369 lines
13 KiB
Perl
369 lines
13 KiB
Perl
# ABSTRACT: Report errors from perspective of caller of a "clan" of modules
|
|
|
|
##
|
|
## Based on Carp.pm from Perl 5.005_03.
|
|
## Last modified 22-May-2016 by Kent Fredric.
|
|
## Should be reasonably backwards compatible.
|
|
##
|
|
## This module is free software and can
|
|
## be used, modified and redistributed
|
|
## under the same terms as Perl itself.
|
|
##
|
|
|
|
@DB::args = (); # Avoid warning "used only once" in Perl 5.003
|
|
|
|
package Carp::Clan; # git description: v6.07-8-g8b5dba6
|
|
|
|
use strict;
|
|
use overload ();
|
|
|
|
# Original comments by Andy Wardley <abw@kfs.org> 09-Apr-1998.
|
|
|
|
# The $Max(EvalLen|(Arg(Len|Nums)) variables are used to specify how
|
|
# the eval text and function arguments should be formatted when printed.
|
|
|
|
our $MaxEvalLen = 0; # How much eval '...text...' to show. 0 = all.
|
|
our $MaxArgLen = 64; # How much of each argument to print. 0 = all.
|
|
our $MaxArgNums = 8; # How many arguments to print. 0 = all.
|
|
|
|
our $Verbose = 0; # If true then make _shortmsg call _longmsg instead.
|
|
|
|
our $VERSION = '6.08';
|
|
|
|
# _longmsg() crawls all the way up the stack reporting on all the function
|
|
# calls made. The error string, $error, is originally constructed from the
|
|
# arguments passed into _longmsg() via confess(), cluck() or _shortmsg().
|
|
# This gets appended with the stack trace messages which are generated for
|
|
# each function call on the stack.
|
|
|
|
sub _longmsg {
|
|
return (@_) if ( ref $_[0] );
|
|
local $_; # Protect surrounding program - just in case...
|
|
my ( $pack, $file, $line, $sub, $hargs, $eval, $require, @parms, $push );
|
|
my $error = join( '', @_ );
|
|
my $msg = '';
|
|
my $i = 0;
|
|
while (
|
|
do {
|
|
{
|
|
|
|
package # hide from PAUSE
|
|
DB;
|
|
( $pack, $file, $line, $sub, $hargs, undef, $eval, $require )
|
|
= caller( $i++ )
|
|
}
|
|
}
|
|
)
|
|
{
|
|
next if ( $pack eq 'Carp::Clan' );
|
|
if ( $error eq '' ) {
|
|
if ( defined $eval ) {
|
|
$eval =~ s/([\\\'])/\\$1/g unless ($require); # Escape \ and '
|
|
$eval
|
|
=~ s/([\x00-\x1F\x7F-\xFF])/sprintf("\\x%02X",ord($1))/eg;
|
|
substr( $eval, $MaxEvalLen ) = '...'
|
|
if ( $MaxEvalLen && length($eval) > $MaxEvalLen );
|
|
if ($require) { $sub = "require $eval"; }
|
|
else { $sub = "eval '$eval'"; }
|
|
}
|
|
elsif ( $sub eq '(eval)' ) { $sub = 'eval {...}'; }
|
|
else {
|
|
@parms = ();
|
|
if ($hargs) {
|
|
$push = 0;
|
|
@parms = @DB::args
|
|
; # We may trash some of the args so we take a copy
|
|
if ( $MaxArgNums and @parms > $MaxArgNums ) {
|
|
$#parms = $MaxArgNums;
|
|
pop(@parms);
|
|
$push = 1;
|
|
}
|
|
for (@parms) {
|
|
if ( defined $_ ) {
|
|
if ( ref $_ ) {
|
|
$_ = overload::StrVal($_);
|
|
}
|
|
else {
|
|
unless ( /^-?\d+(?:\.\d+(?:[eE][+-]\d+)?)?$/
|
|
) # Looks numeric
|
|
{
|
|
s/([\\\'])/\\$1/g; # Escape \ and '
|
|
s/([\x00-\x1F\x7F-\xFF])/sprintf("\\x%02X",ord($1))/eg;
|
|
substr( $_, $MaxArgLen ) = '...'
|
|
if ( $MaxArgLen
|
|
and length($_) > $MaxArgLen );
|
|
$_ = "'$_'";
|
|
}
|
|
}
|
|
}
|
|
else { $_ = 'undef'; }
|
|
}
|
|
push( @parms, '...' ) if ($push);
|
|
}
|
|
$sub .= '(' . join( ', ', @parms ) . ')';
|
|
}
|
|
if ( $msg eq '' ) { $msg = "$sub called"; }
|
|
else { $msg .= "\t$sub called"; }
|
|
}
|
|
else {
|
|
$msg = quotemeta($sub);
|
|
if ( $error =~ /\b$msg\b/ ) { $msg = $error; }
|
|
else {
|
|
if ( $sub =~ /::/ ) { $msg = "$sub(): $error"; }
|
|
else { $msg = "$sub: $error"; }
|
|
}
|
|
}
|
|
$msg .= " at $file line $line\n" unless ( $error =~ /\n$/ );
|
|
$error = '';
|
|
}
|
|
$msg ||= $error;
|
|
$msg =~ tr/\0//d; # Circumvent die's incorrect handling of NUL characters
|
|
$msg;
|
|
}
|
|
|
|
# _shortmsg() is called by carp() and croak() to skip all the way up to
|
|
# the top-level caller's package and report the error from there. confess()
|
|
# and cluck() generate a full stack trace so they call _longmsg() to
|
|
# generate that. In verbose mode _shortmsg() calls _longmsg() so you
|
|
# always get a stack trace.
|
|
|
|
sub _shortmsg {
|
|
my $pattern = shift;
|
|
my $verbose = shift;
|
|
return (@_) if ( ref $_[0] );
|
|
goto &_longmsg if ( $Verbose or $verbose );
|
|
my ( $pack, $file, $line, $sub );
|
|
my $error = join( '', @_ );
|
|
my $msg = '';
|
|
my $i = 0;
|
|
while ( ( $pack, $file, $line, $sub ) = caller( $i++ ) ) {
|
|
next if ( $pack eq 'Carp::Clan' or $pack =~ /$pattern/ );
|
|
if ( $error eq '' ) { $msg = "$sub() called"; }
|
|
else {
|
|
$msg = quotemeta($sub);
|
|
if ( $error =~ /\b$msg\b/ ) { $msg = $error; }
|
|
else {
|
|
if ( $sub =~ /::/ ) { $msg = "$sub(): $error"; }
|
|
else { $msg = "$sub: $error"; }
|
|
}
|
|
}
|
|
$msg .= " at $file line $line\n" unless ( $error =~ /\n$/ );
|
|
$msg =~ tr/\0//d; # Circumvent die's incorrect handling of NUL characters
|
|
return $msg;
|
|
}
|
|
goto &_longmsg;
|
|
}
|
|
|
|
# In the two identical regular expressions (immediately after the two occurrences of
|
|
# "quotemeta") above, the "\b ... \b" helps to avoid confusion between function names
|
|
# which are prefixes of each other, e.g. "My::Class::print" and "My::Class::println".
|
|
|
|
# The following four functions call _longmsg() or _shortmsg() depending on
|
|
# whether they should generate a full stack trace (confess() and cluck())
|
|
# or simply report the caller's package (croak() and carp()), respectively.
|
|
# confess() and croak() die, carp() and cluck() warn.
|
|
|
|
# Following code kept for calls with fully qualified subroutine names:
|
|
# (For backward compatibility with the original Carp.pm)
|
|
|
|
sub croak {
|
|
my $callpkg = caller(0);
|
|
my $pattern = ( $callpkg eq 'main' ) ? '^:::' : "^$callpkg\$";
|
|
die _shortmsg( $pattern, 0, @_ );
|
|
}
|
|
sub confess { die _longmsg(@_); }
|
|
|
|
sub carp {
|
|
my $callpkg = caller(0);
|
|
my $pattern = ( $callpkg eq 'main' ) ? '^:::' : "^$callpkg\$";
|
|
warn _shortmsg( $pattern, 0, @_ );
|
|
}
|
|
sub cluck { warn _longmsg(@_); }
|
|
|
|
# The following method imports a different closure for every caller.
|
|
# I.e., different modules can use this module at the same time
|
|
# and in parallel and still use different patterns.
|
|
|
|
sub import {
|
|
my $pkg = shift;
|
|
my $callpkg = caller(0);
|
|
my $pattern = ( $callpkg eq 'main' ) ? '^:::' : "^$callpkg\$";
|
|
my $verbose = 0;
|
|
my $item;
|
|
my $file;
|
|
|
|
for $item (@_) {
|
|
if ( $item =~ /^\d/ ) {
|
|
if ( $VERSION < $item ) {
|
|
$file = "$pkg.pm";
|
|
$file =~ s!::!/!g;
|
|
$file = $INC{$file};
|
|
die _shortmsg( '^:::', 0,
|
|
"$pkg $item required--this is only version $VERSION ($file)"
|
|
);
|
|
}
|
|
}
|
|
elsif ( $item =~ /^verbose$/i ) { $verbose = 1; }
|
|
else { $pattern = $item; }
|
|
}
|
|
|
|
eval { $pattern = qr/$pattern/ };
|
|
|
|
if ($@) {
|
|
$@ =~ s/\s+$//;
|
|
$@ =~ s/\s+at\s.+$//;
|
|
die _shortmsg( '^:::', 0, $@ );
|
|
}
|
|
|
|
{
|
|
local ($^W) = 0;
|
|
no strict "refs";
|
|
*{"${callpkg}::croak"} = sub { die _shortmsg( $pattern, $verbose, @_ ); };
|
|
*{"${callpkg}::confess"} = sub { die _longmsg ( @_ ); };
|
|
*{"${callpkg}::carp"} = sub { warn _shortmsg( $pattern, $verbose, @_ ); };
|
|
*{"${callpkg}::cluck"} = sub { warn _longmsg ( @_ ); };
|
|
}
|
|
}
|
|
|
|
1;
|
|
|
|
__END__
|
|
|
|
=pod
|
|
|
|
=encoding UTF-8
|
|
|
|
=head1 NAME
|
|
|
|
Carp::Clan - Report errors from perspective of caller of a "clan" of modules
|
|
|
|
=head1 VERSION
|
|
|
|
version 6.08
|
|
|
|
=head1 SYNOPSIS
|
|
|
|
carp - warn of errors (from perspective of caller)
|
|
|
|
cluck - warn of errors with stack backtrace
|
|
|
|
croak - die of errors (from perspective of caller)
|
|
|
|
confess - die of errors with stack backtrace
|
|
|
|
use Carp::Clan qw(^MyClan::);
|
|
croak "We're outta here!";
|
|
|
|
use Carp::Clan;
|
|
confess "This is how we got here!";
|
|
|
|
=head1 DESCRIPTION
|
|
|
|
This module is based on "C<Carp.pm>" from Perl 5.005_03. It has been
|
|
modified to skip all package names matching the pattern given in
|
|
the "use" statement inside the "C<qw()>" term (or argument list).
|
|
|
|
Suppose you have a family of modules or classes named "Pack::A",
|
|
"Pack::B" and so on, and each of them uses "C<Carp::Clan qw(^Pack::);>"
|
|
(or at least the one in which the error or warning gets raised).
|
|
|
|
Thus when for example your script "tool.pl" calls module "Pack::A",
|
|
and module "Pack::A" calls module "Pack::B", an exception raised in
|
|
module "Pack::B" will appear to have originated in "tool.pl" where
|
|
"Pack::A" was called, and not in "Pack::A" where "Pack::B" was called,
|
|
as the unmodified "C<Carp.pm>" would try to make you believe C<:-)>.
|
|
|
|
This works similarly if "Pack::B" calls "Pack::C" where the
|
|
exception is raised, et cetera.
|
|
|
|
In other words, this blames all errors in the "C<Pack::*>" modules
|
|
on the user of these modules, i.e., on you. C<;-)>
|
|
|
|
The skipping of a clan (or family) of packages according to a pattern
|
|
describing its members is necessary in cases where these modules are
|
|
not classes derived from each other (and thus when examining C<@ISA>
|
|
- as in the original "C<Carp.pm>" module - doesn't help).
|
|
|
|
The purpose and advantage of this is that a "clan" of modules can work
|
|
together (and call each other) and throw exceptions at various depths
|
|
down the calling hierarchy and still appear as a monolithic block (as
|
|
though they were a single module) from the perspective of the caller.
|
|
|
|
In case you just want to ward off all error messages from the module
|
|
in which you "C<use Carp::Clan>", i.e., if you want to make all error
|
|
messages or warnings to appear to originate from where your module
|
|
was called (this is what you usually used to "C<use Carp;>" for C<;-)>),
|
|
instead of in your module itself (which is what you can do with a
|
|
"die" or "warn" anyway), you do not need to provide a pattern,
|
|
the module will automatically provide the correct one for you.
|
|
|
|
I.e., just "C<use Carp::Clan;>" without any arguments and call "carp"
|
|
or "croak" as appropriate, and they will automatically defend your
|
|
module against all blames!
|
|
|
|
In other words, a pattern is only necessary if you want to make
|
|
several modules (more than one) work together and appear as though
|
|
they were only one.
|
|
|
|
=head2 Forcing a Stack Trace
|
|
|
|
As a debugging aid, you can force "C<Carp::Clan>" to treat a "croak" as
|
|
a "confess" and a "carp" as a "cluck". In other words, force a detailed
|
|
stack trace to be given. This can be very helpful when trying to
|
|
understand why, or from where, a warning or error is being generated.
|
|
|
|
This feature is enabled either by "importing" the non-existent symbol
|
|
'verbose', or by setting the global variable "C<$Carp::Clan::Verbose>"
|
|
to a true value.
|
|
|
|
You would typically enable it by saying
|
|
|
|
use Carp::Clan qw(verbose);
|
|
|
|
Note that you can both specify a "family pattern" and the string "verbose"
|
|
inside the "C<qw()>" term (or argument list) of the "use" statement, but
|
|
consider that a pattern of packages to skip is pointless when "verbose"
|
|
causes a full stack trace anyway.
|
|
|
|
=head1 BUGS
|
|
|
|
The "C<Carp::Clan>" routines don't handle exception objects currently.
|
|
If called with a first argument that is a reference, they simply
|
|
call "C<die()>" or "C<warn()>", as appropriate.
|
|
|
|
Bugs may be submitted through L<the RT bug tracker|https://rt.cpan.org/Public/Dist/Display.html?Name=Carp-Clan>
|
|
(or L<bug-Carp-Clan@rt.cpan.org|mailto:bug-Carp-Clan@rt.cpan.org>).
|
|
|
|
=head1 AUTHOR
|
|
|
|
Steffen Beyer <STBEY@cpan.org>
|
|
|
|
=head1 CONTRIBUTORS
|
|
|
|
=for stopwords Karen Etheridge Joshua ben Jore Kent Fredric
|
|
|
|
=over 4
|
|
|
|
=item *
|
|
|
|
Karen Etheridge <ether@cpan.org>
|
|
|
|
=item *
|
|
|
|
Joshua ben Jore <jjore@cpan.org>
|
|
|
|
=item *
|
|
|
|
Kent Fredric <kentnl@cpan.org>
|
|
|
|
=back
|
|
|
|
=head1 COPYRIGHT AND LICENSE
|
|
|
|
This software is copyright (c) 2001 by Steffen Beyer, Joshua ben Jore.
|
|
|
|
This is free software; you can redistribute it and/or modify it under
|
|
the same terms as the Perl 5 programming language system itself.
|
|
|
|
=cut
|