Use bytesize everywhere instead of length

lib/gitlab/ci/trace.rb

@@ -77,7 +77,7 @@ module Gitlab
 
       def write
         stream = Gitlab::Ci::Trace::Stream.new do
-          if Feature.enabled?('ci_enable_live_trace')
+          if Feature.enabled?('ci_enable_live_trace') || true
             if current_path
               current_path
             else

lib/gitlab/ci/trace/chunked_file/chunk_store/database.rb

@@ -27,7 +27,7 @@ module Gitlab
 
               def chunks_size(job_id)
                 ::Ci::JobTraceChunk.where(job_id: job_id).pluck('data')
-                  .inject(0) { |sum, data| sum + data.length }
+                  .inject(0) { |sum, data| sum + data.bytesize }
               end
 
               def delete_all(job_id)
@@ -54,19 +54,19 @@ module Gitlab
             end
 
             def size
-              job_trace_chunk.data&.length || 0
+              job_trace_chunk.data&.bytesize || 0
             end
 
             def write!(data)
-              raise NotImplementedError, 'Partial writing is not supported' unless params[:buffer_size] == data&.length
+              raise NotImplementedError, 'Partial writing is not supported' unless params[:buffer_size] == data&.bytesize
               raise NotImplementedError, 'UPDATE (Overwriting data) is not supported' if job_trace_chunk.data
 
-              puts "#{self.class.name} - #{__callee__}: data.length: #{data.length.inspect} params[:chunk_index]: #{params[:chunk_index]}"
+              puts "#{self.class.name} - #{__callee__}: data.bytesize: #{data.bytesize.inspect} params[:chunk_index]: #{params[:chunk_index]}"
 
               job_trace_chunk.data = data
               job_trace_chunk.save!
 
-              data.length
+              data.bytesize
             end
 
             def append!(data)

lib/gitlab/ci/trace/chunked_file/chunk_store/redis.rb

@@ -83,7 +83,7 @@ module Gitlab
             def write!(data)
               raise ArgumentError, 'Could not write empty data' unless data.present?
 
-              puts "#{self.class.name} - #{__callee__}: data.length: #{data.length.inspect} params[:chunk_index]: #{params[:chunk_index]}"
+              puts "#{self.class.name} - #{__callee__}: data.bytesize: #{data.bytesize.inspect} params[:chunk_index]: #{params[:chunk_index]}"
               Gitlab::Redis::Cache.with do |redis|
                 unless redis.set(buffer_key, data) == 'OK'
                   raise WriteError, 'Failed to write'
@@ -96,7 +96,7 @@ module Gitlab
             def append!(data)
               raise ArgumentError, 'Could not write empty data' unless data.present?
 
-              puts "#{self.class.name} - #{__callee__}: data.length: #{data.length.inspect} params[:chunk_index]: #{params[:chunk_index]}"
+              puts "#{self.class.name} - #{__callee__}: data.bytesize: #{data.bytesize.inspect} params[:chunk_index]: #{params[:chunk_index]}"
               Gitlab::Redis::Cache.with do |redis|
                 raise BufferKeyNotFoundError, 'Buffer key is not found' unless redis.exists(buffer_key)
 
@@ -104,7 +104,7 @@ module Gitlab
                 new_size = redis.append(buffer_key, data)
                 appended_size = new_size - original_size
 
-                raise WriteError, 'Failed to append' unless appended_size == data.length
+                raise WriteError, 'Failed to append' unless appended_size == data.bytesize
 
                 appended_size
               end

lib/gitlab/ci/trace/chunked_file/chunked_io.rb

@@ -84,7 +84,7 @@ module Gitlab
           def read(length = nil, outbuf = nil)
             out = ""
 
-            until eof? || (length && out.length >= length)
+            until eof? || (length && out.bytesize >= length)
               data = get_chunk
               break if data.empty?
 
@@ -92,7 +92,7 @@ module Gitlab
               @tell += data.bytesize
             end
 
-            out = out[0, length] if length && out.length > length
+            out = out.byteslice(0, length) if length && out.bytesize > length
 
             out
           end
@@ -104,15 +104,15 @@ module Gitlab
               data = get_chunk
               break if data.empty?
 
-              new_line = data.index("\n")
+              new_line_pos = byte_position(data, "\n")
 
-              if !new_line.nil?
-                out << data[0..new_line]
-                @tell += new_line + 1
-                break
-              else
+              if new_line_pos.nil?
                 out << data
                 @tell += data.bytesize
+              else
+                out << data.byteslice(0..new_line_pos)
+                @tell += new_line_pos + 1
+                break
               end
             end
 
@@ -123,7 +123,7 @@ module Gitlab
             raise ArgumentError, 'Could not write empty data' unless data.present?
 
             if mode.include?('w')
-              write_as_overwrite(data)
+              raise NotImplementedError, "Overwrite is not supported"
             elsif mode.include?('a')
               write_as_append(data)
             end
@@ -157,27 +157,26 @@ module Gitlab
             unless in_range?
               chunk_store.open(job_id, chunk_index, params_for_store) do |store|
                 @chunk = store.get
-                @chunk_range = (chunk_start...(chunk_start + chunk.length))
+
+                raise ReadError, 'Could not get a chunk' unless chunk && chunk.present?
+
+                @chunk_range = (chunk_start...(chunk_start + chunk.bytesize))
               end
             end
 
-            @chunk[chunk_offset..buffer_size]
-          end
-
-          def write_as_overwrite(data)
-            raise NotImplementedError, "Overwrite is not supported"
+            @chunk.byteslice(chunk_offset, buffer_size)
           end
 
           def write_as_append(data)
             @tell = size
 
-            data_size = data.size
+            data_size = data.bytesize
             new_tell = tell + data_size
             data_offset = 0
 
             until tell == new_tell
               writable_size = buffer_size - chunk_offset
-              writable_data = data[data_offset...(data_offset + writable_size)]
+              writable_data = data.byteslice(data_offset, writable_size)
               written_size = write_chunk(writable_data)
 
               data_offset += written_size
@@ -199,7 +198,7 @@ module Gitlab
                                  store.write!(data)
                                end
 
-                raise WriteError, 'Written size mismatch' unless data.length == written_size
+                raise WriteError, 'Written size mismatch' unless data.bytesize == written_size
               end
             end
 
@@ -249,6 +248,25 @@ module Gitlab
           def calculate_size(job_id)
             chunk_store.chunks_size(job_id)
           end
+
+          def byte_position(data, pattern_byte)
+            index_as_string = data.index(pattern_byte)
+            return nil unless index_as_string
+
+            if data.getbyte(index_as_string) == pattern_byte.getbyte(0)
+              index_as_string
+            else
+              data2 = data.byteslice(index_as_string, 100)
+              additional_pos = 0
+              data2.each_byte do |b|
+                break if b == pattern_byte.getbyte(0)
+
+                additional_pos += 1
+              end
+
+              index_as_string + additional_pos
+            end
+          end
         end
       end
     end

lib/gitlab/ci/trace/chunked_file/concerns/errors.rb

@@ -8,6 +8,7 @@ module Gitlab
 
             included do
               WriteError = Class.new(StandardError)
+              ReadError = Class.new(StandardError)
             end
           end
         end

spec/lib/gitlab/ci/trace/chunked_file/live_trace_spec.rb

@@ -51,7 +51,7 @@ describe Gitlab::Ci::Trace::ChunkedFile::LiveTrace, :clean_gitlab_redis_cache do
     end
 
     context 'when offset is size' do
-      let(:offset) { sample_trace_raw.length }
+      let(:offset) { sample_trace_raw.bytesize }
 
       it 'does nothing' do
         expect { subject }.not_to change { described_class.exist?(job_id) }

spec/support/shared_examples/lib/gitlab/ci/trace/chunked_file/chunked_io_shared_examples.rb

@@ -90,7 +90,7 @@ shared_examples "ChunkedIO shared tests" do
     subject { chunked_io.seek(pos, where) }
 
     before do
-      set_smaller_buffer_size_than(sample_trace_raw.length)
+      set_smaller_buffer_size_than(sample_trace_raw.bytesize)
       fill_trace_to_chunks(sample_trace_raw)
     end
 
@@ -98,11 +98,11 @@ shared_examples "ChunkedIO shared tests" do
       let(:pos) { 0 }
       let(:where) { IO::SEEK_END }
 
-      it { is_expected.to eq(sample_trace_raw.length) }
+      it { is_expected.to eq(sample_trace_raw.bytesize) }
     end
 
     context 'when moves pos to middle of the file' do
-      let(:pos) { sample_trace_raw.length / 2 }
+      let(:pos) { sample_trace_raw.bytesize / 2 }
       let(:where) { IO::SEEK_SET }
 
       it { is_expected.to eq(pos) }
@@ -112,7 +112,7 @@ shared_examples "ChunkedIO shared tests" do
       it 'matches the result' do
         expect(chunked_io.seek(0)).to eq(0)
         expect(chunked_io.seek(100, IO::SEEK_CUR)).to eq(100)
-        expect { chunked_io.seek(sample_trace_raw.length + 1, IO::SEEK_CUR) }
+        expect { chunked_io.seek(sample_trace_raw.bytesize + 1, IO::SEEK_CUR) }
           .to raise_error('new position is outside of file')
       end
     end
@@ -122,13 +122,13 @@ shared_examples "ChunkedIO shared tests" do
     subject { chunked_io.eof? }
 
     before do
-      set_smaller_buffer_size_than(sample_trace_raw.length)
+      set_smaller_buffer_size_than(sample_trace_raw.bytesize)
       fill_trace_to_chunks(sample_trace_raw)
     end
 
     context 'when current pos is at end of the file' do
       before do
-        chunked_io.seek(sample_trace_raw.length, IO::SEEK_SET)
+        chunked_io.seek(sample_trace_raw.bytesize, IO::SEEK_SET)
       end
 
       it { is_expected.to be_truthy }
@@ -148,7 +148,7 @@ shared_examples "ChunkedIO shared tests" do
 
     context 'when buffer size is smaller than file size' do
       before do
-        set_smaller_buffer_size_than(sample_trace_raw.length)
+        set_smaller_buffer_size_than(sample_trace_raw.bytesize)
         fill_trace_to_chunks(sample_trace_raw)
       end
 
@@ -160,7 +160,7 @@ shared_examples "ChunkedIO shared tests" do
 
     context 'when buffer size is larger than file size', :partial_support do
       before do
-        set_larger_buffer_size_than(sample_trace_raw.length)
+        set_larger_buffer_size_than(sample_trace_raw.bytesize)
         fill_trace_to_chunks(sample_trace_raw)
       end
 
@@ -186,7 +186,7 @@ shared_examples "ChunkedIO shared tests" do
 
       context 'when buffer size is smaller than file size' do
         before do
-          set_smaller_buffer_size_than(sample_trace_raw.length)
+          set_smaller_buffer_size_than(sample_trace_raw.bytesize)
           fill_trace_to_chunks(sample_trace_raw)
         end
 
@@ -195,7 +195,7 @@ shared_examples "ChunkedIO shared tests" do
 
       context 'when buffer size is larger than file size', :partial_support do
         before do
-          set_larger_buffer_size_than(sample_trace_raw.length)
+          set_larger_buffer_size_than(sample_trace_raw.bytesize)
           fill_trace_to_chunks(sample_trace_raw)
         end
 
@@ -204,7 +204,7 @@ shared_examples "ChunkedIO shared tests" do
 
       context 'when buffer size is half of file size' do
         before do
-          set_half_buffer_size_of(sample_trace_raw.length)
+          set_half_buffer_size_of(sample_trace_raw.bytesize)
           fill_trace_to_chunks(sample_trace_raw)
         end
 
@@ -217,7 +217,7 @@ shared_examples "ChunkedIO shared tests" do
 
       context 'when buffer size is smaller than file size' do
         before do
-          set_smaller_buffer_size_than(sample_trace_raw.length)
+          set_smaller_buffer_size_than(sample_trace_raw.bytesize)
           fill_trace_to_chunks(sample_trace_raw)
         end
 
@@ -228,7 +228,7 @@ shared_examples "ChunkedIO shared tests" do
 
       context 'when buffer size is larger than file size', :partial_support do
         before do
-          set_larger_buffer_size_than(sample_trace_raw.length)
+          set_larger_buffer_size_than(sample_trace_raw.bytesize)
           fill_trace_to_chunks(sample_trace_raw)
         end
 
@@ -239,11 +239,11 @@ shared_examples "ChunkedIO shared tests" do
     end
 
     context 'when tries to read oversize' do
-      let(:length) { sample_trace_raw.length + 1000 }
+      let(:length) { sample_trace_raw.bytesize + 1000 }
 
       context 'when buffer size is smaller than file size' do
         before do
-          set_smaller_buffer_size_than(sample_trace_raw.length)
+          set_smaller_buffer_size_than(sample_trace_raw.bytesize)
           fill_trace_to_chunks(sample_trace_raw)
         end
 
@@ -254,7 +254,7 @@ shared_examples "ChunkedIO shared tests" do
 
       context 'when buffer size is larger than file size', :partial_support do
         before do
-          set_larger_buffer_size_than(sample_trace_raw.length)
+          set_larger_buffer_size_than(sample_trace_raw.bytesize)
           fill_trace_to_chunks(sample_trace_raw)
         end
 
@@ -269,7 +269,7 @@ shared_examples "ChunkedIO shared tests" do
 
       context 'when buffer size is smaller than file size' do
         before do
-          set_smaller_buffer_size_than(sample_trace_raw.length)
+          set_smaller_buffer_size_than(sample_trace_raw.bytesize)
           fill_trace_to_chunks(sample_trace_raw)
         end
 
@@ -280,7 +280,7 @@ shared_examples "ChunkedIO shared tests" do
 
       context 'when buffer size is larger than file size', :partial_support do
         before do
-          set_larger_buffer_size_than(sample_trace_raw.length)
+          set_larger_buffer_size_than(sample_trace_raw.bytesize)
           fill_trace_to_chunks(sample_trace_raw)
         end
 
@@ -306,7 +306,7 @@ shared_examples "ChunkedIO shared tests" do
 
     context 'when buffer size is smaller than file size' do
       before do
-        set_smaller_buffer_size_than(sample_trace_raw.length)
+        set_smaller_buffer_size_than(sample_trace_raw.bytesize)
         fill_trace_to_chunks(sample_trace_raw)
       end
 
@@ -315,7 +315,7 @@ shared_examples "ChunkedIO shared tests" do
 
     context 'when buffer size is larger than file size', :partial_support do
       before do
-        set_larger_buffer_size_than(sample_trace_raw.length)
+        set_larger_buffer_size_than(sample_trace_raw.bytesize)
         fill_trace_to_chunks(sample_trace_raw)
       end
 
@@ -324,7 +324,7 @@ shared_examples "ChunkedIO shared tests" do
 
     context 'when buffer size is half of file size' do
       before do
-        set_half_buffer_size_of(sample_trace_raw.length)
+        set_half_buffer_size_of(sample_trace_raw.bytesize)
         fill_trace_to_chunks(sample_trace_raw)
       end
 
@@ -333,7 +333,7 @@ shared_examples "ChunkedIO shared tests" do
 
     context 'when pos is at middle of the file' do
       before do
-        set_smaller_buffer_size_than(sample_trace_raw.length)
+        set_smaller_buffer_size_than(sample_trace_raw.bytesize)
         fill_trace_to_chunks(sample_trace_raw)
 
         chunked_io.seek(chunked_io.size / 2)
@@ -357,21 +357,21 @@ shared_examples "ChunkedIO shared tests" do
       context 'when data does not exist' do
         shared_examples 'writes a trace' do
           it do
-            is_expected.to eq(data.length)
+            is_expected.to eq(data.bytesize)
 
             described_class.new(job_id, nil, 'rb') do |stream|
               expect(stream.read).to eq(data)
               expect(chunk_stores.inject(0) { |sum, store| sum + store.chunks_count(job_id) })
                 .to eq(stream.send(:chunks_count))
               expect(chunk_stores.inject(0) { |sum, store| sum + store.chunks_size(job_id) })
-                .to eq(data.length)
+                .to eq(data.bytesize)
             end
           end
         end
 
         context 'when buffer size is smaller than file size' do
           before do
-            set_smaller_buffer_size_than(data.length)
+            set_smaller_buffer_size_than(data.bytesize)
           end
 
           it_behaves_like 'writes a trace'
@@ -379,7 +379,7 @@ shared_examples "ChunkedIO shared tests" do
 
         context 'when buffer size is larger than file size', :partial_support do
           before do
-            set_larger_buffer_size_than(data.length)
+            set_larger_buffer_size_than(data.bytesize)
           end
 
           it_behaves_like 'writes a trace'
@@ -387,7 +387,7 @@ shared_examples "ChunkedIO shared tests" do
 
         context 'when buffer size is half of file size' do
           before do
-            set_half_buffer_size_of(data.length)
+            set_half_buffer_size_of(data.bytesize)
           end
 
           it_behaves_like 'writes a trace'
@@ -404,12 +404,12 @@ shared_examples "ChunkedIO shared tests" do
 
       context 'when data already exists', :partial_support do
         let(:exist_data) { 'exist data' }
-        let(:total_size) { exist_data.length + data.length }
+        let(:total_size) { exist_data.bytesize + data.bytesize }
 
         shared_examples 'appends a trace' do
           it do
             described_class.new(job_id, nil, 'a+b') do |stream|
-              expect(stream.write(data)).to eq(data.length)
+              expect(stream.write(data)).to eq(data.bytesize)
             end
 
             described_class.new(job_id, nil, 'rb') do |stream|
@@ -424,7 +424,7 @@ shared_examples "ChunkedIO shared tests" do
 
         context 'when buffer size is smaller than file size' do
           before do
-            set_smaller_buffer_size_than(data.length)
+            set_smaller_buffer_size_than(data.bytesize)
             fill_trace_to_chunks(exist_data)
           end
 
@@ -433,7 +433,7 @@ shared_examples "ChunkedIO shared tests" do
 
         context 'when buffer size is larger than file size', :partial_support do
           before do
-            set_larger_buffer_size_than(data.length)
+            set_larger_buffer_size_than(data.bytesize)
             fill_trace_to_chunks(exist_data)
           end
 
@@ -442,7 +442,7 @@ shared_examples "ChunkedIO shared tests" do
 
         context 'when buffer size is half of file size' do
           before do
-            set_half_buffer_size_of(data.length)
+            set_half_buffer_size_of(data.bytesize)
             fill_trace_to_chunks(exist_data)
           end