pmdtester 1.1.2 → 1.2.0

data/config/projectlist_1_2_0.xsd

@@ -0,0 +1,39 @@
+<?xml version="1.0" ?>
+<!-- version 1.1.0 -->
+<xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema">
+  <xs:element name="projectlist">
+    <xs:complexType>
+      <xs:sequence>
+        <xs:element name="description" type="xs:string" minOccurs="1" maxOccurs="1"/>
+        <xs:element name="project" type="project" minOccurs="1" maxOccurs="unbounded"/>
+      </xs:sequence>
+    </xs:complexType>
+  </xs:element>
+  <xs:complexType name="project">
+    <xs:sequence>
+      <xs:element name="name" type="xs:string" minOccurs="1" maxOccurs="1"/>
+      <xs:element name="type" minOccurs="1" maxOccurs="1">
+        <xs:simpleType>
+          <xs:restriction base="xs:string">
+            <xs:enumeration value="git"/>
+          </xs:restriction>
+        </xs:simpleType>
+      </xs:element>
+      <xs:element name="connection" type="xs:string" minOccurs="1" maxOccurs="1"/>
+      <xs:element name="webview-url" type="xs:string" minOccurs="0" maxOccurs="1"/>
+      <xs:element name="tag" type="xs:string" minOccurs="0" maxOccurs="1"/>
+      <xs:element name="src-subpath" type="xs:string" minOccurs="0" maxOccurs="1" default=".">
+        <xs:annotation>
+          <xs:documentation>
+            Value of the -dir option for the PMD run.
+            The value must be a directory path relative to the root directory of the clone.
+            Defaults to just '.', ie the entire directory.
+          </xs:documentation>
+        </xs:annotation>
+      </xs:element>
+      <xs:element name="exclude-pattern" type="xs:string" minOccurs="0" maxOccurs="unbounded"/>
+      <xs:element name="build-command" type="xs:string" minOccurs="0" maxOccurs="1"/>
+      <xs:element name="auxclasspath-command" type="xs:string" minOccurs="0" maxOccurs="1"/>
+    </xs:sequence>
+  </xs:complexType>
+</xs:schema>

data/lib/pmdtester.rb

@@ -32,7 +32,7 @@ require_relative 'pmdtester/parsers/projects_parser'
 # and unexpected behaviors will not be introduced to PMD project
 # after fixing an issue and new rules can work as expected.
 module PmdTester
-  VERSION = '1.1.2'
+  VERSION = '1.2.0'
   BASE = 'base'
   PATCH = 'patch'
   PR_NUM_ENV_VAR = 'PMD_CI_PULL_REQUEST_NUMBER' # see PmdBranchDetail

data/lib/pmdtester/builders/liquid_renderer.rb

@@ -57,17 +57,74 @@ module PmdTester
       write_file("#{root}/index.html", render_liquid('project_diff_report.html', liquid_env))
       # generate array of violations in json
       write_file("#{root}/project_data.js", dump_violations_json(project))
+      # copy original pmd reports
+      copy_file("#{root}/base_pmd_report.xml", project.report_diff.base_report.file)
+      copy_file("#{root}/patch_pmd_report.xml", project.report_diff.patch_report.file)
+      # render full pmd reports
+      write_file("#{root}/base_pmd_report.html",
+                 render_liquid('project_pmd_report.html', pmd_report_liquid_env(project, BASE)))
+      write_file("#{root}/base_data.js", dump_violations_json(project, BASE))
+      write_file("#{root}/patch_pmd_report.html",
+                 render_liquid('project_pmd_report.html', pmd_report_liquid_env(project, PATCH)))
+      write_file("#{root}/patch_data.js", dump_violations_json(project, PATCH))
     end
 
-    def dump_violations_json(project)
+    def dump_violations_json(project, branch = 'diff')
+      violations_by_file = if branch == BASE
+                             project.report_diff.base_report.violations_by_file.to_h
+                           elsif branch == PATCH
+                             project.report_diff.patch_report.violations_by_file.to_h
+                           else
+                             project.report_diff.violation_diffs_by_file
+                           end
+
       h = {
         'source_link_base' => project.webview_url,
         'source_link_template' => link_template(project),
-        **violations_to_hash(project)
+        **violations_to_hash(project, violations_by_file, branch == 'diff')
       }
 
-      project_data = JSON.fast_generate(h)
+      project_data = JSON.fast_generate(h, indent: '    ', object_nl: "\n", array_nl: "\n")
       "let project = #{project_data}"
     end
+
+    private
+
+    def copy_file(target_file, source_file)
+      if File.exist? source_file
+        FileUtils.cp(source_file, target_file)
+        logger&.info "Written #{target_file}"
+      else
+        logger&.warn "File #{source_file} not found"
+      end
+    end
+
+    def pmd_report_liquid_env(project, branch)
+      report = if branch == BASE
+                 project.report_diff.base_report
+               else
+                 project.report_diff.patch_report
+               end
+      {
+        'project_name' => project.name,
+        'branch' => branch,
+        'report' => report_to_h(project, report)
+      }
+    end
+
+    def report_to_h(project, report)
+      {
+        'violation_counts' => report.violations_by_file.total_size,
+        'error_counts' => report.errors_by_file.total_size,
+        'configerror_counts' => report.configerrors_by_rule.values.flatten.length,
+
+        'execution_time' => PmdReportDetail.convert_seconds(report.exec_time),
+        'timestamp' => report.timestamp,
+
+        'rules' => report.rule_summaries,
+        'errors' => report.errors_by_file.all_values.map { |e| error_to_hash(e, project) },
+        'configerrors' => report.configerrors_by_rule.values.flatten.map { |e| configerror_to_hash(e) }
+      }
+    end
   end
 end

data/lib/pmdtester/builders/pmd_report_builder.rb

@@ -69,7 +69,7 @@ module PmdTester
                       ' -Dmaven.source.skip=true' \
                       ' -Dcheckstyle.skip=true' \
                       ' -Dpmd.skip=true' \
-                      ' -T1C'
+                      ' -T1C -B'
         Cmd.execute(package_cmd)
       end
 
@@ -119,7 +119,7 @@ module PmdTester
       doc = Nokogiri::XML(File.read(@branch_config))
       ruleset = doc.at_css('ruleset')
       ruleset.add_child("\n")
-      project.exclude_pattern.each do |exclude_pattern|
+      project.exclude_patterns.each do |exclude_pattern|
         ruleset.add_child("    <exclude-pattern>#{exclude_pattern}</exclude-pattern>\n")
       end
 

data/lib/pmdtester/builders/project_builder.rb

@@ -17,11 +17,19 @@ module PmdTester
 
       @projects.each do |project|
         logger.info "Start cloning #{project.name} repository"
-        path = project.local_source_path
-        clone_cmd = "#{project.type} clone #{project.connection} #{path}"
+        path = project.clone_root_path
+
         if File.exist?(path)
           logger.warn "Skipping clone, project path #{path} already exists"
         else
+          raise "Unsupported project type '#{project.type}' - only git is supported" unless project.type == 'git'
+
+          # git:
+          # Don't download whole history
+          # Note we don't use --single-branch, because the repo is downloaded
+          # once but may be used with several tags.
+          clone_cmd = "git clone --no-single-branch --depth 1 #{project.connection} #{path}"
+
           Cmd.execute(clone_cmd)
         end
 
@@ -38,7 +46,7 @@ module PmdTester
       logger.info 'Building projects started'
 
       @projects.each do |project|
-        path = project.local_source_path
+        path = project.clone_root_path
         Dir.chdir(path) do
           progress_logger = SimpleProgressLogger.new("building #{project.name} in #{path}")
           progress_logger.start
@@ -87,12 +95,9 @@ module PmdTester
     end
 
     def execute_reset_cmd(type, tag)
-      case type
-      when 'git'
-        reset_cmd = "git checkout #{tag}; git reset --hard #{tag}"
-      when 'hg'
-        reset_cmd = "hg up #{tag}"
-      end
+      raise "Unsupported project type '#{type}' - only git is supported" unless type == 'git'
+
+      reset_cmd = "git checkout #{tag}; git reset --hard #{tag}"
 
       Cmd.execute(reset_cmd)
     end

data/lib/pmdtester/builders/project_hasher.rb

@@ -25,24 +25,14 @@ module PmdTester
       }
     end
 
-    def errors_to_h(project)
-      errors = project.report_diff.error_diffs_by_file.values.flatten
-      errors.map { |e| error_to_hash(e, project) }
-    end
-
-    def configerrors_to_h(project)
-      configerrors = project.report_diff.configerror_diffs_by_rule.values.flatten
-      configerrors.map { |e| configerror_to_hash(e) }
-    end
-
-    def violations_to_hash(project)
+    def violations_to_hash(project, violations_by_file, is_diff)
       filename_index = []
       all_vs = []
-      project.report_diff.violation_diffs_by_file.each do |file, vs|
+      violations_by_file.each do |file, vs|
         file_ref = filename_index.size
         filename_index.push(project.get_local_path(file))
         vs.each do |v|
-          all_vs.push(make_violation_hash(file_ref, v))
+          all_vs.push(make_violation_hash(file_ref, v, is_diff))
         end
       end
 
@@ -52,36 +42,19 @@ module PmdTester
       }
     end
 
-    def link_template(project)
-      l_str = project.type == 'git' ? 'L' : 'l'
-      "#{project.webview_url}/{file}##{l_str}{line}"
-    end
-
-    def violation_type(violation)
-      if violation.changed?
-        '~'
-      elsif violation.branch == 'patch'
-        '+'
-      else
-        '-'
-      end
+    def errors_to_h(project)
+      errors = project.report_diff.error_diffs_by_file.values.flatten
+      errors.map { |e| error_to_hash(e, project) }
     end
 
-    def make_violation_hash(file_ref, violation)
-      h = {
-        't' => violation_type(violation),
-        'l' => violation.line,
-        'f' => file_ref,
-        'r' => violation.rule_name,
-        'm' => violation.changed? ? diff_fragments(violation) : violation.message
-      }
-      h['ol'] = violation.old_line if violation.changed? && violation.line != violation.old_line
-      h
+    def configerrors_to_h(project)
+      configerrors = project.report_diff.configerror_diffs_by_rule.values.flatten
+      configerrors.map { |e| configerror_to_hash(e) }
     end
 
-    def diff_fragments(violation)
-      diff = Differ.diff_by_word(violation.message, violation.old_message)
-      diff.format_as(:html)
+    def link_template(project)
+      l_str = project.type == 'git' ? 'L' : 'l'
+      "#{project.webview_url}/{file}##{l_str}{line}"
     end
 
     def error_to_hash(error, project)
@@ -122,5 +95,34 @@ module PmdTester
         'added'
       end
     end
+
+    private
+
+    def violation_type(violation)
+      if violation.changed?
+        '~'
+      elsif violation.branch == PATCH
+        '+'
+      else
+        '-'
+      end
+    end
+
+    def make_violation_hash(file_ref, violation, is_diff = TRUE)
+      h = {
+        't' => is_diff ? violation_type(violation) : '+',
+        'l' => violation.line,
+        'f' => file_ref,
+        'r' => violation.rule_name,
+        'm' => is_diff && violation.changed? ? diff_fragments(violation) : violation.message
+      }
+      h['ol'] = violation.old_line if is_diff && violation.changed? && violation.line != violation.old_line
+      h
+    end
+
+    def diff_fragments(violation)
+      diff = Differ.diff_by_word(violation.message, violation.old_message)
+      diff.format_as(:html)
+    end
   end
 end

data/lib/pmdtester/builders/rule_set_builder.rb

@@ -33,6 +33,10 @@ module PmdTester
       rule_refs
     end
 
+    def calculate_filter_set
+      output_filter_set(ALL_CATEGORIES)
+    end
+
     def output_filter_set(rule_refs)
       if rule_refs == ALL_CATEGORIES
         if @options.mode == Options::ONLINE

data/lib/pmdtester/builders/summary_report_builder.rb

@@ -80,6 +80,7 @@ module PmdTester
           'sha' => details.branch_last_sha,
           'message' => details.branch_last_message
         },
+        'timestamp' => details.timestamp,
         'execution_time' => PmdReportDetail.convert_seconds(details.execution_time),
         'jdk_info' => details.jdk_version,
         'locale' => details.language,

data/lib/pmdtester/parsers/options.rb

@@ -29,6 +29,7 @@ module PmdTester
     attr_reader :threads
     attr_reader :html_flag
     attr_reader :auto_config_flag
+    attr_reader :filter_with_patch_config
     attr_reader :debug_flag
     attr_accessor :filter_set
     attr_reader :keep_reports
@@ -48,6 +49,7 @@ module PmdTester
       @threads = options[:t]
       @html_flag = options[:f]
       @auto_config_flag = options[:a]
+      @filter_with_patch_config = options.filter_with_patch_config?
       @debug_flag = options[:d]
       @filter_set = nil
       @keep_reports = options.keep_reports?
@@ -100,6 +102,9 @@ module PmdTester
         o.bool '-a', '--auto-gen-config',
                'whether to generate configurations automatically based on branch differences,' \
                'this option only works in online and local mode'
+        o.bool '--filter-with-patch-config',
+               'whether to use patch config to filter baseline result as if --auto-gen-config ' \
+               'has been used. This option only works in online mode.'
         o.bool '--keep-reports',
                'whether to keep old reports and skip running PMD again if possible'
         o.bool '-d', '--debug',

data/lib/pmdtester/parsers/pmd_report_document.rb

@@ -8,14 +8,12 @@ module PmdTester
     attr_reader :violations
     attr_reader :errors
     attr_reader :configerrors
-    attr_reader :infos_by_rules
 
     def initialize(branch_name, working_dir, filter_set = nil)
       @violations = CollectionByFile.new
       @errors = CollectionByFile.new
       @configerrors = Hash.new { |hash, key| hash[key] = [] }
 
-      @infos_by_rules = {}
       @current_violations = []
       @current_violation = nil
       @current_error = nil

data/lib/pmdtester/parsers/projects_parser.rb

@@ -21,7 +21,7 @@ module PmdTester
     end
 
     def schema_file_path
-      ResourceLocator.locate('config/projectlist_1_1_0.xsd')
+      ResourceLocator.locate('config/projectlist_1_2_0.xsd')
     end
   end
 

data/lib/pmdtester/pmd_branch_detail.rb

@@ -10,6 +10,8 @@ module PmdTester
     attr_accessor :branch_last_sha
     attr_accessor :branch_last_message
     attr_accessor :branch_name
+    # Start of the regression report
+    attr_accessor :timestamp
     # The branch's execution time on all standard projects
     attr_accessor :execution_time
     attr_accessor :jdk_version
@@ -26,6 +28,7 @@ module PmdTester
       @branch_name = branch_name
       branch_filename = PmdBranchDetail.branch_filename(branch_name)
       @base_branch_dir = "target/reports/#{branch_filename}" unless @branch_name.nil?
+      @timestamp = Time.now
       @execution_time = 0
       # the result of command 'java -version' is going to stderr
       @jdk_version = Cmd.stderr_of('java -version')
@@ -42,11 +45,13 @@ module PmdTester
         details.branch_last_sha = hash['branch_last_sha']
         details.branch_last_message = hash['branch_last_message']
         details.branch_name = hash['branch_name']
+        details.timestamp = hash['timestamp']
         details.execution_time = hash['execution_time']
         details.jdk_version = hash['jdk_version']
         details.language = hash['language']
         details.pull_request = hash['pull_request']
       else
+        details.timestamp = Time.now
         details.jdk_version = ''
         details.language = ''
         logger&.warn "#{details.path_to_save_file} doesn't exist!"
@@ -58,6 +63,7 @@ module PmdTester
       hash = { branch_last_sha: @branch_last_sha,
                branch_last_message: @branch_last_message,
                branch_name: @branch_name,
+               timestamp: @timestamp,
                execution_time: @execution_time,
                jdk_version: @jdk_version,
                language: @language,

data/lib/pmdtester/pmd_tester_utils.rb

@@ -29,6 +29,7 @@ module PmdTester
       parser.parse_file(report_file) if File.exist?(report_file)
       Report.new(
         report_document: doc,
+        file: report_file,
 
         timestamp: report_details.timestamp,
         exec_time: report_details.execution_time

data/lib/pmdtester/project.rb

@@ -1,5 +1,7 @@
 # frozen_string_literal: true
 
+require 'pathname'
+
 module PmdTester
   # This class represents all the information about the project
   class Project
@@ -12,7 +14,8 @@ module PmdTester
     attr_reader :connection
     attr_reader :webview_url
     attr_reader :tag
-    attr_reader :exclude_pattern
+    attr_reader :exclude_patterns
+    attr_reader :src_subpath
     attr_accessor :report_diff
     # key: pmd branch name as String => value: local path of pmd report
     attr_reader :build_command
@@ -20,7 +23,7 @@ module PmdTester
     # stores the auxclasspath calculated after cloning/preparing the project
     attr_accessor :auxclasspath
 
-    def initialize(project)
+    def initialize(project) # rubocop:disable Metrics/CyclomaticComplexity, Metrics/PerceivedComplexity
       @name = project.at_xpath('name').text
       @type = project.at_xpath('type').text
       @connection = project.at_xpath('connection').text
@@ -31,9 +34,10 @@ module PmdTester
       @webview_url = default_webview_url
       @webview_url = webview_url_element.text unless webview_url_element.nil?
 
-      @exclude_pattern = []
+      @src_subpath = project.at_xpath('src-subpath')&.text || '.'
+      @exclude_patterns = []
       project.xpath('exclude-pattern').each do |ep|
-        @exclude_pattern.push(ep.text)
+        @exclude_patterns.push(ep.text)
       end
 
       @build_command = project.at_xpath('build-command')&.text
@@ -56,17 +60,17 @@ module PmdTester
     # Change the file path from 'LOCAL_DIR/SOURCE_CODE_PATH' to
     # 'WEB_VIEW_URL/SOURCE_CODE_PATH'
     def get_webview_url(file_path)
-      file_path.gsub(%r{/#{local_source_path}}, @webview_url)
+      file_path.gsub(%r{/#{clone_root_path}}, @webview_url)
     end
 
     # Change the file path from 'LOCAL_DIR/SOURCE_CODE_PATH' to
     # 'PROJECT_NAME/SOURCE_CODE_PATH'
     def get_path_inside_project(file_path)
-      file_path.gsub(%r{/#{local_source_path}}, @name)
+      file_path.gsub(%r{/#{clone_root_path}}, @name)
     end
 
     def get_local_path(file_path)
-      file_path.sub(%r{/#{local_source_path}/}, '')
+      file_path.sub(%r{/#{clone_root_path}/}, '')
     end
 
     def get_pmd_report_path(branch_name)
@@ -93,6 +97,19 @@ module PmdTester
       end
     end
 
+    ##
+    # Path to the sources to analyze (below or equal to clone_root_path)
+    def local_source_path
+      # normalize path
+      Pathname.new("#{clone_root_path}/#{src_subpath}").cleanpath
+    end
+
+    ##
+    # Path to the clone directory
+    def clone_root_path
+      "#{REPOSITORIES_PATH}/#{@name}"
+    end
+
     def get_project_target_dir(branch_name)
       branch_filename = PmdBranchDetail.branch_filename(branch_name)
       dir = "target/reports/#{branch_filename}/#{@name}"
@@ -100,10 +117,6 @@ module PmdTester
       dir
     end
 
-    def local_source_path
-      "#{REPOSITORIES_PATH}/#{@name}"
-    end
-
     def compute_report_diff(base_branch, patch_branch, filter_set)
       self.report_diff = build_report_diff(get_pmd_report_path(base_branch),
                                            get_pmd_report_path(patch_branch),