[whimsy] branch master updated: Add privacy check

[se...@apache.org]

This is an automated email from the ASF dual-hosted git repository.

sebb pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/whimsy.git


The following commit(s) were added to refs/heads/master by this push:
     new 28af1a52 Add privacy check
28af1a52 is described below

commit 28af1a526380ec193d27b9bb31ce31c746295733
Author: Sebb <se...@apache.org>
AuthorDate: Fri Apr 29 13:18:31 2022 +0100

    Add privacy check
---
 lib/whimsy/sitestandards.rb | 9 +++++++++
 tools/site-scan.rb          | 4 ++--
 2 files changed, 11 insertions(+), 2 deletions(-)

diff --git a/lib/whimsy/sitestandards.rb b/lib/whimsy/sitestandards.rb
index 0f97d32c..769b5c82 100644
--- a/lib/whimsy/sitestandards.rb
+++ b/lib/whimsy/sitestandards.rb
@@ -110,6 +110,15 @@ module SiteStandards
       CHECK_DOC => 'All website content SHOULD include a copyright notice for the ASF.',
     },
 
+    'privacy' => { # link_check
+      CHECK_TEXT => %r{Privacy Policy}i,
+      CHECK_CAPTURE => %r{(Privacy)}i,
+      CHECK_VALIDATE => %r{\Ahttps://privacy.apache.org/policies/privacy-policy-public.html\z}i,
+      CHECK_TYPE => true,
+      CHECK_POLICY => 'https://lists.apache.org/thread/bz0lv2wlh90t58n73wr85qoof7vdvdxt',
+      CHECK_DOC => 'All websites must link to the Privacy Policy.',
+    },
+
     'image' => { # Custom: merely looks in IMAGE_DIR for #{id}.*
       CHECK_TEXT => nil,
       CHECK_CAPTURE => nil,
diff --git a/tools/site-scan.rb b/tools/site-scan.rb
index 4aa8063a..7e76cd51 100755
--- a/tools/site-scan.rb
+++ b/tools/site-scan.rb
@@ -108,7 +108,7 @@ def parse(id, site, name)
       end
     end
 
-    %w(thanks security sponsorship).each do |check|
+    %w(thanks security sponsorship privacy).each do |check|
       if a_text =~ SiteStandards::COMMON_CHECKS[check][SiteStandards::CHECK_CAPTURE]
         begin
           data[check.to_sym] = uri + a_href
@@ -157,7 +157,7 @@ $cache = Cache.new(dir: 'site-scan')
 $verbose = ARGV.delete '--verbose'
 
 # USAGE:
-# site-scan.rb https://whimsical.apache.org [whimsy] [whimsy-scan.json] - to scan one project
+# site-scan.rb https://whimsical.apache.org [Whimsy] [whimsy-scan.json] - to scan one project
 # site-scan.rb [project-output.json] [podlings-output.json] [projname podlingname ...]
 # If additional projname|podlingname are provided, only scans those sites
 if ARGV.first =~ /^https?:\/\/\w/