Debug::ShowStuff - A collection of handy debugging routines for displaying the values of variables with a minimum of coding.
Here's a sampling of a few of my favorite functions in this module.
use Debug::ShowStuff ':all'; # display values of a hash or hash reference showhash %hash; showhash $hashref; # display values of an array or array reference showarr @arr; showarr $arrref; # show all the params received through CGI showcgi(); # A particularly fancy utility: display STDERR at top of web page my $warnings = showstderr;
Debug::ShowStuff can be installed with the usual routine:
Debug::ShowStuff
perl Makefile.PL make make test make install
Debug::ShowStuff grew dynamically from my needs in debugging code. I found myself doing the same tasks over and over... displaying the keys and values in a hash, displaying the elements in an array, displaying the output of STDERR in a web page, etc. Debug::ShowStuff began as two or three of my favorite routines and grew as I added to that collection. Finally I decided to publish these tools in the hope that other Perl programmers will find them useful.
Debug::ShowStuff is for debugging, not for production work. It does not always output the actual value of something, but rather information about the value. For example, the following code outputs the actual value in the first line, but a note about the value in the second.
println 'my value'; println undef;
which outputs
my value [undef]
I would discourage you from using Debug::ShowStuff in production code. Debug::ShowStuff is only for quick-n-dirty displays of variable values in order to debug your code.
The functions in Debug::ShowStuff are designed to output either in plain text mode (like if you're running the script from a command prompt), or in web mode (like from a CGI). If the script appears to be running in a CGI or other web mode (see the inweb function) then values are output using HTML, with special HTML characters escaped for proper display. Othewise the values are output as they are.
inweb
Generally you won't need to bother telling Debug::ShowStuff if you're in text or web mode... it figures it out on its own.
NOTE: The dynamic behavior of "show" functions has changed since earlier versions of Debug::ShowStuff. "show" functions now always outputs to STDOUT or STDERR unless $Debug::ShowStuff::always_void is set to false. By default $always_void is true.
If $always_void is false, then the following applies:
The functions that start with "show" dynamically either output to STDOUT or STDERR or return a string, depending on the context in which the functions are called. For example, if you call showhash in a void context:
showhash %myhash;
then the contents of %myhash are output to STDOUT. On the other hand, if the function is called in scalar context:
my $var = showhash(%myhash);
then the same string that would have been output to STDOUT is instead returned and stored in $var.
By default, output is sent to STDOUT, not STDERR. You can change the default output to STDERR using the setoutput command. See the docs for that command for more detail.
setoutput
To facilitate telling the difference between [undef] and an empty string, functions output the strings "[undef]" and "[empty string]". So, for example, this code:
[undef]
println undef; println "";
produces this:
[undef] [empty string]
println was the original Debug::ShowStuff function. It simply outputs the given values.
println
In text mode it adds a newline to the end.
For example, this code:
println "hello world";
produces this output, including the newline:
hello world
In web mode it puts the output inside a <p> element. The values are HTML escaped so that HTML-significant characters like < and > are actually displayed as < and >. The <p> element has CSS styles set so that the characters are always black, the background is always white, text is left-aligned, and the <p> element has a black border, regardless of the styles of the surrounding elements. So, for example, in web mode, the following code:
println 'whatever';
outputs the following HTML:
<p style="background-color:white;color:black;text-align:left">whatever</p>
Like other "show" functions, undef is output as the string "[undef]" and an empty string is output as the string "[empty string]".
Values in the arguments array are output concatenated together with no spaces in between. Each value is evaluated independently for if it is undef, empty string, or a string with visible content. So, for example, this code:
println "whatever", "", undef, "dude";
outputs this:
whatever[empty string][undef]dude
indent() is for situations where you're outputting a lot of stuff and you want to tidy up the putput with indentation. In text mode the output is indented with 3 spaces per indentation level. In web mode the output is indented with 20px per indentation level.
indent()
indent() must be assigned to a variable or it has no effect. indent() increases the indent level by one. When the variable goes out of scope, the indent level is decreased by one.
For example suppose you want to display the values of records from a database. You might loop through the records, outputting them like this:
while (my $record = $sth->fetchrow_hashref) { println $record->{'name_nick'}; my $indent = indent(); showhash $record; }
That would produce output something like this:
Ricky --------------------------------------- name_first = Rick name_last = Adams --------------------------------------- Dan --------------------------------------- name_first = Daniel name_last = Bradley ---------------------------------------
By default, three spaces are used to indent. To change that set $Debug::ShowStuff::indent_tab to whatever string you want to use for indentation.
option: bottom_space
The bottom_space option indicates to output an extra line at the bottom of the indented output, just to give some extra division before the next batch of code. For example, the following code does not use bottom_space:
bottom_space
foreach my $name (qw[Larry Moe]) { println $name; my $indent = indent(bottom_space=>1); println 'years: ', length($name); }
and so produces the following output:
Larry years: 5 Moe years: 3
But this code:
produces this output:
This function turns on/off most of the functions in this module, with one important exception explained below. The function also returns the state of whether or not Debug::ShowStuff is on/off.
If a parameter is sent, that param is used to turn display on/off. The value is stored in the global variable $Debug::ShowStuff::active.
The function is also used by most subroutines to determine if they should actually output anything. $Debug::ShowStuff::active is not the only criteria used to determine if Debug::ShowStuff is active. The algorithm is as follows:
- If the environment variable $ENV{'SHOWSTUFF'} is defined and false then showstuff() returns false regardless of the state of $active.
- If the environment variable $ENV{'SHOWSTUFF'} is not defined or is defined and true then showstuff() uses $Debug::ShowStuff::active to determine on/off.
The purpose of this algorithm is to allow the use of debugging display in situations where one perl script calls another, such as in regression testing.
For example, suppose you have a script as follows:
#!/usr/bin/perl -w use strict; use Debug::ShowStuff ':all'; my ($rv); println 'running my_function()'; $rv = my_function(); println 'the returned value is: ', $rv; $rv or die 'error!';
The output of the script might look something like this:
running my_function() 1
Now suppose you call that and other scripts from some OTHER script, and you don't want the screen cluttered with all that debugging, you just want to see if all those scripts run successfully. You could use $ENV{'SHOWSTUFF'} to turn off showing stuff:
#!/usr/bin/perl -w use strict; use Debug::ShowStuff ':all'; my @tests = ("./script1.pl", "./script2.pl", "./script3.pl"); $ENV{'SHOWSTUFF'} = 0; foreach my $test () { system($test) and die "$test failed"; }
In that case, none of the stuff from the test scripts would be output.
Works like println but doesn't add trailing newline. In web environment uses <span> instead of <p>.
Displays the keys and values in a hash. Input is either a single hash reference or a regular hash. The key=values pairs are sorted by the names of the keys.
So, for example, the following code:
my %hash = ( Larry => 'curly headed guy', Curly => 'funny bald guy', Moe => 'guy in charge', ); showhash %hash;
Produces the following output. Notice that the keys are in alphabetic order:
--------------------------------------- Curly = funny bald guy Larry = curly headed guy Moe = guy in charge ---------------------------------------
This code, using a hash reference, produces exactly the same output:
my $hash = { Larry => 'curly headed guy', Curly => 'funny bald guy', Moe => 'guy in charge', }; showhash $hash;
If the hash is empty, then that fact is output. So, this code:
showhash {};
--------------------------------------- [empty hash] ---------------------------------------
If an undef value is sent instead of a hashref, then that fact is displayed instead of a hash. For example consider the following code that uses a variable that is undef:
my ($hash); showhash $hash;
That code produces this output:
--------------------------------------- Only one element input and it was undefined ---------------------------------------
Optional arguments only come into play if the first argument is a hashref.
option: title => "string"
If this option is sent, the string is displayed at the top of the display of the hash values. So this code:
my $hash = { Larry => 'curly headed guy', Curly => 'funny bald guy', Moe => 'guy in charge', }; showhash $hash, title=>'Stooges';
--- Stooges --------------------------------- Curly = funny bald guy Larry = curly headed guy Moe = guy in charge ---------------------------------------------
option: line_cut => 1
If the line_cut option is sent, then each value is truncated after the first newline if there is one. The fact that there is more output is mentioned. So the following code:
line_cut
my $hash = { Larry => "curly\nheaded guy", Curly => "funny\nbald guy", Moe => "guy\nin charge", }; showhash $hash, line_cut =>1;
produces this output.
--------------------------------------- Curly = funny [more lines...] Larry = curly [more lines...] Moe = guy [more lines...] ---------------------------------------
Several other options do exactly the same thing: linecut, line_chop, and first_line.
Displays the values of an array. c<showarr> and c<showarray>
Each element is displayed in a table row (in web mode) or on a separate line (in text mode).
If showarray receives exactly one argument, and if that item is an array reference, then the routine assumes that you want to display the elements in the referenced array. Therefore, the following blocks of code display the same thing:
showarray
showarray @myarr; showarray \@myarr;
Works just like showarray, except that in text mode displays a solid line between each element of the array.
Outputs the value of a scalar. The name is slightly innaccurate: you can input an array. The array will be joined together to form a single scalar.
Actually, I hardly ever use showscalar, but it seemed unbalanced to have showhash and showarray without showscalar.
showscalar
showhash
Displays the CGI parameter keys and values. This sub always outputs HTML.
There are several optional parameters, described in the following sections.
option: q
The optional parameter q, may be a CGI query object:
q
my $query = CGI->new(); showcgi q => $query;
If q is not sent, then a CGI object is created on the fly.
option: skipempty
If the optional parameter skipempty is true:
skipempty
showcgi skipempty => 1;
then CGI params that are empty (i.e. do not have at least one non-space character) are not displayed.
option: skip
skip sets a list of parameters to not display. For example, if you don't want to see the choices or toppings params, then call showcgi like this:
skip
choices
toppings
showcgi skip => ['choices', 'toppings'];
Single item lists can be passed in directly without putting them in an anonymous array:
showcgi skip => 'choices';
Displays a hash, array, or scalar references, treeing down through other references it contains. So, for example, the following code:
my $ob = { name => 'Raha', email => 'raha@idocs.com', friends => [ 'Shalom', 'Joe', 'Furocha', ], }; showref $ob;
produces the following output:
/-----------------------------------------------------------\ friends = ARRAY Shalom Joe Furocha email = raha@idocs.com name = Raha \-----------------------------------------------------------/
The values of the hash or arrays being referenced are only displayed once, so you're safe from infinite recursion.
option: maxhash
The maxhash option allows you to indicate the maximum number of hash elements to display. If a hash has more then maxhash elements then none of them are displayed or recursed through, and instead an indicator of how many elements there are is output. So, for example, the following command only displays the hash values if there are 10 or fewer elements in the hash:
maxhash
showref $myob, maxhash=>10;
If maxhash is not sent then there is no maximum.
option: maxarr
The maxarr option allows you to indicate the maximum number of array elements to display. If an array has more then maxarr elements then none of them are displayed or recursed through, and instead an indicator of how many elements there are is output. If maxarr is not sent then there is no maximum.
maxarr
option: depth
The depth option allows you to indicate a maximum depth to display in the tree. If depth is not sent then there is no maximum depth.
depth
A list of hash elements to skip. Only applies to the top element and only if it is a hash.
option: skipall
Works the same as skip, but applies to all hashes in the structure, not just the top-level hash.
Prints a horizontal rule. Handy for dividing up multiple println's.
In text mode, the horizontal rule is a set of 80 dashes. In web mode, the output is either a <hr> element or a <p> element, depending on the title option (see "title" below).
So, for example, the following line outputs a simple horizontal rule:
option: title
If the titleoption is sent, the title is embedded in the horizontal rule. So, for example, the following code produces a horizontal rule with with the string "test" embedded in it:
title
printhr title=>'test';
If only one param is sent, it is assumed that param is the title. So the' following code produces exactly the same thing as the example above:
printhr 'test';
In web mode, a title changes the HTML element that is output. If no title is given then printhr outputs an <hr> element. If a title is given the output is p element with the title as the content. The <p> element has a gray background and a black border.
p
option: dash
If the dashoption is sent, the given character is used as the separator. This param only applies to text mode;
dash
Outputs the given values inside a <pre> element. If not in a web environment, works just like preln.
Works like the die command, except it always adds an end-of-line to the input array so you never get those "at line blah-blah-blah" additions.
die
Works like dieln except it prepends 'dev exit: ' to the end of the string. If no string is sent, just outputs "dev exit".
dieln
Displays an array, then dies using dieln.
For use at the command line. Outputs a prompt to "press enter to continue", then waits for you to do exactly that.
Prompts the user for a y or n. Exits quietly if y is pressed.
Outputs a text/html HTTP header. Not useful if you're using mod_perl.
This function allows you to see, in the web page produced by a CGI, everything the CGI output to STDERR.
To use showstderr, assign the return value of the function to a variable that is scoped to the entire CGI script:
showstderr
my $stderr = showstderr();
You need not do anything more with that variable. The object reference by your variable holds on to everything output to both STDOUT and STDERR. When the variable goes out of scope, the object outputs the STDOUT content with the STDERR content at the top of the web page.
By default, Debug::Showstuff guesses if it should be in text or web mode. These functions are for when you want to explicitly tell Debug::ShowStuff what mode it should be in. forcetext forces text mode. forceweb forces web mode. forcenone tells Debug::Showstuff that you don't want to force either mode.
forcetext
forceweb
forcenone
Returns a guess on if we're in a web environment. The guess is pretty simple: if the environment variable REQUEST_URI is true (in the Perlish sense) then this function returns true.
REQUEST_URI
If the global $Debug::ShowStuff::forceenv is defined, this function returns the value of $Debug::ShowStuff::forceenv.
$Debug::ShowStuff::forceenv
Sends Debug::ShowStuff output to a file instead of to STDOUT or STDERR. The value of this function must be assigned to a variable or it has no effect. Don't do anything with the returned value... it is NOT a file handle. The returned value is an object that, when it goes out of scope, closes the output file handle.
For example, the following code will output to three files names Larry.txt, Curly.txt, and Moe.txt:
foreach my $name (qw[Larry Curyl Moe]) { my $output = output_to_file("$name.txt"); println $name; println 'length: ', length($name); }
Sets the default output handle. By default, routines in Debug::ShowStuff output to STDOUT. With this command you can set the default output to STDERR, back to STDOUT, to a filehandle you specify, or to a Debug::ShowStuff::SeparatePrint file handle.
The following command sets default output to STDERR:
setoutput 'stderr';
This command sets output back to STDOUT:
setoutput 'stdout';
This command sends output to a file handle of your choice:
setoutput $fh;
This command sends output to a Debug::ShowStuff::SeparatePrint file handle. This option is a good way to create a simple log file. When you print to this type of file handle, the file is opened separately just for that write, an exclusive lock on the file is obtained, and the end of the file is sought.
Note that the next parameter must be the path to the output file:
setoutput 'separateprint', $file_path;
option: new
If the new parameter is true, then the output file is open in non-append mode, which means any existing contents are removed.
new
Takes a single argument. If that argument is undefined, returns an empty string. Otherwise returns the argument exactly as it is.
Given one or more file names, searches @INC for where they are located. Returns an array of full file names.
Given an array of values, shows which are tainted and which are not. If the first argument is a hashref, displays the tainted status of each element value.
Outputs a table of all rows in the given DBI statement handle. Note that this function "uses up" the statement handle.
showsql output the results of an SQL statement. showsql takes three parameters: the database handle, the SQL statement, and an array-ref of parameters for the SQL statement.
explainsql outputs the result of an SQL EXPLAIN call. This function works much like showsql. The parameters are the database handle, the SQL statement, and an array-ref of SQL parameters. explainsql prepends "EXPLAIN ANALYZE" to your SQL, runs the statement, then outputs the results.
I have only used explainsql with PostGresql. I would be interested hear about how it works with other database management systems and how it might be improved to work in those environments.
Temporarily turn showstuff on or off. Create a variable in the lexical scope where you want the tempoary change, like this:
my $temp = tempshowstuff(1)
When the variable goes out of scope, showstuff will revert back to its previous state.
Outputs the ISA hierarchy of an object or class. For example, the following code outputs the ISA hierarchy for a Xapat object (Xapat is a web server written in Perl and which uses Net::Server).
$xapat = Xapat->new(%opts); showisa $xapat;
which outputs:
------------------------------------ Xapat Net::Server::HTTP Net::Server::MultiType Net::Server ------------------------------------
Note that showisa loads Class::ISA, which is available on CPAN.
This function is for when you want to display how long it took for your code to run. Assign the return value of this function to a variable. When the variable goes out of scope then the difference between the start and end time is displayed in seconds. For example, the following code:
do { my $timer = timer(); sleep 3; };
start timer duration: 3 seconds
If you send the title option, then that title will be displayed with the beginning and ending output. For example, this code:
do { my $timer = timer(title=>'my block'); sleep 3; };
start timer - my block duration - my block: 3 seconds
method: $timer->elapsed
Returns the difference between when the timer was started and the current time.
method: $timer->silence
Turns off the timer so that it doesn't display anything when it dies.
Copyright (c) 2010-2013 by Miko O'Sullivan. All rights reserved. This program is free software; you can redistribute it and/or modify it under the same terms as Perl itself. This software comes with NO WARRANTY of any kind.
Miko O'Sullivan miko@idocs.com
Initial public release.
Setting sort order of hash keys to case insensitive.
After seven years, decided to update the version on CPAN.
Fixed prerequisite requirement for MemHandle and Taint. Added timer() functions. Some minor documentation fixes and tidying up.
Changing from using Taint module, which has had a lot of problems, to Scalar::Util, which is more (but not completely) stable.
Fixed bug in prerequisites for Scalar::Util.
Added showsth, showsql, and explainsql. Added the separateprint option to setoutput. Tidied up documentation. Fixed problems with prerequisites. Probably added many other features since the last time I uploaded this module, but can't remember tham all.
To install Debug::ShowStuff, copy and paste the appropriate command in to your terminal.
cpanm
cpanm Debug::ShowStuff
CPAN shell
perl -MCPAN -e shell install Debug::ShowStuff
For more information on module installation, please visit the detailed CPAN module installation guide.