diff --git a/lib/html_proofer/attribute/url.rb b/lib/html_proofer/attribute/url.rb
index 5c5c59f3..46bdd169 100644
--- a/lib/html_proofer/attribute/url.rb
+++ b/lib/html_proofer/attribute/url.rb
@@ -220,11 +220,24 @@ def without_hash
@url.to_s.sub(/##{hash}/, "")
end
- # catch any obvious issues, like strings in port numbers
+ # catch any obvious issues
private def clean_url!
- return if @url =~ /^([!#{Regexp.last_match(0)}-;=?-\[\]_a-z~]|%[0-9a-fA-F]{2})+$/
-
- @url = Addressable::URI.parse(@url).normalize.to_s
+ parsed_url = Addressable::URI.parse(@url)
+ url = if parsed_url.scheme.nil?
+ parsed_url
+ else
+ parsed_url.normalize
+ end.to_s
+
+ # normalize strips this off, which causes issues with cache
+ @url = if @url.end_with?("/") && !url.end_with?("/")
+ "#{url}/"
+ elsif !@url.end_with?("/") && url.end_with?("/")
+ url.chop
+ else
+ url
+ end
+ rescue Addressable::URI::InvalidURIError # rubocop:disable Lint/SuppressedException; error will be reported at check time
end
private def swap_urls!
diff --git a/spec/html-proofer/fixtures/links/unicode_domain.html b/spec/html-proofer/fixtures/links/unicode_domain.html
index 608bb04f..c7f5bcae 100644
--- a/spec/html-proofer/fixtures/links/unicode_domain.html
+++ b/spec/html-proofer/fixtures/links/unicode_domain.html
@@ -2,6 +2,6 @@
-A real link
+A real link
diff --git a/spec/html-proofer/fixtures/vcr_cassettes/links/unicode_domain_html_log_level_error_type_file_.yml b/spec/html-proofer/fixtures/vcr_cassettes/links/unicode_domain_html_log_level_error_type_file_.yml
index 26cfc5e6..2bc6e344 100644
--- a/spec/html-proofer/fixtures/vcr_cassettes/links/unicode_domain_html_log_level_error_type_file_.yml
+++ b/spec/html-proofer/fixtures/vcr_cassettes/links/unicode_domain_html_log_level_error_type_file_.yml
@@ -2,13 +2,13 @@
http_interactions:
- request:
method: head
- uri: https://xn--mxaaaiil1bdgepgr1bpt0d.gr/
+ uri: https://xn--gran-8qa.fi
body:
encoding: US-ASCII
string: ''
headers:
User-Agent:
- - Mozilla/5.0 (compatible; HTML Proofer/3.19.3; +https://github.com/gjtorikian/html-proofer)
+ - Mozilla/5.0 (compatible; HTML Proofer/5.0.8; +https://github.com/gjtorikian/html-proofer)
Accept:
- application/xml,application/xhtml+xml,text/html;q=0.9, text/plain;q=0.8,image/png,*/*;q=0.5
Expect:
@@ -16,33 +16,39 @@ http_interactions:
response:
status:
code: 200
- message: OK
+ message: ''
headers:
- Date:
- - Fri, 31 Dec 2021 16:37:41 GMT
- Content-Type:
- - text/html; charset=UTF-8
- Connection:
- - keep-alive
- x-powered-by:
- - PHP/8.0.13
- x-turbo-charged-by:
- - LiteSpeed
- CF-Cache-Status:
- - DYNAMIC
- Report-To:
- - '{"endpoints":[{"url":"https:\/\/a.nel.cloudflare.com\/report\/v3?s=vHwiFt05sw2oM%2BxOEY0YbwWDVJ7lNnaWzqt3YVttfW73qJj%2FOicx%2F2u6FF84ul4dFoAdtntXjMsZhHCWx2Q%2F2%2B842ECjjDJOLjPrkVdaHxBaMKnuzqJWI8j2aXzIKuRDADaWiYS6XcGYTXuEAXZg1g%3D%3D"}],"group":"cf-nel","max_age":604800}'
- NEL:
- - '{"success_fraction":0,"report_to":"cf-nel","max_age":604800}'
- Server:
- - cloudflare
- CF-RAY:
- - 6c64ef1218b18c4b-EWR
+ accept-ranges:
+ - bytes
+ access-control-allow-origin:
+ - "*"
+ age:
+ - '2084372'
+ cache-control:
+ - public, max-age=0, must-revalidate
+ content-disposition:
+ - inline
+ content-type:
+ - text/html; charset=utf-8
+ date:
+ - Mon, 06 May 2024 22:16:55 GMT
+ etag:
+ - '"17bf20f1307387d12393425f77714bc2"'
+ server:
+ - Vercel
+ strict-transport-security:
+ - max-age=63072000
+ x-vercel-cache:
+ - HIT
+ x-vercel-id:
+ - sfo1::jvrgm-1715033815794-3d4492f2b00f
+ content-length:
+ - '11213'
body:
encoding: ASCII-8BIT
string: ''
- http_version: '1.1'
+ http_version: '2'
adapter_metadata:
- effective_url: http://xn--mxaaaiil1bdgepgr1bpt0d.gr/
- recorded_at: Fri, 31 Dec 2021 16:37:41 GMT
+ effective_url: https://www.xn--gran-8qa.fi/
+ recorded_at: Mon, 06 May 2024 22:16:55 GMT
recorded_with: VCR 2.9.3