Add backend for merge conflicts reading

2 files changed, 254 insertions, 0 deletions

diff --git a/spec/lib/gitlab/conflict/file_spec.rb b/spec/lib/gitlab/conflict/file_spec.rb
new file mode 100644
index 00000000000..318d0d249d6
--- /dev/null
+++ b/spec/lib/gitlab/conflict/file_spec.rb
@@ -0,0 +1,76 @@
+require 'spec_helper'
+
+describe Gitlab::Conflict::File, lib: true do
+  let(:project) { create(:project) }
+  let(:repository) { project.repository }
+  let(:rugged) { repository.rugged }
+  let(:their_ref) { their_commit.oid }
+  let(:their_commit) { rugged.branches['conflict-a'].target }
+  let(:our_ref) { our_commit.oid }
+  let(:our_commit) { rugged.branches['conflict-b'].target }
+  let(:index) { rugged.merge_commits(our_commit, their_commit) }
+  let(:conflict) { index.conflicts.last }
+  let(:merge_file) { index.merge_file('files/ruby/regex.rb') }
+  let(:conflict_file) { Gitlab::Conflict::File.new(merge_file, conflict, their_ref, our_ref, repository) }
+
+  describe '#highlighted_lines' do
+    def html_to_text(html)
+      CGI.unescapeHTML(ActionView::Base.full_sanitizer.sanitize(html))
+    end
+
+    it 'returns lines with rich_text' do
+      conflict_file.highlighted_lines.each do |line|
+        expect(line).to have_attributes(rich_text: an_instance_of(String))
+      end
+    end
+
+    it 'returns lines with rich_text matching the text content of the line' do
+      conflict_file.highlighted_lines.each do |line|
+        expect(line.text).to eq(html_to_text(line.rich_text))
+      end
+    end
+  end
+
+  describe '#sections' do
+    it 'returns match lines when there is a gap between sections' do
+      section = conflict_file.sections[5]
+      match_line = section[:lines][0]
+
+      expect(section[:conflict]).to be_falsey
+      expect(match_line.type).to eq('match')
+      expect(match_line.text).to eq('@@ -46,53 +46,53 @@')
+    end
+
+    it 'does not return match lines when there is no gap between sections' do
+      conflict_file.sections.each_with_index do |section, i|
+        unless i == 5
+          expect(section[:lines][0].type).not_to eq(5)
+        end
+      end
+    end
+
+    it 'sets conflict to false for sections with only unchanged lines' do
+      conflict_file.sections.reject { |section| section[:conflict] }.each do |section|
+        without_match = section[:lines].reject { |line| line.type == 'match' }
+
+        expect(without_match).to all(have_attributes(type: nil))
+      end
+    end
+
+    it 'only includes a maximum of CONTEXT_LINES (plus an optional match line) in context sections' do
+      conflict_file.sections.reject { |section| section[:conflict] }.each do |section|
+        without_match = section[:lines].reject { |line| line.type == 'match' }
+
+        expect(without_match.length).to be <= Gitlab::Conflict::File::CONTEXT_LINES * 2
+      end
+    end
+
+    it 'sets conflict to true for sections with only changed lines' do
+      conflict_file.sections.select { |section| section[:conflict] }.each do |section|
+        section[:lines].each do |line|
+          expect(line.type).to be_in(['new', 'old'])
+        end
+      end
+    end
+  end
+end
diff --git a/spec/lib/gitlab/conflict/parser_spec.rb b/spec/lib/gitlab/conflict/parser_spec.rb
new file mode 100644
index 00000000000..335da58e13c
--- /dev/null
+++ b/spec/lib/gitlab/conflict/parser_spec.rb
@@ -0,0 +1,178 @@
+require 'spec_helper'
+
+describe Gitlab::Conflict::Parser, lib: true do
+  let(:parser) { Gitlab::Conflict::Parser.new }
+
+  describe '#parse' do
+    context 'when the file has valid conflicts' do
+      let(:text) do
+        <<CONFLICT
+module Gitlab
+  module Regexp
+    extend self
+
+    def username_regexp
+      default_regexp
+    end
+
+<<<<<<< files/ruby/regex.rb
+    def project_name_regexp
+      /\A[a-zA-Z0-9][a-zA-Z0-9_\-\. ]*\z/
+    end
+
+    def name_regexp
+      /\A[a-zA-Z0-9_\-\. ]*\z/
+=======
+    def project_name_regex
+      %r{\A[a-zA-Z0-9][a-zA-Z0-9_\-\. ]*\z}
+    end
+
+    def name_regex
+      %r{\A[a-zA-Z0-9_\-\. ]*\z}
+>>>>>>> files/ruby/regex.rb
+    end
+
+    def path_regexp
+      default_regexp
+    end
+
+<<<<<<< files/ruby/regex.rb
+    def archive_formats_regexp
+      /(zip|tar|7z|tar\.gz|tgz|gz|tar\.bz2|tbz|tbz2|tb2|bz2)/
+=======
+    def archive_formats_regex
+      %r{(zip|tar|7z|tar\.gz|tgz|gz|tar\.bz2|tbz|tbz2|tb2|bz2)}
+>>>>>>> files/ruby/regex.rb
+    end
+
+    def git_reference_regexp
+      # Valid git ref regexp, see:
+      # https://www.kernel.org/pub/software/scm/git/docs/git-check-ref-format.html
+      %r{
+        (?!
+           (?# doesn't begins with)
+           \/|                    (?# rule #6)
+           (?# doesn't contain)
+           .*(?:
+              [\/.]\.|            (?# rule #1,3)
+              \/\/|               (?# rule #6)
+              @\{|                (?# rule #8)
+              \\                  (?# rule #9)
+           )
+        )
+        [^\000-\040\177~^:?*\[]+  (?# rule #4-5)
+        (?# doesn't end with)
+        (?<!\.lock)               (?# rule #1)
+        (?<![\/.])                (?# rule #6-7)
+      }x
+    end
+
+    protected
+
+<<<<<<< files/ruby/regex.rb
+    def default_regexp
+      /\A[.?]?[a-zA-Z0-9][a-zA-Z0-9_\-\.]*(?<!\.git)\z/
+=======
+    def default_regex
+      %r{\A[.?]?[a-zA-Z0-9][a-zA-Z0-9_\-\.]*(?<!\.git)\z}
+>>>>>>> files/ruby/regex.rb
+    end
+  end
+end
+CONFLICT
+      end
+
+      let(:lines) { parser.parse(text, 'files/ruby/regex.rb', 'files/ruby/regex.rb') }
+
+      it 'sets our lines as new lines' do
+        expect(lines[8..13]).to all(have_attributes(type: 'new'))
+        expect(lines[26..27]).to all(have_attributes(type: 'new'))
+        expect(lines[56..57]).to all(have_attributes(type: 'new'))
+      end
+
+      it 'sets their lines as old lines' do
+        expect(lines[14..19]).to all(have_attributes(type: 'old'))
+        expect(lines[28..29]).to all(have_attributes(type: 'old'))
+        expect(lines[58..59]).to all(have_attributes(type: 'old'))
+      end
+
+      it 'sets non-conflicted lines as both' do
+        expect(lines[0..7]).to all(have_attributes(type: nil))
+        expect(lines[20..25]).to all(have_attributes(type: nil))
+        expect(lines[30..55]).to all(have_attributes(type: nil))
+        expect(lines[60..62]).to all(have_attributes(type: nil))
+      end
+
+      it 'sets consecutive line numbers for index, old_pos, and new_pos' do
+        old_line_numbers = lines.select { |line| line.type != 'new' }.map(&:old_pos)
+        new_line_numbers = lines.select { |line| line.type != 'old' }.map(&:new_pos)
+
+        expect(lines.map(&:index)).to eq(0.upto(62).to_a)
+        expect(old_line_numbers).to eq(1.upto(53).to_a)
+        expect(new_line_numbers).to eq(1.upto(53).to_a)
+      end
+    end
+
+    context 'when the file contents include conflict delimiters' do
+      let(:path) { 'README.md' }
+
+      def parse_text(text)
+        parser.parse(text, path, path)
+      end
+
+      it 'raises UnexpectedDelimiter when there is a non-start delimiter first' do
+        expect { parse_text('=======') }.
+          to raise_error(Gitlab::Conflict::Parser::UnexpectedDelimiter)
+
+        expect { parse_text('>>>>>>> README.md') }.
+          to raise_error(Gitlab::Conflict::Parser::UnexpectedDelimiter)
+
+        expect { parse_text('>>>>>>> some-other-path.md') }.
+          not_to raise_error
+      end
+
+      it 'raises UnexpectedDelimiter when a start delimiter is followed by a non-middle delimiter' do
+        start_text = "<<<<<<< README.md\n"
+        end_text = "\n=======\n>>>>>>> README.md"
+
+        expect { parse_text(start_text + '>>>>>>> README.md' + end_text) }.
+          to raise_error(Gitlab::Conflict::Parser::UnexpectedDelimiter)
+
+        expect { parse_text(start_text + start_text + end_text) }.
+          to raise_error(Gitlab::Conflict::Parser::UnexpectedDelimiter)
+
+        expect { parse_text(start_text + '>>>>>>> some-other-path.md' + end_text) }.
+          not_to raise_error
+      end
+
+      it 'raises UnexpectedDelimiter when a middle delimiter is followed by a non-end delimiter' do
+        start_text = "<<<<<<< README.md\n=======\n"
+        end_text = "\n>>>>>>> README.md"
+
+        expect { parse_text(start_text + '=======' + end_text) }.
+          to raise_error(Gitlab::Conflict::Parser::UnexpectedDelimiter)
+
+        expect { parse_text(start_text + start_text + end_text) }.
+          to raise_error(Gitlab::Conflict::Parser::UnexpectedDelimiter)
+
+        expect { parse_text(start_text + '>>>>>>> some-other-path.md' + end_text) }.
+          not_to raise_error
+      end
+
+      it 'raises MissingEndDelimiter when there is no end delimiter at the end' do
+        start_text = "<<<<<<< README.md\n=======\n"
+
+        expect { parse_text(start_text) }.
+          to raise_error(Gitlab::Conflict::Parser::MissingEndDelimiter)
+
+        expect { parse_text(start_text + '>>>>>>> some-other-path.md') }.
+          to raise_error(Gitlab::Conflict::Parser::MissingEndDelimiter)
+      end
+    end
+
+    context 'when lines is blank' do
+      it { expect(parser.parse('', 'README.md', 'README.md')).to eq([]) }
+      it { expect(parser.parse(nil, 'README.md', 'README.md')).to eq([]) }
+    end
+  end
+end