From fbf95216c55adac8af18bc709b4e3f456641013a Mon Sep 17 00:00:00 2001 From: Adam Niedzielski Date: Thu, 25 Jul 2024 16:51:44 +0200 Subject: [PATCH] Fix LinkCrawlWorker NoMethodError: undefined method `[]' for nil (#31144) --- app/lib/link_details_extractor.rb | 2 +- spec/lib/link_details_extractor_spec.rb | 18 ++++++++++++++++++ 2 files changed, 19 insertions(+), 1 deletion(-) diff --git a/app/lib/link_details_extractor.rb b/app/lib/link_details_extractor.rb index d81f4a306..bd78aef7a 100644 --- a/app/lib/link_details_extractor.rb +++ b/app/lib/link_details_extractor.rb @@ -101,7 +101,7 @@ class LinkDetailsExtractor end def json - @json ||= root_array(Oj.load(@data)).find { |obj| SUPPORTED_TYPES.include?(obj['@type']) } || {} + @json ||= root_array(Oj.load(@data)).compact.find { |obj| SUPPORTED_TYPES.include?(obj['@type']) } || {} end end diff --git a/spec/lib/link_details_extractor_spec.rb b/spec/lib/link_details_extractor_spec.rb index 2a4df70a8..b1e5cedce 100644 --- a/spec/lib/link_details_extractor_spec.rb +++ b/spec/lib/link_details_extractor_spec.rb @@ -129,6 +129,24 @@ RSpec.describe LinkDetailsExtractor do include_examples 'structured data' end + context 'with the first tag is null' do + let(:html) { <<~HTML } + + + + + + + + HTML + + include_examples 'structured data' + end + context 'with preceding block of unsupported LD+JSON' do let(:html) { <<~HTML }