You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
#1211 added new command-line arguments for choosing which result files benchexec should produce, fixing #983. However, the default remains the same and can be confusing. For the next major revision of BenchExec we should finish this:
Choose a default mode and implement it. Likely the default mode should be either the same as --results-per-rundefinition or --results-per-rundefinition --results-per-taskset.
Tighten rules for uniqueness and presence of names of run definitions and task sets. Add parameters for choosing which result XML files to write #1211 added such rules for cases when the new parameters are used, but in the future they should always be enforced: no mix of named and unnamed tags, and no duplicate tag names (at least whenever those names would end up in the file name of a results file).
Get rid of the distinction between "name" and "real name" in benchexec.model.RunSet. This is no longer necessary once we have consistent names.
Basically we should go through the diff of #1211 and resolve all TODOs and all cases where new code is currently guarded behind either --results-per-rundefinition or --results-per-taskset being present.
#1211 added new command-line arguments for choosing which result files
benchexecshould produce, fixing #983. However, the default remains the same and can be confusing. For the next major revision of BenchExec we should finish this:--results-per-rundefinitionor--results-per-rundefinition --results-per-taskset.benchexec.model.RunSet. This is no longer necessary once we have consistent names.Basically we should go through the diff of #1211 and resolve all TODOs and all cases where new code is currently guarded behind either
--results-per-rundefinitionor--results-per-tasksetbeing present.