Fix form.t compatibility with HTTP::Message 7.01+

[Copilot]

Problem

The form.t test file fails with HTTP::Message version 7.01 and later due to a change in how HTTP requests are stringified. In HTTP::Message 7.01+, the as_string() method no longer includes a trailing newline, causing test failures when comparing expected vs actual output.

This is the same issue reported in libwww-perl/libwww-perl#491.

Solution

Added a helper function is_maybe_nl() that handles both versions of HTTP::Message by detecting when the actual output is missing a trailing newline that exists in the expected output, and normalizing them before comparison:

sub is_maybe_nl {
    my ($got, $expected, $message) = @_;
    
    local $Test::Builder::Level = $Test::Builder::Level + 1;
    
    if ($got !~ m{\n$} && $expected =~ m{\n$}) {
        chomp($expected);
    }
    
    is($got, $expected, $message);
}

Updated 5 test assertions that check POST requests with message bodies to use is_maybe_nl() instead of is(). These tests now pass with both older (< 7.01) and newer (≥ 7.01) versions of HTTP::Message.

Changes

• Added is_maybe_nl() helper function for version-agnostic string comparison
• Replaced is() with is_maybe_nl() for POST request assertions (lines 100, 480, 489, 619, 634)
• Added trailing newline to done_testing; for consistency
• Updated .gitignore to exclude MYMETA build artifacts

Fixes compatibility with HTTP::Message 7.01+ while maintaining backward compatibility with earlier versions.

Original prompt

---

This section details on the original issue you should resolve

<issue_title>form.t fails with HTTP::Message 7.01</issue_title>
<issue_description>same problem as libwww-perl/libwww-perl#491

HTML-Form-6.12.patch fixes the tests</issue_description>

Comments on the Issue (you are @copilot in this section)

Fixes #56

---

💬 We'd love your input! Share your thoughts on Copilot coding agent in our 2 minute survey.

[haarg]

Newlines should not be optional. The HTTP::Message prereq should be bumped.

[oalders]

Newlines should not be optional. The HTTP::Message prereq should be bumped.

I see that was in the original patch. I'll rework it to do just the bump. That was a helpful start, but I'll update it.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 85.68%. Comparing base (90132a6) to head (b086572).
⚠️ Report is 5 commits behind head on master.

Additional details and impacted files

@@           Coverage Diff           @@
##           master      #57   +/-   ##
=======================================
  Coverage   85.68%   85.68%           
=======================================
  Files           9        9           
  Lines         545      545           
  Branches      165      165           
=======================================
  Hits          467      467           
  Misses         38       38           
  Partials       40       40           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[haarg]

This still seems wrong to me. Using chomp means the newline is acceptable if it exists or not. I think it should be testing the exact output.

[oalders]

We chomp the expected output to strip the trailing newline out of the HEREDOC and then test what we actually got using is() -- so this should be a straight check for equality on the exact output.