory knowledge of Perl internals. Type space to continue, `q' to quit. (Actually, you just want to skip to the next section.)

Here is the itemized list of Perl allocations performed during parsing of this file:

 !!! "after" at test.pl line 3.
    Id  subtot   4   8  12  16  20  24  28  32  36  40  48  56  64  72  80 80+
  0 02   13752   .   .   .   . 294   .   .   .   .   .   .   .   .   .   .   4
  0 54    5545   .   .   8 124  16   .   .   .   1   1   .   .   .   .   .   3
  5 05      32   .   .   .   .   .   .   .   1   .   .   .   .   .   .   .   .
  6 02    7152   .   .   .   .   .   .   .   .   .   . 149   .   .   .   .   .
  7 02    3600   .   .   .   .   . 150   .   .   .   .   .   .   .   .   .   .
  7 03      64   .  -1   .   1   .   .   2   .   .   .   .   .   .   .   .   .
  7 04    7056   .   .   .   .   .   .   .   .   .   .   .   .   .   .   .   7
  7 17   38404   .   .   .   .   .   .   .   1   .   . 442 149   .   . 147   .
  9 03    2078  17 249  32   .   .   .   .   2   .   .   .   .   .   .   .   .

To see this list, insert two CWwarn('!...') statements around the call:

  warn('!');
  do 'lib/auto/POSIX/autosplit.ix';
  warn('!!! "after"');

and run it with Perl's -DL option. The first warn() will print memory allocation info before parsing the file and will memorize the statistics at this point (we ignore what it prints). The second warn() prints increments with respect to these memorized data. This is the printout shown above.

Different Ids on the left correspond to different subsystems of the perl interpreter. They are just the first argument given to the perl memory allocation API named New(). To find what CW9 03 means, just grep the perl source for CW903. You'll find it in util.c, function savepvn(). (I know, you wonder why we told you to grep and then gave away the answer. That's because grepping the source is good for the soul.) This function is used to store a copy of an existing chunk of memory. Using a C debugger, one can see that the function was called either directly from gv_init() or via sv_magic(), and that gv_init() is called from gv_fetchpv()--which was itself called from newSUB(). Please stop to catch your breath now.

NOTE: To reach this point in the debugger and skip the calls to savepvn() during the compilation of the main program, you should set a C breakpoint in Perl_warn(), continue until this point is reached, and then set a C breakpoint in Perl_savepvn(). Note that you may need to skip a handful of Perl_savepvn() calls that do not correspond to mass production of CVs (there are more CW903 allocations than 146 similar lines of lib/auto/POSIX/autosplit.ix). Note also that CWPerl_ prefixes are added by macroization code in perl header files to avoid conflicts with external libraries.

Anyway, we see that CW903 ids correspond to creation of globs, twice per glob - for glob name, and glob stringification magic.

Here are explanations for other Ids above: Creates bigger CWXPV* structures. In the case above, it creates 3 CWAVs per subroutine, one for a list of lexical variable names, one for a scratchpad (which contains lexical variables and CWtargets), and one for the array of scratchpads needed for recursion. It also creates a CWGV and a CWCV per subroutine, all called from start_subparse(). Creates a C array corresponding to the CWAV of scratchpads and the scratchpad itself. The first fake entry of this scratchpad is created though the subroutine itself is not defined yet. It also creates C arrays to keep data for the stash. This is one HV, but it grows; thus, there are 4 big allocations: the big chunks are not freed, but are kept as additional arenas for CWSV allocations. Creates a CWHEK for the name of the glob for the subroutine. This name is a key in a stash. Big allocations with this Id correspond to allocations of new arenas to keep CWHE. Creates a CWGP for the glob for the subroutine. Creates the CWMAGIC for the glob for the subroutine. Creates arenas which keep SVs.

If Perl is run with -DL option, then warn()s that start with `!' behave specially. They print a list of categories of memory allocations, and statistics of allocations of different sizes for these categories.

If warn() string starts with print changed categories only, print the differences in counts of allocations. print grown categories only; print the absolute values of counts, and totals. print nonempty categories, print the absolute values of counts and totals.

If an extension or external library does not use the Perl API to allocate memory, such allocations are not counted.

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