Darcs Support: Parses logs of newer version of darcs. Use hashes as t…

…okens instead of patch names.

lib/ohloh_scm/adapters/darcs/cat_file.rb

@@ -10,7 +10,7 @@ def cat_file_parent(commit, diff)
 		end
 
 		def cat(revision, path)
-			out, err = run_with_err("cd '#{url}' && darcs show contents -p '#{revision}' #{escape(path)}")
+			out, err = run_with_err("cd '#{url}' && darcs show contents -h '#{revision}' #{escape(path)}")
 			# show contents gives no error for non-existent paths
 			#return nil if err =~ /No such file in rev/
 			raise RuntimeError.new(err) unless err.to_s == ''

lib/ohloh_scm/adapters/darcs/commits.rb

@@ -2,19 +2,19 @@ module OhlohScm::Adapters
 	class DarcsAdapter < AbstractAdapter
 
 		# Return the number of commits in the repository following +since+.
-		def commit_count(since=nil)
-			commit_tokens(since).size
+		def commit_count(opts={})
+			commit_tokens(opts).size
 		end
 
 		# Return the list of commit tokens following +since+.
-		def commit_tokens(since=nil, up_to=nil)
-      from = since ? " --from-patch #{since}" : ""
-      to   = up_to ? " --to-patch #{up_to}" : ""
-			tokens = string_to_patch_names(run("cd '#{self.url}' && darcs changes#{from}#{to}")).reverse
+		def commit_tokens(opts={})
+			after = opts[:after] ? " --from-match 'hash #{opts[:after]}'" : ""
+			up_to = opts[:up_to] ? " --to-match 'hash #{opts[:up_to]}'" : ""
+			tokens = string_to_patch_tokens(run("cd '#{self.url}' && darcs changes#{after}#{up_to}")).reverse
 
 			# Darcs returns everything after *and including* since.
 			# We want to exclude it.
-			if tokens.any? && tokens.first == since
+			if tokens.any? && tokens.first == opts[:after]
 				tokens[1..-1]
 			else
 				tokens
@@ -25,11 +25,11 @@ def commit_tokens(since=nil, up_to=nil)
 		# Not including the diffs is meant to be a memory savings when we encounter massive repositories.
 		# If you need all commits including diffs, you should use the each_commit() iterator, which only holds one commit
 		# in memory at a time.
-		def commits(since=nil)
-			from = since ? " --from-patch #{since}" : ""
-			log = run("cd '#{self.url}' && darcs changes#{from} --reverse")
+		def commits(opts={})
+			after = opts[:after] ? " --from-match 'hash #{opts[:after]}'" : ""
+			log = run("cd '#{self.url}' && darcs changes#{after} --reverse")
 			a = OhlohScm::Parsers::DarcsParser.parse(log)
-			if a.any? && a.first.token == since
+			if a.any? && a.first.token == opts[:after]
 				a[1..-1]
 			else
 				a
@@ -38,40 +38,41 @@ def commits(since=nil)
 
 		# Returns a single commit, including its diffs
 		def verbose_commit(token)
-			log = run("cd '#{self.url}' && darcs changes -v -p '#{token}'")
+			log = run("cd '#{self.url}' && darcs changes -v -h '#{token}'")
 			OhlohScm::Parsers::DarcsParser.parse(log).first
 		end
 
 		# Yields each commit after +since+, including its diffs.
 		# The log is stored in a temporary file.
 		# This is designed to prevent excessive RAM usage when we encounter a massive repository.
 		# Only a single commit is ever held in memory at once.
-		def each_commit(since=nil)
-			open_log_file(since) do |io|
+		def each_commit(opts={})
+			open_log_file(opts) do |io|
 				OhlohScm::Parsers::DarcsParser.parse(io) do |commit|
-					yield commit if block_given? && commit.token != since
+					yield commit if block_given? && commit.token != opts[:after]
 				end
 			end
 		end
 
 		# Not used by Ohloh proper, but handy for debugging and testing
-		def log(since=nil)
-      from = since ? " --from-patch #{since}" : ""
-			run "cd '#{url}' && darcs changes -s#{from}"
+		def log(opts={})
+      after = opts[:after] ? " --from-match 'hash #{opts[:after]}'" : ""
+			run "cd '#{url}' && darcs changes -s#{after}"
 		end
 
 		# Returns a file handle to the log.
 		# In our standard, the log should include everything AFTER +since+. However, darcs doesn't work that way;
 		# it returns everything after and INCLUDING +since+. Therefore, consumers of this file should check for
 		# and reject the duplicate commit.
-		def open_log_file(since=nil)
+		def open_log_file(opts={})
+			after = opts[:after] ? " --from-match 'hash #{opts[:after]}'" : ''
 			begin
-				if since == head_token # There are no new commits
+				if opts[:after] == head_token # There are no new commits
 					# As a time optimization, just create an empty file rather than fetch a log we know will be empty.
 					File.open(log_filename, 'w') { }
 				else
-          from = since ? " --from-patch #{since}" : ""
-					run "cd '#{url}' && darcs changes --reverse -v#{from} > #{log_filename}"
+					after = opts[:after] ? " --from-match 'hash #{opts[:after]}'" : ""
+					run "cd '#{url}' && darcs changes --reverse -v#{after} > #{log_filename}"
 				end
 				File.open(log_filename, 'r') { |io| yield io }
 			ensure

lib/ohloh_scm/adapters/darcs/head.rb

@@ -1,23 +1,23 @@
 module OhlohScm::Adapters
 	class DarcsAdapter < AbstractAdapter
 		def head_token
-		  string_to_patch_names(run("cd '#{url}' && darcs changes --last 1"))[0]
+		  string_to_patch_tokens(run("cd '#{url}' && darcs changes --last 1"))[0]
 		end
 
 		def head
 			verbose_commit(head_token)
 		end
 
 		def parent_tokens(commit)
-		  string_to_patch_names(run("cd '#{url}' && darcs changes --to-patch #{commit.token}"))[1..-1]
+		  string_to_patch_tokens(run("cd '#{url}' && darcs changes --to-match 'hash #{commit.token}'"))[1..-1]
 		end
 
 		def parents(commit)
 			parent_tokens(commit).map {|token| verbose_commit(token)}
 		end
 
-		def string_to_patch_names(s)
-		  s.split(/\n/).select {|s| s =~ /^  \* /}.map {|s| s.sub(/^  \* /,'')}
+		def string_to_patch_tokens(s)
+		  s.split(/\n/).select {|s| s =~ /^patch /}.map {|s| s.sub(/^patch /,'')}
 		end
 	end
 end

lib/ohloh_scm/adapters/darcs/misc.rb

@@ -10,11 +10,11 @@ def exist?
 		end
 
 		def ls_tree(token)
-			run("cd '#{path}' && darcs show files -p '#{token}'").split("\n")
+			run("cd '#{path}' && darcs show files --no-pending -h '#{token}'").split("\n")
 		end
 
 		def export(dest_dir, token=nil)
-			p = token ? " -p '#{token}'" : ""
+			p = token ? " -h '#{token}'" : ""
 			run("cd '#{path}' && darcs dist#{p} && mv darcs.tar.gz '#{dest_dir}'")
 		end
 	end

lib/ohloh_scm/adapters/darcs/patch.rb

@@ -1,7 +1,7 @@
 module OhlohScm::Adapters
   class DarcsAdapter < AbstractAdapter
     def patch_for_commit(commit)
-      run("cd '#{url}' && darcs changes -p'#{commit.token}' -v")
+      run("cd '#{url}' && darcs changes -h'#{commit.token}' -v")
     end
   end
 end

lib/ohloh_scm/commit.rb

@@ -34,7 +34,7 @@ class Commit
 		# For Git, the token is the commit SHA1 hash.
 		# For CVS, which does not support atomic commits with unique IDs, we use
 		# the approximate timestamp of the change.
-		# For Darcs, the token is the patch name, and it may not be unique. XXX
+		# For Darcs, the hash will be used as the token here.
 		attr_accessor :token
 
 		# A pointer back to the adapter that contains this commit.

lib/ohloh_scm/parsers/darcs_parser.rb

@@ -8,34 +8,37 @@ def self.scm
 		def self.internal_parse(buffer, opts)
 			e = nil
 			state = :patch
+#			email_match = Regexp.new(Regexp.quote('/^(?!(?>(?1)"?(?>\\\[ -~]|[^"])"?(?1)){255,})(?!(?>(?1)"?(?>\\\[ -~]|[^"])"?(?1)){65,}@)((?>(?>(?>((?>(?>(?>\x0D\x0A)?[\x09 ])+|(?>[\x09 ]*\x0D\x0A)?[\x09 ]+)?)(\((?>(?2)(?>[\x01-\x08\x0B\x0C\x0E-\'*-\[\]-\x7F]|\\\[\x00-\x7F]|(?3)))*(?2)\)))+(?2))|(?2))?)([!#-\'*+\/-9=?^-~-]+|"(?>(?2)(?>[\x01-\x08\x0B\x0C\x0E-!#-\[\]-\x7F]|\\\[\x00-\x7F]))*(?2)")(?>(?1)\.(?1)(?4))*(?1)@(?!(?1)[a-z0-9-]{64,})(?1)(?>([a-z0-9](?>[a-z0-9-]*[a-z0-9])?)(?>(?1)\.(?!(?1)[a-z0-9-]{64,})(?1)(?5)){0,126}|\[(?:(?>IPv6:(?>([a-f0-9]{1,4})(?>:(?6)){7}|(?!(?:.*[a-f0-9][:\]]){8,})((?6)(?>:(?6)){0,6})?::(?7)?))|(?>(?>IPv6:(?>(?6)(?>:(?6)){5}:|(?!(?:.*[a-f0-9]:){6,})(?8)?::(?>((?6)(?>:(?6)){0,4}):)?))?(25[0-5]|2[0-4][0-9]|1[0-9]{2}|[1-9]?[0-9])(?>\.(?9)){3}))\])(?1)$/isD'))
 
 			buffer.each_line do |l|
 				#print "\n#{state}"
 				next_state = state
 				if state == :patch
 					case l
-					when /^([^ ]...........................)  (.*)$/
+					when /^patch ([0-9a-f]*)/
 						yield e if e && block_given?
 						e = Scm::Commit.new
-						e.author_date = Time.parse($1).utc
-						nameemail = $2
+						e.diffs = []
+						e.token = $1
+					when /^Author: (.*)/
+						nameemail = $1
 						case nameemail
-						when /^([^<]*) <(.*)>$/
-						  e.author_name = $1
-						  e.author_email = $2
+						when /^(\b.*) <([^>]*)>/
+							e.author_name = $1
+							e.author_email = $2
 						when /^([^@]*)$/
-						  e.author_name = $1
-						  e.author_email = nil
+							e.author_name = $1
+							e.author_email = nil
 						else
-						  e.author_name = nil
-						  e.author_email = nameemail
+							e.author_name = nil
+							e.author_email = nameemail
 						end
-						e.diffs = []
+					when /^Date:   ([^ ]...........................)/
+						e.author_date = Time.parse($1).utc
 					when /^  \* (.*)/
-						e.token = ($1 || '')
+						e.message = ($1 || '')
 						next_state = :long_comment_or_prims
 					end
-
 				elsif state == :long_comment_or_prims
 					case l
 					when /^    addfile\s+(.+)/

test/data/darcs_patch.diff

@@ -1,4 +1,6 @@
-Wed Nov  3 15:49:53 PDT 2010  Simon Michael <[email protected]>
+patch bd7e455d648b784ce4be2db26a4e62dfe734dd66
+Author: Simon Michael <[email protected]>
+Date:   Wed Nov  3 18:49:53 EDT 2010
   * add helloworld.c
     addfile ./helloworld.c
     hunk ./helloworld.c 1

test/unit/darcs_cat_file_test.rb

@@ -21,9 +21,9 @@ def test_cat_file
 EXPECTED
 
 				# The file was deleted by the "remove..." patch. Check that it does not exist now, but existed in parent.
-				assert_equal nil, darcs.cat_file(Scm::Commit.new(:token => 'remove helloworld.c'), Scm::Diff.new(:path => 'helloworld.c'))
-				assert_equal expected, darcs.cat_file_parent(Scm::Commit.new(:token => 'remove helloworld.c'), Scm::Diff.new(:path => 'helloworld.c'))
-				assert_equal expected, darcs.cat_file(Scm::Commit.new(:token => 'add helloworld.c'), Scm::Diff.new(:path => 'helloworld.c'))
+				assert_equal nil, darcs.cat_file(Scm::Commit.new(:token => '1007b5ad4831769283213d47e1fd5f6d30ac97f0'), Scm::Diff.new(:path => 'helloworld.c'))
+				assert_equal expected, darcs.cat_file_parent(Scm::Commit.new(:token => '1007b5ad4831769283213d47e1fd5f6d30ac97f0'), Scm::Diff.new(:path => 'helloworld.c'))
+				assert_equal expected, darcs.cat_file(Scm::Commit.new(:token => 'bd7e455d648b784ce4be2db26a4e62dfe734dd66'), Scm::Diff.new(:path => 'helloworld.c'))
 			end
 		end
 

test/unit/darcs_commits_test.rb

@@ -6,17 +6,17 @@ class DarcsCommitsTest < OhlohScm::Test
 		def test_commit
 			with_darcs_repository('darcs') do |darcs|
 				assert_equal 2, darcs.commit_count
-				assert_equal 1, darcs.commit_count('add helloworld.c')
-				assert_equal 0, darcs.commit_count('remove helloworld.c')
-				assert_equal ['add helloworld.c', 'remove helloworld.c'], darcs.commit_tokens
-				assert_equal ['remove helloworld.c'], darcs.commit_tokens('add helloworld.c')
-				assert_equal [], darcs.commit_tokens('remove helloworld.c')
-				assert_equal ['add helloworld.c',
-											'remove helloworld.c'], darcs.commits.collect { |c| c.token }
-				assert_equal ['remove helloworld.c'], darcs.commits('add helloworld.c').collect { |c| c.token }
+				assert_equal 1, darcs.commit_count(:after => 'bd7e455d648b784ce4be2db26a4e62dfe734dd66')
+				assert_equal 0, darcs.commit_count(:after => '1007b5ad4831769283213d47e1fd5f6d30ac97f0')
+				assert_equal ['bd7e455d648b784ce4be2db26a4e62dfe734dd66', '1007b5ad4831769283213d47e1fd5f6d30ac97f0'], darcs.commit_tokens
+				assert_equal ['1007b5ad4831769283213d47e1fd5f6d30ac97f0'], darcs.commit_tokens(:after => 'bd7e455d648b784ce4be2db26a4e62dfe734dd66')
+				assert_equal [], darcs.commit_tokens(:after => '1007b5ad4831769283213d47e1fd5f6d30ac97f0')
+				assert_equal ['bd7e455d648b784ce4be2db26a4e62dfe734dd66',
+											'1007b5ad4831769283213d47e1fd5f6d30ac97f0'], darcs.commits.collect { |c| c.token }
+				assert_equal ['1007b5ad4831769283213d47e1fd5f6d30ac97f0'], darcs.commits(:after => 'bd7e455d648b784ce4be2db26a4e62dfe734dd66').collect { |c| c.token }
 				# Check that the diffs are not populated
-				assert_equal [], darcs.commits('add helloworld.c').first.diffs
-				assert_equal [], darcs.commits('remove helloworld.c')
+				assert_equal [], darcs.commits(:after => 'bd7e455d648b784ce4be2db26a4e62dfe734dd66').first.diffs
+				assert_equal [], darcs.commits(:after => '1007b5ad4831769283213d47e1fd5f6d30ac97f0')
 			end
 		end
 
@@ -37,8 +37,8 @@ def test_each_commit
 				assert !FileTest.exist?(darcs.log_filename) # Make sure we cleaned up after ourselves
 
 				# Verify that we got the commits in forward chronological order
-				assert_equal ['add helloworld.c',
-											'remove helloworld.c'], commits.map {|c| c.token}
+				assert_equal ['bd7e455d648b784ce4be2db26a4e62dfe734dd66',
+											'1007b5ad4831769283213d47e1fd5f6d30ac97f0'], commits.map {|c| c.token}
 			end
 		end
 	end

test/unit/darcs_head_test.rb

@@ -5,11 +5,11 @@ class DarcsHeadTest < OhlohScm::Test
 
 		def test_head_and_parents
 			with_darcs_repository('darcs') do |darcs|
-				assert_equal 'remove helloworld.c', darcs.head_token
-				assert_equal 'remove helloworld.c', darcs.head.token
+				assert_equal '1007b5ad4831769283213d47e1fd5f6d30ac97f0', darcs.head_token
+				assert_equal '1007b5ad4831769283213d47e1fd5f6d30ac97f0', darcs.head.token
 				assert darcs.head.diffs.any? # diffs should be populated
 
-				assert_equal 'add helloworld.c', darcs.parents(darcs.head).first.token
+				assert_equal 'bd7e455d648b784ce4be2db26a4e62dfe734dd66', darcs.parents(darcs.head).first.token
 				assert darcs.parents(darcs.head).first.diffs.any?
 			end
 		end

test/unit/darcs_misc_test.rb

@@ -14,14 +14,14 @@ def test_exist
 
 		def test_ls_tree
 			with_darcs_repository('darcs') do |darcs|
-				assert_equal ['.','./helloworld.c'], darcs.ls_tree('add helloworld.c').sort
+				assert_equal ['.','./helloworld.c'], darcs.ls_tree('bd7e455d648b784ce4be2db26a4e62dfe734dd66').sort
 			end
 		end
 
 		def test_export
 			with_darcs_repository('darcs') do |darcs|
 				Scm::ScratchDir.new do |dir|
-					darcs.export(dir, 'add helloworld.c')
+					darcs.export(dir, 'bd7e455d648b784ce4be2db26a4e62dfe734dd66')
 					assert_equal ['.', '..', 'darcs.tar.gz'], Dir.entries(dir).sort
 				end
 			end