kataba 1.1.3 → 1.1.4

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 0c1a21b706cc4e327925cd9d936247e6f5ce17ea3964853f1b442e5ee8ecae3b
-  data.tar.gz: 11ad87d5a2baede9403e4b83e2e5db01f5f2b8fe9702371228018cb96534ec22
+  metadata.gz: 8ae88ae5dafc92e24f90d773300f44d096a5efb5d42e82c047e48c413ffcf852
+  data.tar.gz: 6954878cb4e2fb52646f294efda7a53dd0c11ac97e9a08554691d298ac5e961e
 SHA512:
-  metadata.gz: e2bdb712cd9c063469ee35b6dd5c3f1321018fc490ed25cb095c63715aaf63829915db1b99bfac94d87e574c88e301145fde7359cc57139dd1a997d61954bddc
-  data.tar.gz: d21c6a7c83b2a93c89115305856cf2824a62cfa75e7aaa32bda7ea0f4d648bb4d8e6e4db3d4e45ea734a2bc495ea33d3737eb62bc85b3fd076958375d3d0e6fa
+  metadata.gz: 4be9c7f61034d3d0dbabcb65ccd56169f1f1f7e5aafeaa583a6cd5ab5cf348a5dac1c378d44a715f8b2aeecac8059a1af39a47a1cce53120bf2a5f3153d20630
+  data.tar.gz: bf9b63fcf72d6ae6c4ee1e515ff81854ae4f88d79b0a4496bdc2bdccd63fce49bcded29d8eb0fddc7c22f83860956716c7c114d101fe4e18a5b02e5f2e735c7e

data/lib/kataba/fetcher.rb

@@ -87,7 +87,7 @@ module Kataba
     def request_with_redirects(uri, redirect_depth:)
       raise FetchError, "too many redirects fetching #{@original_uri}" if redirect_depth > MAX_REDIRECTS
 
-      parsed = URI.parse(uri)
+      parsed = parse_http_uri(uri)
       http = Net::HTTP.new(parsed.host, parsed.port)
       http.use_ssl = (parsed.scheme == 'https')
       http.open_timeout = @open_timeout
@@ -116,6 +116,17 @@ module Kataba
       end
     end
 
+    # A bare filename parses as a URI with no host, and Net::HTTP then fails
+    # with NoMethodError, which no consumer's fetch-failure branch rescues.
+    def parse_http_uri(uri)
+      parsed = URI.parse(uri)
+      return parsed if parsed.is_a?(URI::HTTP) && parsed.host
+
+      raise FetchError, "not an absolute http(s) URI: #{uri}"
+    rescue URI::InvalidURIError => e
+      raise FetchError, "not a valid URI: #{uri} (#{e.message})"
+    end
+
     def resolve_redirect(from, location)
       raise FetchError, "redirect with no Location header from #{from}" if location.nil? || location.empty?
 

data/lib/kataba.rb

@@ -3,6 +3,8 @@ require 'tmpdir'
 require 'digest/md5'
 require 'fileutils'
 require 'yaml'
+require 'uri'
+require 'set'
 require 'kataba/fetcher'
 
 module Kataba
@@ -116,12 +118,17 @@ module Kataba
 
   private
 
-    def self.download_xsd(xsd_uri_array)
+    def self.download_xsd(xsd_uri_array, visited = Set.new)
       new_xsd_uris = []
       file_paths = []
 
       # Download files
       xsd_uri_array.each do |xsd_uri|
+        # Schemas may import each other (tdwg_dwc_simple.xsd and
+        # tdwg_dwc_extensions.xsd do), so fetch each URI once per call or the
+        # recursion never ends.
+        next unless visited.add?(xsd_uri)
+
         uri_md5 = Digest::MD5.hexdigest(xsd_uri)
 
         dir_name = "#{self.configuration.offline_storage}"
@@ -137,18 +144,22 @@ module Kataba
         # Resolve mirror first, if configured, then fetch. Fetching BEFORE
         # opening tmp_path means a failed network fetch can't leave an
         # orphaned 0-byte .part on disk — the file simply isn't created.
-        fetch_uri = xsd_uri
+        mirror = nil
         if !self.configuration.mirror_list.to_s.empty?
           # YAML.load_file returns nil for a comments-only or empty file.
           # Treat that as "no mirror configured" rather than NoMethodError.
           mirror_list = YAML.load_file(self.configuration.mirror_list) || {}
-          mirror = mirror_list[xsd_uri]
-          fetch_uri = mirror unless mirror.to_s.empty?
+          mirror = mirror_list[xsd_uri].to_s
+          mirror = nil if mirror.empty?
         end
 
-        body = Kataba::Fetcher.new(fetch_uri,
-                                   open_timeout: self.configuration.open_timeout,
-                                   read_timeout: self.configuration.read_timeout).fetch
+        body = if mirror && !mirror.match?(%r{\Ahttps?://}i)
+                 read_local_mirror(mirror)
+               else
+                 Kataba::Fetcher.new(mirror || xsd_uri,
+                                     open_timeout: self.configuration.open_timeout,
+                                     read_timeout: self.configuration.read_timeout).fetch
+               end
 
         # Write to a .part file first; only rename to the final cache path
         # after we've confirmed the bytes parse as XML. Without this, a
@@ -165,20 +176,20 @@ module Kataba
         end
 
         File.rename(tmp_path, file_path)
-        file_paths << file_path
+        file_paths << [file_path, xsd_uri]
       end
 
       # Search inside for other schemaLocations
-      file_paths.each do |file_path|
-        new_xsd_uris = find_schemas(file_path)
+      file_paths.each do |file_path, xsd_uri|
+        new_xsd_uris.concat(find_schemas(file_path, xsd_uri))
       end
 
       if !new_xsd_uris.reject(&:empty?).empty?
-        download_xsd(new_xsd_uris)
+        download_xsd(new_xsd_uris, visited)
       end
     end
 
-    def self.find_schemas(xml_file_path)
+    def self.find_schemas(xml_file_path, base_uri)
       xsd_uri_array = []
 
       # Open XML file
@@ -186,10 +197,11 @@ module Kataba
       # search for schemaLocations
       doc.xpath("//@schemaLocation").each do |node|
         if !node.value.to_s.empty?
+          schema_uri = resolve_schema_location(node.value, base_uri)
           # Add to array
-          xsd_uri_array << node.value
+          xsd_uri_array << schema_uri
           # Get MD5
-          uri_md5 = Digest::MD5.hexdigest(node.value)
+          uri_md5 = Digest::MD5.hexdigest(schema_uri)
           # Reassign attribute value
           node.value = "#{uri_md5}.xsd"
         end
@@ -200,4 +212,29 @@ module Kataba
 
       return xsd_uri_array
     end
+
+    # A mirror value that is not an http(s) URL is a file, so a consumer can
+    # ship a corrected schema with its own code. A relative path resolves
+    # against the mirror list's directory, never the working directory, which
+    # fetch_schema changes while it parses.
+    def self.read_local_mirror(path)
+      full_path = File.expand_path(path, File.dirname(self.configuration.mirror_list))
+      File.read(full_path)
+    rescue SystemCallError => e
+      raise Kataba::Fetcher::FetchError, "could not read mirror file #{full_path} (#{e.class})"
+    end
+
+    # An absolute location is returned byte for byte. Its MD5 is the cache
+    # filename and the mirror_list key, so normalising it would orphan every
+    # cached schema and miss every mirror. A relative one resolves against the
+    # canonical URI of the schema importing it, never a mirror or a redirect
+    # target, so its cache name does not depend on where the bytes came from.
+    def self.resolve_schema_location(location, base_uri)
+      return location if URI.parse(location).absolute?
+
+      URI.join(base_uri, location).to_s
+    rescue URI::Error
+      # Left as written, so the Fetcher reports it as a FetchError.
+      location
+    end
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: kataba
 version: !ruby/object:Gem::Version
-  version: 1.1.3
+  version: 1.1.4
 platform: ruby
 authors:
 - David Cliff
@@ -14,56 +14,56 @@ dependencies:
   name: nokogiri
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - '>='
+    - - ">="
       - !ruby/object:Gem::Version
         version: '1.19'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - '>='
+    - - ">="
       - !ruby/object:Gem::Version
         version: '1.19'
 - !ruby/object:Gem::Dependency
   name: rake
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - '>='
+    - - ">="
       - !ruby/object:Gem::Version
         version: '0'
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - '>='
+    - - ">="
       - !ruby/object:Gem::Version
         version: '0'
 - !ruby/object:Gem::Dependency
   name: minitest
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - '>='
+    - - ">="
       - !ruby/object:Gem::Version
         version: '0'
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - '>='
+    - - ">="
       - !ruby/object:Gem::Version
         version: '0'
 - !ruby/object:Gem::Dependency
   name: webmock
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - '>='
+    - - ">="
       - !ruby/object:Gem::Version
         version: '0'
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - '>='
+    - - ">="
       - !ruby/object:Gem::Version
         version: '0'
 description: Kataba allows for mirroring and offline storage of XSD files, to enhance
@@ -85,16 +85,16 @@ require_paths:
 - lib
 required_ruby_version: !ruby/object:Gem::Requirement
   requirements:
-  - - '>='
+  - - ">="
     - !ruby/object:Gem::Version
       version: '3.2'
 required_rubygems_version: !ruby/object:Gem::Requirement
   requirements:
-  - - '>='
+  - - ">="
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.0.9
+rubygems_version: 3.2.33
 signing_key:
 specification_version: 4
 summary: XML Schema Definition (XSD) mirroring and offline validation for Nokogiri