magsilva / proteum Goto Github PK

Proteum cannot parse an IOL file it has created (file attached to this issue). Error message:

tcase: Error Reading File: Prime.IOL

Prime.TXT

A long time ago, I faced a issue when importing test cases. For instance, I tried to exhaustively create and add test cases to current project (test set):

for month in `seq 1 12`; do
  for year in `seq 1 9999`; do
     tcase -add -D /media/disk/Cal -E cal  -p "$month $year"  -EE
__cal_inst -trace cal;
  done;
done

However, only 512 test cases could be imported. It would be important to remove such limit.

Currently, to keep working with more than 512 mutants, we can workaround this by importing as much as 512 test cases, run the mutants, and run tcase with the option to remove test cases that could not kill mutants. This should free some space to import new test cases.

Currently, lib/gerais.c has some static variables such as 'ap_from_buf' and 'bufaux'. Remove them and rewrite the functions that depends upon them: 'inic_from_buf' and 'from_buf'.

Proteum fails to recognize functions within a software under testing. That's quite frequent. For instance, if a function makes use of 'sprintf', it is not recognized by Proteum. See Righetto's report:

"""
a ferramenta não consegue reconhecer todas as funções do código; não sei ao certo o porque isso acontece, mas fiz alguns testes e utilizar a função "sprintf" faz com que a função não seja reconhecida, ou utilizar acesso de vetor ou criação estática em uma função diferente da main. Porém, esta ultima afirmação eu não tenho total certeza porque em programas como QuadTree existe uma sctruct com declaração estática de vetor e funciona.
"""

When running test cases, move all error messages to a log file instead of sending them to a stderr. Alternatively, we could still send messages to stderr, but we must review the code to assure that, for every execution on test cases, we provide different stderr and stdout file descriptors.