English

CIn - Centro de Informática UFPE




Eventos Relacionados

Defesa de Dissertação de Mestrado Nº 1.241: "Dataflow Analysis for Software Product Lines"

O aluno Társis Wanderley Tolêdo irá defender seu trabalho nesta quinta (21), às 10h, na sala D224 Início: 21/02/2013 às 10:00 Término: 21/02/2013 às 12:00 Local: Sala D224

Pós-Graduação em Ciência da Computação – UFPE
Defesa de Dissertação de Mestrado Nº  1.241
 
Aluno: Társis Wanderley Tolêdo
Orientador: Prof. Paulo Henrique Monteiro Borba
Título: Dataflow Analysis for Software Product Lines
Data: 21/02/2013
Hora/Local: 10:00h – Sala D224
Banca Examinadora: 
Prof. Marcelo Bezerra D'Amorin  (UFPE / CIn)
Prof. Dalton Dario Serey Guerrero (UFCG / Departamento de Sistemas e Computação)
Prof. Paulo Henrique Monteiro Borba  (UFPE / CIn)
 
RESUMO:
 
Software product lines (SPLs) that use preprocessor directives, such as #ifdefs, to define code-level variability of features suffer from the lack of techniques of dataflow analyses. Developers are forced to explicitly generate all products from the SPL in order to apply dataflow analyses, and that becomes prohibitive when the number of possible products increases due to the combinatorial nature of SPLs. In this work, we describe how two different dataflow analsysis frameworks, Kildall's and IFDS/IDE, are lifted to support feature-sensitive dataflow analyses without using brute force to explicitly generate all possible products in the product line. Specifically, we describe four different ways of performing intraprocedural analyses within Kildall's framework and compare these to the brute-force approach; and how interprodecural analyses can also be lifted to feature-sensitivity within the IFDS/IDE framework. We experimented with the performance of the proposed techniques by applying feature-sensitive dataflow analyses to five different SPLs. We found that it is possible to significantly speed up intraprocedural and interprocedural analyses, the latter by several orders of magnitude.
 
Palavras-chave: Software Product Lines, Dataflow Analysis

  • © Centro de Informática UFPE - Todos os direitos reservados
    Tel +55 81 2126.8430 - Cidade Universitária - 50740-560 - Recife/PE
Plano4 Consultoria Web