| Popis: |
V magistrskem delu smo predstavili področje tehničnega dolga in pomanjkljive kode ter raziskali povezavo med tipi pomanjkljive kode zaznanimi z izbranimi orodji in privzetimi pravili v orodju SonarQube. Za raziskavo omenjene povezave smo klasificirali 12 pravil orodja SonarQube v različne tipe pomanjkljive kode. Na podlagi kriterijev smo izbrali 32 projektov, orodje JSpIRIT in orodje JDeodorant ter tri najbolj pogosto analizirane tipe pomanjkljive kode. Empirični podatki dobljeni z analizo izbranih projektov so statistično analizirani in nakazujejo na težave pri preslikavi pravil orodja SonarQube v tipe pomanjkljive kode. Posledično je skoraj nemogoče definirati povezavo med zaznanimi pomanjkljivimi kodami v izbranih orodjih in pravili v orodju SonarQube. In this master thesis, we present the area of code smells and technical debt. We focus our research on the correlation between code smells detected by chosen tools and the default rules defined in SonarQube. To help establish the correlation we classified 12 rules defined in SonarQube into different types of code smells. Based on the defined criteria we selected 32 projects, tools JSpIRIT and JDeodorant and three most analyzed code smell types. Empirical data obtained through analysis of chosen projects is statistically analyzed and it indicates on problems when mapping SonarQube’s rules into different types of code smells. As a result it is almost impossible to define a correlation between code smells detected by chosen tools and rules defined in SonarQube. |
