ffdc3f5 introduced some counter variables as well as printf statements
into the testcases fac, anagram, huff_enc and susan.
However, neither of those testcases actually provide a printf
implementation themselves, so instead the libc variant was linked and
used. This happend within the designated analysis entrypoints, and was
thus part of the benchmark. This breaks those benchmarks for
sourcelevel WCET-analyzers, so I've removed those calls there.
Note: I've only tested the kernel and sequential benchmarks, it might
be that similar things might still be found within the app or parallel
benchmarks as well.
Please let me know in case you have any further questions.
ffdc3f5 introduced some counter variables as well as printf statements into the testcases fac, anagram, huff_enc and susan. However, neither of those testcases actually provide a printf implementation themselves, so instead the libc variant was linked and used. This happend within the designated analysis entrypoints, and was thus part of the benchmark. This breaks those benchmarks for sourcelevel WCET-analyzers, so I've removed those calls there.
Note: I've only tested the
kernel
andsequential
benchmarks, it might be that similar things might still be found within theapp
orparallel
benchmarks as well.Please let me know in case you have any further questions.