Handle new changes to site 2025-12

lib/faexport/scraper.rb

@@ -1318,7 +1318,7 @@
       rescue OpenURI::HTTPError => e
         $http_errors.increment(labels: { page_type: page_type })
         # Detect and handle known errors
-        if e.io.status[0] == "403" || e.io.status[0] == "503"
+        if e.io.status[0] == "403" || e.io.status[0] == "503" || e.io.status[0] == "400"
           raw = e.io.read
           html = Nokogiri::HTML(raw.encode("UTF-8", invalid: :replace, undef: :replace).delete("\000"))
 
@@ -1334,6 +1334,18 @@
             $slowdown_errors.increment(labels: { page_type: page_type })
             raise FASlowdownError.new(url)
           end
+
+          # Handle user not found errors
+          if e.io.status[0] == "400"
+            head = html.xpath("//head//title").first
+            if head.content == "System Error"
+              error_msg = html.at_css("table.maintable td.alt1 font").content
+              # Handle user profile not found, and user not found on journal listing
+              if error_msg.include?("This user cannot be found") || error_msg.include?("User not found!")
+                raise FANoUserError.new(url)
+              end
+            end
+          end
         end
         # Retry some types of error
         if e.io.status[0] == "502" || e.io.status[0] == "520"
@@ -1428,6 +1440,9 @@
       if maintable_content.include?("has voluntarily disabled access to their account and all of its contents.")
         raise FAAccountDisabledError.new(url)
       end
+      if maintable_content.include?("Access has been disabled to the account and contents of user")
+        raise FAAccountDisabledError.new(url)
+      end
 
       # Handle user not existing (this version of the error is raised by watchers lists and galleries)
       if maintable_content.include?("Provided username not found in the database.") ||

tests/integration/check_helper.rb

@@ -38,7 +38,7 @@
 RSpec::Matchers.define :be_valid_date_and_match_iso do |iso_string|
   match do |date_string|
     expect(date_string).not_to be_blank
-    expect(date_string).to match(/[A-Z][a-z]{2} [0-9]+([a-z]{2})?, [0-9]{4},? [0-9]{2}:[0-9]{2}( ?[AP]M)?/)
+    expect(date_string).to match(/[A-Z][a-z]{2,} [0-9]+([a-z]{2})?, [0-9]{4},? [0-9]{2}:[0-9]{2}(:[0-9]{2})?( ?[AP]M)?/)
     expect(iso_string).not_to be_blank
     expect(iso_string).to eql(Time.parse("#{date_string} UTC").iso8601)
   end