akita-har_logger 0.2.4 → 0.2.8

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 51fcd2da35cb17ecb365988d7975f746a717f12b002d3a20f1d0c3eab245b10c
-  data.tar.gz: 4972d841dfc7c8dcd41085137d45faedeb450836d46e6d3f14d584785d997a24
+  metadata.gz: 4c48c0c45ad4fc65a7dcfa32c042ff24039025111dead2a610c5b956e1cf24e9
+  data.tar.gz: 85becb20d70489a5d8e48d275e894b40325acf010714f1d9d8031a90ead03996
 SHA512:
-  metadata.gz: 8fdffa25d122a573e8c72526934e5e713fbdc573c0d4ebfa13383d749ea856f9b01d2e5bc43e89b1c152df64f97cd965ad54bd5a4ebab924d222a1829bbbec60
-  data.tar.gz: 8ca06a98ce89ec4e9c3e6ac7f3b10ec98f73c44aacab1af0bac8c81649f342ec721017ee57ff6590c93a24234cce9d40c9b341aa8277a39a59bc86a3a9c3484e
+  metadata.gz: 02b3c2b375f134945aa2c7e20552771bc4829c7644b3e5d04561177e4806c2f0245b8219808e4fd608bd4da379187d7d4a2bb14e3b58ca72e69b737c14c83e57
+  data.tar.gz: 40bdf185b1dda6a8b83b124599931597f4363f61470ab5253645a4ff3ba028470f50254faffab63b7bf1b3334a96a86c01eb7ac005c39c1b589d98c3899ebeb5

data/Gemfile.lock

@@ -1,7 +1,7 @@
 PATH
   remote: .
   specs:
-    akita-har_logger (0.2.4)
+    akita-har_logger (0.2.8)
 
 GEM
   remote: https://rubygems.org/

data/lib/akita/har_logger/har_utils.rb

@@ -5,12 +5,31 @@ module Akita
     class HarUtils
       # Rack apparently uses 8-bit ASCII for everything, even when the string
       # is not 8-bit ASCII. This reinterprets 8-bit ASCII strings as UTF-8.
+      #
+      # If we are unable to do this reinterpretation, return the string
+      # unchanged, but log a warning that points to the caller.
       def self.fixEncoding(v)
-        if v == nil || v.encoding != Encoding::ASCII_8BIT then
-          v
-        else
-          String.new(v).force_encoding(Encoding::UTF_8)
+        if v == nil then
+          return v
         end
+
+        if !(v.is_a? String) then
+          Rails.logger.warn "[#{caller_locations(1, 1)}] fixEncoding was not given a string. This might cause JSON serialization to fail."
+          return v
+        end
+
+        # Only re-interpret 8-bit ASCII.
+        if v.encoding != Encoding::ASCII_8BIT then
+          return v
+        end
+
+        forced = String.new(v).force_encoding(Encoding::UTF_8)
+        if forced.valid_encoding? then
+          return forced
+        end
+
+        Rails.logger.warn "[#{caller_locations(1, 1)}] Unable to fix encoding: not a valid UTF-8 string. This will likely cause JSON serialization to fail."
+        v
       end
 
       # Converts a Hash into a list of Hash objects. Each entry in the given
@@ -20,7 +39,7 @@ module Akita
         hash.reduce([]) { |accum, (k, v)|
           accum.append({
             name: fixEncoding(k),
-            value: fixEncoding(v),
+            value: fixEncoding(v.to_s),
           })
         }
       end

data/lib/akita/har_logger/http_request.rb

@@ -85,7 +85,7 @@ module Akita
           return Encoding::ISO_8859_1
         end
 
-        Encoding.default_external
+        Encoding::UTF_8
       end
 
       # Obtains the posted data from an HTTP environment.
@@ -117,8 +117,27 @@ module Akita
             # body when the request specifies UTF-8. Reinterpret the content
             # body according to what the request says it is, and re-encode into
             # UTF-8.
-            result[:text] = req.body.string.encode(Encoding::UTF_8,
-                getPostDataCharSet(env))
+            #
+            # Gracefully handle any characters that are invalid in the source
+            # encoding and characters that have no UTF-8 representation by
+            # replacing with '?'. Log a warning when this happens.
+            source = req.body.string.force_encoding(getPostDataCharSet(env))
+            utf8EncodingSuccessful = false
+            if source.valid_encoding? then
+              begin
+                result[:text] = source.encode(Encoding::UTF_8)
+                utf8EncodingSuccessful = true
+              rescue Encoding::UndefinedConversionError
+                Rails.logger.warn "[#{caller_locations(0, 1)}] Unable to losslessly convert request body from #{source.encoding} to UTF-8. Characters undefined in UTF-8 will be replaced with '?'."
+              end
+            else
+              Rails.logger.warn "[#{caller_locations(0, 1)}] Request body is not valid #{source.encoding}. Invalid characters and characters undefined in UTF-8 will be replaced with '?'."
+            end
+
+            if !utf8EncodingSuccessful then
+              result[:text] = source.encode(Encoding::UTF_8,
+                  invalid: :replace, undef: :replace, replace: '?')
+            end
           end
 
           result

data/lib/akita/har_logger/http_response.rb

@@ -74,14 +74,51 @@ module Akita
       end
 
       def getContent(headers, body)
-        # XXX Handle compression & encoding.
+        # XXX Handle compression
+        # XXX Figure out how to properly join together multi-part bodies.
 
+        # Try to convert the body into UTF-8. If this fails, assume the body is
+        # binary data.
+        # XXX TODO Take charset part of Content-Type header into account.
         text = +""
+        haveBinaryData = false
         body.each { |part|
-          # XXX Figure out how to join together multi-part bodies.
-          text << (HarUtils.fixEncoding part);
+          partStr = part.to_s
+
+          if partStr.encoding == Encoding::ASCII_8BIT then
+            # Have 8-bit ASCII data. Try to interpret as UTF-8. If this fails,
+            # treat as binary data.
+            forced = String.new(partStr).force_encoding(Encoding::UTF_8)
+            if forced.valid_encoding? then
+              text << forced
+              next
+            end
+
+            haveBinaryData = true
+            break
+          end
+
+          if !partStr.valid_encoding? then
+            # Source encoding is not valid. Treat as binary data.
+            haveBinaryData = true
+            break
+          end
+
+          # Try to re-encode as UTF-8. If this fails, treat as binary data.
+          begin
+            text << partStr.encode(Encoding::UTF_8)
+          rescue Encoding::UndefinedConversionError
+            haveBinaryData = true
+            break
+          end
         }
 
+        if haveBinaryData then
+          # TODO Encode binary body data with base64.
+          # XXX Omit for now.
+          text = ""
+        end
+
         {
           size: getBodySize(body),
 
@@ -116,7 +153,7 @@ module Akita
         # for the CRLF on the blank line.
         headers.reduce(status_length + 2) { |accum, (k, v)|
           # Header-Name: header value<CR><LF>
-          accum + k.length + 2 + v.length + 2
+          accum + k.length + 2 + v.to_s.length + 2
         }
       end
 

data/lib/akita/har_logger/version.rb

@@ -2,6 +2,6 @@
 
 module Akita
   module HarLogger
-    VERSION = "0.2.4"
+    VERSION = "0.2.8"
   end
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: akita-har_logger
 version: !ruby/object:Gem::Version
-  version: 0.2.4
+  version: 0.2.8
 platform: ruby
 authors:
 - Jed Liu
-autorequire:
+autorequire: 
 bindir: exe
 cert_chain: []
-date: 2021-07-11 00:00:00.000000000 Z
+date: 2021-07-15 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rspec
@@ -54,7 +54,7 @@ licenses:
 metadata:
   homepage_uri: https://akitasoftware.com/
   source_code_uri: https://github.com/akitasoftware/akita-rails-har-logger
-post_install_message:
+post_install_message: 
 rdoc_options: []
 require_paths:
 - lib
@@ -69,8 +69,8 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.2.21
-signing_key:
+rubygems_version: 3.1.2
+signing_key: 
 specification_version: 4
 summary: Rails middleware for HAR logging
 test_files: []