Re: 5.8.0 -> 5.8.1 breakage on Debian: HTML::Parser failure

Walt Mankowski <waltman@pobox.com> writes:

> I don't know enough XS to understand much of that code, but I did see
> that report_tags was defined as an HV*.  I had a hunch that it might
> be related to the hash randomization that was introduced in 5.8.1, so
> I tried setting PERL_HASH_SEED to 0 and rerunning make test.  Sure
> enough, all the tests for WWW::Mechanize now passed!

If the hash is initialized in an 5.8.0 extension and then used 5.8.1
core which hash randomization then the hash values might not match any
more.  Hash lookups will then fail.

Regards,
Gisle

• 5.8.0 -> 5.8.1 breakage on Debian: HTML::Parser failure by Chip Salzenberg
• Re: 5.8.0 -> 5.8.1 breakage on Debian: HTML::Parser failure by Walt Mankowski
• Re: 5.8.0 -> 5.8.1 breakage on Debian: HTML::Parser failure by Gisle Aas
• Re: 5.8.0 -> 5.8.1 breakage on Debian: HTML::Parser failure by Chip Salzenberg
• Re: 5.8.0 -> 5.8.1 breakage on Debian: HTML::Parser failure by Gisle Aas