I am getting an expected error

Why not test for it in that case? Pick one of Test::Warnings, Test::Warn, Test::Trap, etc.

Why not test for it in that case?

Simply because I didn't know about those Test modules...

In that case may I suggest starting with Test::Warn? I find it the simplest of the three (and these are not the only three of course) as it has a very logically straightforward approach. Simply:

use strict;
use warnings;

use Test::More tests => 1;
use Test::Warn;

# Your code which gives an expected warning
sub foo {
    warn "Stripe Webhook Error: Invalid Stripe Signature\n";
}

# Test your code
warning_is { foo (); } "Stripe Webhook Error: Invalid Stripe Signature
+\n",
    'Invalid sig warning issued';
[download]

Task::Kensho recommends Test::Warnings which is fine but a little more abstract. Both modules are widely used in other dists. I have found Test::Trap useful on occasion but it can take a bit of wrangling to set up properly.

---

🦛

G'day Bod,

The way I normally deal with this is along these lines:

my $expected_error = 1;
my $got_error = 0;

eval {
    # code expected to die with error here
    1;
}
or do {
    $got_error = 1;
};

is $got_error, $expected_error, 'test name';
[download]

Thanks kcott but that still generates the rogue output to STDERR

I've tried this:

my $got_warn = 0;
eval {
    $webhook_fail->process();
} or do {
    $got_warn = 1;
};
ok( $got_warn, "Signature warning generated");
ok( !$webhook_fail->success, "Signature error" );
[download]

Both tests pass but I still get the output to STDERR when the process method is called.

I guess I'm just looking for a way to suppress the output...

"... expected error ... Webhook Error ... Signature error ... an error displayed ..."

Originally, you repeatedly referred to an error; now you seem to have changed that to a warning. My solution traps errors, which you asked for; it doesn't trap warnings.

$ alias perle
alias perle='perl -Mstrict -Mwarnings -Mautodie=:all -MCarp::Always -E
+'

$ perle '
    my $expected_error = 1;
    my $got_error = 0;

    eval {
        my @rgb = qw{#ff0000 #00ff00 #0000ff};
        open my $fh, "<", "nonexistent_file";
        1;
    }
    or do {
        $got_error = 1;
    };

    say "\$expected_error[$expected_error] \$got_error[$got_error]";
'
Possible attempt to put comments in qw() list at -e line 6.
$expected_error[1] $got_error[1]
[download]

Perhaps "warnings: Fatal Warnings" would help you.

— Ken

Update: Problem already found in Re^4: STDERR in Test Results.

---

Both tests pass but I still get the output to STDERR when the process method is called. I guess I'm just looking for a way to suppress the output...

Or perhaps for a place where someone prints to STDERR or warns.

Looking at B::S::W 1.11 at CPAN, B::S::W::process() calls $self->_error() with the error message shown. And now look at that method:

sub _error {
    my ($self, $message) = @_;
     
    $self->{'error'} = $message;
    if (defined &{$self->{'error'}}) {
        &{$self->{'error'}}($message);
    } else {
        STDERR->print("Stripe Webhook Error: $message\n");
    }
}
[download]

That's why you shall use warn and die instead of printing to STDERR. sub _warning has the same problem.

---

Oh, by the way: eval catches errors, not warnings, so your variable would be better named $got_error instead of $got_warn.

Alexander

--
Today I will gladly share my knowledge and experience, for there are no sweeter words than "I told you so". ;-)

It's an extra test phase dependency but if you're determined to make sure that auth error triggers, you could verify it with something like Test::Output.