Closed jgroote closed 12 years ago
According to the Aldebaran standard, lps2lts produces correct output; the assumption when parsing .aut files that state numbers are smaller than the number of states is incorrect. Fixing this in lps2lts is impossible, as this requires too much memory.
Log: Resolves bug #1024.
A .aut file with non consecutive state numbers is now read. This means that a file written using bit hashing can now be read.
The rational test has been adapted. The ltsconvert tests should not be run, because lps2lts does not work. This does not work with -rjittyp. As lps2lts does not work not .aut, .fsm and .lts files are generated. This means ltsinfo, ltsconvert and the ltscompare tests can not be run. For reasons that I do not understand the ltscompare tests are still running on my machine.
Modified: trunk/libraries/lts/source/liblts_aut.cpp trunk/scripts/MCRL2TestTargets.cmake
Issue migrated from trac ticket # 1024
component: lps2lts | priority: minor | resolution: fixed
2012-06-04 11:51:24: scranen@win.tue.nl created the issue
When using bithashing, the state indices in generated .aut files are incorrect. E.g., when generating a random trace using the following command:
Viewing the resulting .aut file gives the following error: