Detailed Analysis

Cross-category patterns and system-wide analysis derived from individual component analyses.

Table of Contents

Analysis Methodology

Analysis Methodology

Applied Concepts & Principles

Analysis Phase Concepts/Principles
Data Collection Domain mapping from sequence diagrams, SRP analysis, Call frequency analysis
Individual File Fowler's Code Smells, Suryanarayana's Design Smells (OOP-only excluded)
Design Symptoms Martin's 7 Design Symptoms
System-Level Architecture Smells, Cross-pipeline comparison

Chapter 1: Common Symptoms by Category

1.1 Helper Common Symptoms

Data Sources:

Component Source
generalHelper 01-generalHelper/03-design-smells-symptoms.md
jsHelper 02-jsHelper/03-design-smells-symptoms.md
unityHelper 03-unityHelper/03-design-smells-symptoms.md

Summary:

Symptom generalHelper jsHelper unityHelper
Rigidity High Low High
Fragility High Low High
Immobility High Medium Medium
Viscosity Medium Medium High
Needless Complexity - - -
Needless Repetition High Medium High
Opacity Medium Medium Medium

Common Patterns: - All 3 Helpers exhibit Rigidity and Needless Repetition - No Needless Complexity found (no over-engineering) - generalHelper most severe: 7 domains, highest Immobility - unityHelper highest Viscosity: Stage conditionals make proper design harder than copy-paste - jsHelper least severe overall

1.2 Jenkinsfile Common Symptoms

Data Sources:

Component Source
DLX CI 04-DLXJenkins-Jenkinsfile/03-design-smells-symptoms.md
DLX CD 05-DLXJenkins-JenkinsfileDeployment/03-design-smells-symptoms.md
JS CI 06-JsJenkins-Jenkinsfile/03-design-smells-symptoms.md
JS CD 07-JsJenkins-JenkinsfileDeployment/03-design-smells-symptoms.md
Jenkins CI 08-PipelineForJenkins-Jenkinsfile/03-design-smells-symptoms.md

Summary:

Symptom DLX CI DLX CD JS CI JS CD Jenkins CI
Rigidity High High High High Medium
Fragility High High High High High
Immobility High High High High High
Needless Complexity - - Low - -
Viscosity Medium High High High Medium
Needless Repetition High High High High High
Opacity Medium Medium High Medium High

Common Patterns: - All 5 Jenkinsfiles exhibit Immobility and Needless Repetition (High) - Needless Complexity nearly absent (only JS CI has commented-out code) - CD pipelines (DLX CD, JS CD) have higher Viscosity due to deployment pattern duplication - JS CI has highest Opacity due to undefined variable and 112-line Long Method - Jenkins CI has Opacity issue due to catchError parameter typo (potential bug)

Chapter 2: Cross-Category Patterns

2.1 Abstraction Inconsistency

Data Sources:

Source Extracted Data
Pipeline Sequence Diagrams Pipeline Comparison tables

Evidence:

Function Helper Usage Direct Call Pipelines
Git operations cloneOrUpdateRepo git clone/checkout/reset/pull CI vs CD
Build status init initializeEnvironment sendBuildStatus direct CI vs CD
Web Server cleanMergedBranchFromWebServer SSH/SCP direct DLX CD (same file)

Analysis: - Same functionality implemented differently between CI and CD pipelines - Even within same file (DLX CD), Web Server operations use both Helper and direct calls

2.2 Shotgun Surgery Locations

Data Sources:

Source Extracted Data
01-generalHelper/02-software-smells-analysis.md Cross-file Shotgun Surgery
03-unityHelper/02-software-smells-analysis.md Cross-file + Within-file Shotgun Surgery
04-08 Jenkinsfile analyses Within-file Shotgun Surgery

Cross-file (Helper signature change):

Change Trigger Affected Files Total Locations
sendBuildStatus generalHelper + 5 Jenkinsfiles 18
runUnityStage unityHelper + 2 Jenkinsfiles 9
parseJson generalHelper + 5 Jenkinsfiles 10

Within-file (Pattern duplication):

Change Trigger File Locations
stageName/errorMsg DLX CI, DLX CD 9
Adding new Unity Stage unityHelper 7
Server/Client report pattern JS CI 6
Server/Client deploy pattern JS CD 4
Groovy/Jenkinsfile lint pattern Jenkins CI 4

Analysis: - sendBuildStatus change requires 18 modifications across 6 files - Pattern duplication within Jenkinsfiles creates additional Shotgun Surgery risk

Chapter 3: System-wide Patterns

3.1 Pipeline Duplication Analysis

Data Sources:

Source Extracted Data
Jenkinsfile source code DLX CI, DLX CD, JS CI, JS CD, Jenkins CI

Lines by Section:

Section DLX CI DLX CD JS CI JS CD Jenkins CI Total
Boilerplate
Variables 6 5 4 5 3 23
Tools 3 0 3 1 5 12
Parameters 11 9 10 8 10 48
Triggers 18 17 17 16 16 84
Environment 13 10 16 13 17 69
Post 22 22 22 26 22 114
Boilerplate Subtotal 73 63 72 69 73 350
Stages
Delete Merged Branch - 28 - 26 - 54
Prepare WORKSPACE 53 36 35 30 33 187
Install Dependencies - - 16 16 - 32
Linting 43 24 12 12 67 158
EditMode Tests 18 17 - - - 35
PlayMode Tests 14 13 - - - 27
Unit Testing - - 82 14 - 96
Run Unit Tests - - - - 5 5
Publish Test Results - - - - 5 5
Code Coverage 24 - - - - 24
Build Project 30 14 - - - 44
Static Analysis - - 35 33 42 110
Check Build Condition - - - 24 - 24
Deploy Build - 32 - - - 32
Server Build and Deploy - - - 56 - 56
Client Build and Deploy - - - 56 - 56
Generate Groovydoc - - - - 28 28
Stages Subtotal 182 164 180 267 180 973
Total 255 227 252 336 253 1,323

Duplication Analysis:

Comparison Common Lines Total Lines Duplication %
All 5 Pipelines (Boilerplate + Prepare WORKSPACE) 537 1,323 41%
DLX CI vs DLX CD 398 482 83%
JS CI vs JS CD 426 588 72%

Note: - Includes Stages and Boilerplate that exist across pipelines. - Includes sections that can be abstracted into one, even if not completely identical.

Code Duplication Trend:

Baseline Commit Date Duplication %
Earlier analysis 54479b2 2025-02-21 37%
Current analysis 74fc356 2025-03-20 41%

Chapter 4: Why Untestable?

Data Sources:

Source Extracted Data
architecture-smells-analysis.md Section 2.1 Hub-like Dependency Diagram
architecture-smells-analysis.md Section 2.3 External calls (113), Environment variables (10)

Evidence:

Factor Count Impact
External calls via sh 113 Cannot mock in unit tests
Hidden environment variables 10 Runtime-only values
Hub-like dependencies 5 pipelines → generalHelper No isolation possible
Direct external calls from pipelines 50 Bypasses any abstraction

Analysis: - All Helpers and Pipelines directly call external systems via sh - Tight coupling to external systems without Dependency Injection - Testing requires actual Jenkins + Git + Bitbucket + Unity + Docker environment

References