perldebug - Perl debugging
First of all, have you tried using the -w switch?
``As soon as we started programming, we found to our surprise that it wasn't as easy to get programs right as we had thought. Debugging had to be discovered. I can remember the exact instant when I realized that a large part of my life from then on was going to be spent in finding mistakes in my own programs.''
--Maurice Wilkes, 1949
If you invoke Perl with the -d switch, your script runs under the Perl source debugger. This works like an interactive Perl environment, prompting for debugger commands that let you examine source code, set breakpoints, get stack backtraces, change the values of variables, etc. This is so convenient that you often fire up the debugger all by itself just to test out Perl constructs interactively to see what they do. For example:
perl -d -e 42
In Perl, the debugger is not a separate program as it usually is in the typical compiled environment. Instead, the -d flag tells the compiler to insert source information into the parse trees it's about to hand off to the interpreter. That means your code must first compile correctly for the debugger to work on it. Then when the interpreter starts up, it preloads a Perl library file containing the debugger itself.
The program will halt right before the first run-time executable statement (but see below regarding compile-time statements) and ask you to enter a debugger command. Contrary to popular expectations, whenever the debugger halts and shows you a line of code, it always displays the line it's about to execute, rather than the one it has just executed.
Any command not recognized by the debugger is directly executed (eval'd) as Perl code in the current package. (The debugger uses the DB package for its own state information.)
Leading white space before a command would cause the debugger to think it's NOT a debugger command but for Perl, so be careful not to do that.
The debugger understands the following commands:
If you supply another debugger command as an argument to the h command, it prints out the description for just that command. The special
argument of h h produces a more compact help listing, designed to fit together on one
screen.
If the output of the h command (or any command, for that matter) scrolls past your screen, either
precede the command with a leading pipe symbol so it's run through your
pager, as in
DB> |h
You may change the pager which is used via O pager=... command.
print {$DB::OUT} expr in the current package. In particular, because this is just Perl's own print function, this means that nested data structures and objects are not
dumped, unlike with the x command.
The DB::OUT filehandle is opened to /dev/tty, regardless of where
STDOUT may be redirected to.
The details of printout are governed by multiple Options.
main
package) using a data pretty-printer (hashes show their keys and values so
you see what's what, control characters are made printable, etc.). Make
sure you don't put the type specifier (like $) there, just the symbol names, like this:
V DB filename line
Use ~pattern and !pattern for positive and negative regexps.
Nested data structures are printed out in a legible fashion, unlike the print function.
The details of printout are governed by multiple Options.
V currentpackage [vars].
n or s command.
incr+1 lines starting at min.
min through max. l - is synonymous to -.
filename
is not a full filename as found in values of
%INC, it is considered as a regexp.
AutoTrace Option).
$ 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
or, with the Option frame=2 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
b 237 $x > 30
b 237 ++$count237 < 11
b 33 /pattern/i
1. check for a breakpoint at this line 2. print the line if necessary (tracing) 3. do any actions associated with that line 4. prompt user if at a breakpoint or in single-step 5. evaluate line
For example, this will print out $foo every time line 53 is
passed:
a 53 print "DB FOUND $foo\n"
!.
| character.) By default,
$ENV{PAGER} will be used.
To disable this behaviour, set these values to 0. If dieLevel is 2, then the messages which will be caught by surrounding eval are also printed.
t command, but can be put into
PERLDB_OPTS).
|visual_perl_db), then a short, ``emacs like'' message is used.
r command.
frame & 2 is false, messages are printed on entry only. (Printing on exit may be useful if
inter(di)spersed with other messages.)
If frame & 4, arguments to functions are printed as well as the context and caller
info. If frame & 8, overloaded stringify and
tied FETCH are enabled on the printed arguments. If frame &
16, the return value from the subroutine is printed as well.
The length at which the argument list is truncated is governed by the next option:
frame option's bit 4 is set.
The following options affect what happens with V, X, and x
commands:
compactDump, short array may be printed on one line.
quote is auto, one can enable either double-quotish dump, or single-quotish by setting
it to " or '. By default, characters with high bit set are printed
as is.
During startup options are initialized from $ENV{PERLDB_OPTS}. You can put additional initialization options TTY, noTTY,
ReadLine, and NonStop there.
Example rc file:
&parse_options("NonStop=1 LineInfo=db.out AutoTrace");
The script will run without human intervention, putting trace information
into the file db.out. (If you interrupt it, you would better reset
LineInfo to something ``interactive''!)
NonStop mode, and would not connect to a
TTY. If interrupt (or if control goes to debugger via explicit setting of $DB::signal or $DB::single from the Perl script), connects to a
TTY specified by the
TTY option at startup, or to a
TTY found at runtime using Term::Rendezvous module of your choice.
This module should implement a method new which returns an object with two methods: IN and OUT, returning two filehandles to use for debugging input and output
correspondingly. Method new may inspect an argument which is a value of $ENV{PERLDB_NOTTY} at startup, or is "/tmp/perldbtty$$" otherwise.
Here's an example of using the $ENV{PERLDB_OPTS} variable:
$ PERLDB_OPTS="N f=2" perl -d myprogram
will run the script myprogram without human intervention, printing out the call tree with entry and exit
points. Note that N f=2 is equivalent to NonStop=1 frame=2. Note also that at the moment when this documentation was written all the
options to the debugger could be uniquely abbreviated by the first letter
(with exception of
Dump* options).
Other examples may include
$ PERLDB_OPTS="N f A L=listing" perl -d myprogram
- runs script noninteractively, printing info on each entry into a
subroutine and each executed line into the file listing. (If you interrupt it, you would better reset LineInfo to something ``interactive''!)
$ env "PERLDB_OPTS=R=0 TTY=/dev/ttyc" perl -d myprogram
may be useful for debugging a program which uses Term::ReadLine
itself. Do not forget detach shell from the
TTY in the window which corresponds to /dev/ttyc, say, by issuing a command like
$ sleep 1000000
See Debugger Internals below for more details.
command is missing, resets the list of actions.
command is missing, resets the list of actions.
command is missing, resets the list of actions.
O recallCommand, too.
O shellBang too.
Set an Option inhibit_exit to 0 if you want to be able to step
off the end the script. You may also need to set $finished to 0 at some moment if you want to step through global destruction.
Currently the following setting are preserved: history, breakpoints,
actions, debugger Options, and the following command line options: -w, -I, and -e.
|dbcmd but
DB::OUT is temporarily selected as well. Often used with commands that would otherwise produce long
output, such as
|V main
= quit q
or list current aliases.
DB<8>
or even
DB<<17>>
where that number is the command number, which you'd use to access with the
builtin csh-like history mechanism, e.g., !17 would repeat command number 17. The number of angle brackets indicates the
depth of the debugger. You could get more than one set of brackets, for
example, if you'd already at a breakpoint and then printed out the result
of a function call that itself also has a breakpoint, or you step into an
expression via s/n/t expression command.
DB<1> for (1..4) { \
cont: print "ok\n"; \
cont: }
ok
ok
ok
ok
Note that this business of escaping a newline is specific to interactive commands typed into the debugger.
T command might look like:
$ = main::infested called from file `Ambulation.pm' line 10
@ = Ambulation::legs(1, 2, 3, 4) called from file `camel_flea' line 7
$ = main::pests('bactrian', 4) called from file `camel_flea' line 4
The left-hand character up there tells whether the function was called in a
scalar or list context (we bet you can tell which is which). What that says
is that you were in the function main::infested when you ran the stack dump, and that it was called in a scalar context
from line 10 of the file Ambulation.pm, but without any arguments at all, meaning it was called as &infested. The next stack frame shows that the function Ambulation::legs was called in a list context from the
camel_flea file with four arguments. The last stack frame shows that
main::pests was called in a scalar context, also from camel_flea, but from line 4.
Note that if you execute T command from inside an active use
statement, the backtrace will contain both require
frame and an eval) frame.
l command looks like this:
DB<<13>> l
101: @i{@i} = ();
102:b @isa{@i,$pack} = ()
103 if(exists $i{$prevpack} || exists $isa{$pack});
104 }
105
106 next
107==> if(exists $isa{$pack});
108
109:a if ($extra-- > 0) {
110: %isa = ($pack,1);
Note that the breakable lines are marked with :, lines with breakpoints are marked by b, with actions by a, and the next executed line is marked by ==>.
frame option is set, debugger would print entered (and optionally exited)
subroutines in different styles.
What follows is the start of the listing of
env "PERLDB_OPTS=f=n N" perl -d -V
for different values of n:
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
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
in $=main::BEGIN() from /dev/nul: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/nul:0
in $=Exporter::export('Config', 'main', 'myconfig', 'config_vars') from li
in @=Config::myconfig() from /dev/nul: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), 'PATCHLEVEL') from lib/Config.pm:574
in $=Config::FETCH(ref(Config), '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
in $=main::BEGIN() from /dev/nul: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/nul: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/nul:0
out $=main::BEGIN() from /dev/nul:0
in @=Config::myconfig() from /dev/nul: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), 'PATCHLEVEL') from lib/Config.pm:574
out $=Config::FETCH(ref(Config), 'PATCHLEVEL') from lib/Config.pm:574
in $=Config::FETCH(ref(Config), 'SUBVERSION') from lib/Config.pm:574
in $=main::BEGIN() from /dev/nul: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/nul: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/nul:0
out $=main::BEGIN() from /dev/nul:0
in @=Config::myconfig() from /dev/nul:0
in $=Config::FETCH('Config=HASH(0x1aa444)', 'package') from lib/Config.pm:574
out $=Config::FETCH('Config=HASH(0x1aa444)', 'package') from lib/Config.pm:574
in $=Config::FETCH('Config=HASH(0x1aa444)', 'baserev') from lib/Config.pm:574
out $=Config::FETCH('Config=HASH(0x1aa444)', 'baserev') from lib/Config.pm:574
in $=CODE(0x15eca4)() from /dev/null:0
in $=CODE(0x182528)() from lib/Config.pm:2
Package lib/Exporter.pm.
out $=CODE(0x182528)() from lib/Config.pm:0
scalar context return from CODE(0x182528): 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 the cases indentation of lines shows the call tree, if bit 2 of
frame is set, then a line is printed on exit from a subroutine as well, if bit 4
is set, then the arguments are printed as well as 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 as well.
When a package is compiled, a line like this
Package lib/Carp.pm.
is printed with proper indentation.
If you have any compile-time executable statements (code within a
BEGIN block or a use statement), these will NOT be stopped by debugger, although requires will (and compile-time statements can be traced with AutoTrace option set in PERLDB_OPTS). From your own Perl code, however, you can transfer control back to the
debugger using the following statement, which is harmless if the debugger
is not running:
$DB::single = 1;
If you set $DB::single to the value 2, it's equivalent to having just typed the n command, whereas a value of 1 means the s
command. The $DB::trace variable should be set to 1 to simulate having typed the t command.
Another way to debug compile-time code is to start debugger, set a breakpoint on load of some module thusly
DB<7> b load f:/perllib/lib/Carp.pm Will stop on load of `f:/perllib/lib/Carp.pm'.
and restart debugger by R command (if possible). One can use b
compile subname for the same purpose.
Most probably you do not want to modify the debugger, it contains enough
hooks to satisfy most needs. You may change the behaviour of debugger from
the debugger itself, using Options, from the command line via
PERLDB_OPTS environment variable, and from customization files.
You can do some customization by setting up a .perldb file which contains initialization code. For instance, you could make aliases like these (the last one is one people expect to be there):
$DB::alias{'len'} = 's/^len(.*)/p length($1)/';
$DB::alias{'stop'} = 's/^stop (at|in)/b/';
$DB::alias{'ps'} = 's/^ps\b/p scalar /';
$DB::alias{'quit'} = 's/^quit(\s*)/exit\$/';
One changes options from .perldb file via calls like this one;
parse_options("NonStop=1 LineInfo=db.out AutoTrace=1 frame=2");
(the code is executed in the package DB). Note that .perldb is processed before processing PERLDB_OPTS. If .perldb defines the subroutine afterinit, it is called after all the debugger initialization ends. .perldb may be contained in the current directory, or in the LOGDIR/HOME directory.
If you want to modify the debugger, copy perl5db.pl from the Perl library to another name and modify it as necessary. You'll also want to set your PERL5DB environment variable to say something like this:
BEGIN { require "myperl5db.pl" }
As the last resort, one can use PERL5DB to customize debugger by directly setting internal variables or calling debugger functions.
As shipped, the only command line history supplied is a simplistic one that checks for leading exclamation points. However, if you install the Term::ReadKey and Term::ReadLine modules from CPAN, you will have full editing capabilities much like GNU readline(3) provides. Look for these in the modules/by-module/Term directory on CPAN.
A rudimentary command line completion is also available. Unfortunately, the names of lexical variables are not available for completion.
If you have GNU emacs installed on your system, it can interact with the Perl debugger to provide an integrated software development environment reminiscent of its interactions with C debuggers.
Perl is also delivered with a start file for making emacs act like a syntax-directed editor that understands (some of) Perl's syntax. Look in the emacs directory of the Perl source distribution.
(Historically, a similar setup for interacting with vi and the X11 window system had also been available, but at the time of this writing, no debugger support for vi currently exists.)
If you wish to supply an alternative debugger for Perl to run, just invoke your script with a colon and a package argument given to the -d flag. One of the most popular alternative debuggers for Perl is DProf, the Perl profiler. As of this writing, DProf is not included with the standard Perl distribution, but it is expected to be included soon, for certain values of ``soon''.
Meanwhile, you can fetch the Devel::Dprof module from CPAN. Assuming it's properly installed on your system, to profile your Perl program in the file mycode.pl, just type:
perl -d:DProf mycode.pl
When the script terminates the profiler will dump the profile information to a file called tmon.out. A tool like dprofpp (also supplied with the Devel::DProf package) can be used to interpret the information which is in that profile.
When you call the caller function (see caller) from the package DB, Perl sets the array @DB::args to contain the arguments the corresponding stack frame was called with.
If perl is run with -d option, the following additional features are enabled (cf. $^P):
Perl inserts the contents of $ENV{PERL5DB} (or BEGIN {require
'perl5db.pl'} if not present) before the first line of the application.
@{"_<$filename"} is the line-by-line contents of $filename for all the compiled
files. Same for evaled strings which contain subroutines, or which are currently executed. The $filename
for evaled strings looks like (eval 34).
%{"_<$filename"} contains breakpoints and action (it is keyed by line number), and
individual entries are settable (as opposed to the whole hash). Only
true/false is important to Perl, though the values used by perl5db.pl have the form
"$break_condition\0$action". Values are magical in numeric context: they are zeros if the line is not
breakable.
Same for evaluated strings which contain subroutines, or which are
currently executed. The $filename for evaled strings looks like
(eval 34).
${"_<$filename"} contains "_<$filename". Same for evaluated strings which contain subroutines, or which are
currently executed. The $filename for evaled strings looks like (eval
34).
DB::postponed(*{"_<$filename"}) is called (if subroutine
DB::postponed exists). Here the $filename is the expanded name of the required file (as found in values of
%INC).
subname is compiled existence of
$DB::postponed{subname} is checked. If this key exists,
DB::postponed(subname) is called (if subroutine DB::postponed
exists).
%DB::sub is maintained, with keys being subroutine names, values having the form filename:startline-endline. filename has the form (eval 31) for subroutines defined inside evals.
DB::DB() is performed if any one of variables $DB::trace, $DB::single, or
$DB::signal is true. (Note that these variables are not localizable.) This feature is disabled when the control is inside DB::DB() or functions called from it (unless
$^D & (1<<30)).
&DB::sub(args) is performed instead, with $DB::sub being the name of the called subroutine. (Unless the subroutine is compiled
in the package DB.)
Note that if &DB::sub needs some external data to be setup for it to work, no subroutine call is
possible until this is done. For the standard debugger $DB::deep (how many levels of recursion deep into the debugger you can go before a
mandatory break) gives an example of such a dependency.
The minimal working debugger consists of one line
sub DB::DB {}
which is quite handy as contents of PERL5DB environment variable:
env "PERL5DB=sub DB::DB {}" perl -d your-script
Another (a little bit more useful) minimal debugger can be created with the only line being
sub DB::DB {print ++$i; scalar <STDIN>}
This debugger would print the sequential number of encountered statement,
and would wait for your CR to continue.
The following debugger is quite functional:
{
package DB;
sub DB {}
sub sub {print ++$i, " $sub\n"; &$sub}
}
It prints the sequential number of subroutine call and the name of the
called subroutine. Note that &DB::sub should be compiled into the package DB.
At the start, the debugger reads your rc file (./.perldb or
~/.perldb under Unix), which can set important options. This file may define a
subroutine &afterinit to be executed after the debugger is initialized.
After the rc file is read, the debugger reads environment variable
PERLDB_OPTS and parses it as a rest of O ... line in debugger prompt.
It also maintains magical internal variables, such as @DB::dbline,
%DB::dbline, which are aliases for @{"::_<current_file"}
%{"::_<current_file"}. Here current_file is the currently selected (with the debugger's f command, or by flow of execution) file.
Some functions are provided to simplify customization. See Debugger Customization for description of DB::parse_options(string). The function DB::dump_trace(skip[, count]) skips the specified number of frames, and returns a list containing info
about the caller frames (all if count is missing). Each entry is a hash with keys
context ($ or @), sub (subroutine name, or info about eval), args (undef or a reference to an array), file, and
line.
The function DB::print_trace(FH, skip[, count[, short]]) prints formatted info about caller frames. The last two functions may be
convenient as arguments to <, << commands.
You did try the -w switch, didn't you?
You cannot get the stack frame information or otherwise debug functions that were not compiled by Perl, such as C or C++ extensions.
If you alter your @_ arguments in a subroutine (such as with shift
or pop, the stack backtrace will not show the original values.
Perl is very frivolous with memory. There is a saying that to estimate memory usage of Perl, assume a reasonable algorithm of allocation, and multiply your estimages by 10. This is not absolutely true, but may give you a good grasp of what happens.
Say, 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 much 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 by another 20 bytes.
A sloppy
malloc() implementation will make these numbers yet more.
On the opposite end of the scale, a declaration like
sub foo;
may take (on some versions of perl) up to 500 bytes of memory.
Off-the-cuff anecdotal estimates of a code bloat give a factor around 8. This means that the compiled form of reasonable (commented indented etc.) code will take approximately 8 times more than the disk space the code takes.
There are two Perl-specific ways to analyze the memory usage:
$ENV{PERL_DEBUG_MSTATS} and -DL switch. First one is available only if perl is compiled with Perl's
malloc(), the second one only
if Perl compiled with -DDEBUGGING (as with giving -D optimise=-g
option to Configure).
$ENV{PERL_DEBUG_MSTATS}
If your perl is using Perl's
malloc(), and compiled with
correct switches (this is the default), then it will print memory usage
statistics after compiling your code (if $ENV{PERL_DEBUG_MSTATS} > 1), and before termination of the script (if
$ENV{PERL_DEBUG_MSTATS} >= 1). The report format is similar to one in the following example:
env PERL_DEBUG_MSTATS=2 perl -e "require Carp"
Memory allocation statistics after compilation: (buckets 4(4)..8188(8192)
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(8192)
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 moment by usind Devel::Peek::mstats() (module Devel::Peek is available on CPAN).
Here is the explanation of different parts of the format:
malloc() uses bucketed
allocations. Every request is rounded up to the closest bucket size
available, and a bucket of these size is taken from the pool of the buckets
of this size.
The above line describes limits of buckets currently in use. Each bucket has two sizes: memory footprint, and the maximal size of user data which may be put into this bucket. Say, in the above example the smallest bucket is both sizes 4. The biggest bucket has usable size 8188, and the memory footprint 8192.
With debugging Perl some buckets may have negative usable size. This means
that these buckets cannot (and will not) be used. For greater buckets the
memory footprint may be one page greater than a power of 2. In such a case
the corresponding power of two is printed instead in the APPROX field above.
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 memory footprints of two
buckets ``above''.
Say, with the above example the memory footprints are (with current algorith)
free: 8 16 32 64 128 256 512 1024 2048 4096 8192
4 12 24 48 80
With non-DEBUGGING perl the buckets starting from 128-long ones have 4-byte overhead, thus 8192-long bucket may take up to
8188-byte-long allocations.
sbrk()ed, and number of
sbrk()s used. The third number is what perl thinks about continuity of returned chunks. As far as this number is positive,
malloc() will assume that it is probable that
sbrk() will provide continuous memory.
The amounts
sbrk()ed by external
libraries is not counted.
sbrk()ed memory needed to
keep buckets aligned.
malloc() may want to
subdivide a bigger bucket into smaller buckets. If only a part of the
deceased-bucket is left non-subdivided, the rest is kept as an element of a
linked list. This field gives the total size of these chunks.
sbrk()s
malloc() asks for more memory. This field gives the size of the yet-unused part, which is
sbrk()ed, but never touched.
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 ;
Note: the discussion below supposes 32-bit architecture. In the newer versions of perl the memory usage of the constructs discussed here is much improved, but the story discussed below is a real-life story. This story is very terse, and assumes more than cursory knowledge of Perl internals.
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 warn('!...') statements around the call:
warn('!');
do 'lib/auto/POSIX/autosplit.ix';
warn('!!! "after"');
and run it with -DL option. The first
warn() will print memory allocation info before the parsing of the file, and will memorize the statistics at this point (we ignore what it prints). The second
warn() will print increments w.r.t. this memorized statistics. This is the above printout.
Different Ids on the left correspond to different subsystems of perl interpreter, they are just first argument given to perl memory allocation
API
New(). To find what
9 03 means grep the perl source for 903. You will see that it is util.c, function
savepvn(). This function is used to store a copy of existing chunk of memory. Using
C debugger, one can see that it is called either directly from
gv_init(), or via
sv_magic(), and
gv_init() is called from
gv_fetchpv() - which is called from
newSUB().
Note: to reach this place in debugger and skip all the calls to savepvn during the compilation of the main script, set a
C breakpoint in
Perl_warn(),
continue this point is reached, then set breakpoint in
Perl_savepvn(). Note that you may need to skip a handful of
Perl_savepvn() which do not correspond to mass production of CVs (there are more
903 allocations than 146 similar lines of
lib/auto/POSIX/autosplit.ix). Note also that Perl_ prefixes are added by macroization code in perl header files to avoid
conflicts with external libraries.
Anyway, we see that 903 ids correspond to creation of globs, twice per glob - for glob name, and
glob stringification magic.
Here are explanations for other Ids above:
is for creation of bigger XPV* structures. In the above case it creates 3 AV per subroutine, one for a list of lexical variable names, one for a
scratchpad (which contains lexical variables and
targets), and one for the array of scratchpads needed for recursion.
It also creates a GV and a CV per subroutine (all called from
start_subparse()).
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 freeed, but are kept as additional arenas for SV allocations).
HEK 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 HE.
GP for the glob for the subroutine.
MAGIC for the glob for the subroutine.
If Perl is run with -DL option, then
warn()s which 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
If an extension or an external library does not use Perl API to allocate memory, these allocations are not counted.
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.
Otherwise, one can use re 'debug', which has effects both at compile time, and at run time (and is not lexically scoped).
The debugging output for the compile time looks like this:
compiling RE `[bc]d(ef*g)+h[ij]k$'
size 43 first at 1
1: ANYOF(11)
11: EXACT <d>(13)
13: CURLYX {1,32767}(27)
15: OPEN1(17)
17: EXACT <e>(19)
19: STAR(22)
20: EXACT <f>(0)
22: EXACT <g>(24)
24: CLOSE1(26)
26: WHILEM(0)
27: NOTHING(28)
28: EXACT <h>(30)
30: ANYOF(40)
40: EXACT <k>(42)
42: EOL(43)
43: END(0)
anchored `de' at 1 floating `gh' at 3..2147483647 (checking floating)
stclass `ANYOF' minlen 7
The first line shows the pre-compiled form of the regexp, and the second shows the size of the compiled form (in arbitrary units, usually 4-byte words) and the label id of the first node which does a match.
The last line (split into two lines in the above) contains the optimizer
info. In the example shown, the optimizer found that the match should
contain a substring de at the offset 1, and substring gh
at some offset between 3 and infinity. Moreover, when checking for these
substrings (to abandon impossible matches quickly) it 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 the match cannot be shorter than 7 chars.
The fields of interest which may appear in the last line are
\G;
x+y);
.*;
TYPE being
BOL, 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 info is used to avoid entering (a slow)
RE engine on strings which will definitely not match.
If isall flag is set, a call to the
RE engine may be avoided even when optimizer found an
appropriate place for the match.
The rest of the output contains the list of nodes of the compiled form of the RE. Each line has format
<PRE>
</PRE>
id: TYPE OPTIONAL-INFO (next-id)
Here is the list of 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 parenths.
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-RE.
IFTHEN off 1 1 Switch, should be preceeded by switcher .
GROUPP num 1 Whether the group matched.
# Support for long RE
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.
First of all, when doing a match, one may get no run-time output even if debugging is enabled. this means that the RE engine was never entered, all of the job was done by the optimizer.
If RE 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 RE which is currently being tested against the target string. The format of these lines is
<PRE>
</PRE>
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.
If rather than formatting bugs, you encounter substantive content errors in these documents, such as mistakes in the explanations or code, please use the perlbug utility included with the Perl distribution.