vcs2json 1.0.0 → 1.0.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 2800a945a5ca9ced91f1dc9b6e7dfeea9dd035d2
-  data.tar.gz: dec9e8a47a8a76d552d13a7946afc8050a863249
+  metadata.gz: 13238e079776d7f77e6f301c9778f1cd93383e7c
+  data.tar.gz: ad0e55fbe2d9b7bba56d343fb51c8e2a4d08ed6c
 SHA512:
-  metadata.gz: 1491d0a803e27febfe7e2cc04e569e5ebc02af45b104be8812ae6bc4d44ced3008ab3944b27557c78ff22d4e54b212c8ffb36276b57eb113aec0291fd0d36b93
-  data.tar.gz: 30df5f5ffc7f6973a03d0be6ca7bcf7f9f9b48a9bb3c2b666b2c0ac2f63bd96e8e27dc19e33e52534194b8159d4ba4aa7b345ee24bf6a912022284a94933911c
+  metadata.gz: 77ad00f756a282c3b97fd471f1628bf78048f9fa2416ac0ae5b09e8ea430706d1edd67a97f2e19c7d6a88152446b5acde560135076d8c399b4d62386be083c95
+  data.tar.gz: 0ecea201b2eecbeae823c4a3fb3d31e37de37a07cd5b4d602efc7bd631f33c8bccf58e18e303e372a64641c28088e57d03bac1f3fcb8f3aeaa53c50ea926564b

data/lib/exceptions/encoding_error.rb

@@ -0,0 +1,2 @@
+class EncodingError < StandardError
+end

data/lib/vcs2json/git.rb

@@ -20,22 +20,26 @@ module Vcs2Json
 
         def execute
             # recursively add commits as long as we have less than :number and there are still more commits to search
-            add_commits(@opts)
-            add_integer_mapping
+            begin
+                add_commits(@opts)
+                add_integer_mapping
 
-            # sort on date and prune excessive commits
-            sorted_and_pruned = @commits.sort_by {|id,commit| commit[:date]}.reverse.map {|(_,commit)| commit}.first(@opts[:number])
+                # sort on date and prune excessive commits
+                sorted_and_pruned = @commits.sort_by {|id,commit| commit[:date]}.reverse.map {|(_,commit)| commit}.first(@opts[:number])
 
-            # print commits to stdout as json
-            $stdout.puts JSON.pretty_generate(sorted_and_pruned)
+                # print commits to stdout as json
+                $stdout.puts JSON.pretty_generate(sorted_and_pruned)
 
-            # print ids of empty commits to stderr
-            if !@empty_commits.empty?
-                STDERR.puts "EMPTY COMMITS"
-                STDERR.puts @empty_commits
+                # print ids of empty commits to stderr
+                if !@empty_commits.empty?
+                    STDERR.puts "EMPTY COMMITS"
+                    STDERR.puts @empty_commits
+                end
+                # print additional info to stderr
+                STDERR.puts "\n\nExtracted #{sorted_and_pruned.size} commits."
+            rescue EncodingError => e
+                puts e
             end
-            # print additional info to stderr
-            STDERR.puts "\n\nExtracted #{sorted_and_pruned.size} commits."
         end
 
         def ignore
@@ -115,7 +119,24 @@ module Vcs2Json
 
         def add_meta_information(opts)
             raw_commits = `git log #{hash_2_gitoptions(opts)} --pretty=format:'%H#{FIELD_SEP}%cn#{FIELD_SEP}%ce#{FIELD_SEP}%cd#{FIELD_SEP}%ad#{FIELD_SEP}%B#{COMMIT_SEP}'`
-            commits_info = raw_commits.encode('UTF-8', :invalid => :replace).split(COMMIT_SEP)
+
+            begin
+                encoded = ''
+                # try encoding to utf8
+                encoded = raw_commits.encode('UTF-8', 'binary', invalid: :replace, undef: :replace, replace: '')
+                # need to expliceitely check if the encoding is valid for ruby <= 2.0
+                # utf8 -> utf8 will not do anything even with invalid bytes
+                # http://stackoverflow.com/questions/24036821/ruby-2-0-0-stringmatch-argumenterror-invalid-byte-sequence-in-utf-8
+                if !encoded.valid_encoding?
+                    # encode to utf16 first and then back to utf8
+                    encoded.encode!("UTF-16be", invalid: :replace, undef: :replace, :replace=>'')
+                    encoded.encode!('UTF-8', 'binary', invalid: :replace, undef: :replace, replace: '')
+                end
+                # split into individual commits
+                commits_info = encoded.split(COMMIT_SEP)
+            rescue ArgumentError
+                raise EncodingError.new, "Unable to encode input as UTF-8"
+            end
 
             commits_info.each do |commit|
                 fields = commit.split(FIELD_SEP)

data/lib/vcs2json/version.rb

@@ -1,3 +1,3 @@
 module Vcs2Json
-  VERSION = "1.0.0"
+  VERSION = "1.0.1"
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: vcs2json
 version: !ruby/object:Gem::Version
-  version: 1.0.0
+  version: 1.0.1
 platform: ruby
 authors:
 - Thomas Rolfsnes
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2016-06-08 00:00:00.000000000 Z
+date: 2016-06-10 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler
@@ -131,6 +131,7 @@ files:
 - lib/exceptions/ambigious_date.rb
 - lib/exceptions/could_not_find_format_of_date.rb
 - lib/exceptions/could_not_parse_date.rb
+- lib/exceptions/encoding_error.rb
 - lib/exceptions/no_base_path.rb
 - lib/exceptions/no_date_field.rb
 - lib/exceptions/no_file_section.rb