jpf.properties

# JPF configuration for "jpf-core" component

# path elements can either be relative to the property file location, or
# use the JPF component name (e.g. "jpf-core") as a variable prefix
# (e.g. "jpf-core.sourcepath=${jpf-core}/src/examples").
# If a jpf.properties is to be used within NB, it has to use the variable prefix
# (prepending the project root is only done during JPF init)

# we use the ';' separator here because it is recognized by NetBeans (as opposed
# to ',')

# 'config_path' is set automatically by gov.nasa.jpf.Config during JPF init.
# If used from within an Ant build system, 'jpf-core' has to be set explicitly
# (ant does ${..} property expansion, but ignores property redefinition, so the
# following line will be ignored)

jpf-core = /home/m/java/java_pathfinder_r615/jpf-core

jpf-core.native_classpath=\
  ${jpf-core}/build/jpf.jar;\
  ${jpf-core}/build/jpf-annotations.jar;\
  ${jpf-core}/lib/junit-4.6.jar

jpf-core.classpath=\
  ${jpf-core}/build/jpf-classes.jar;\
  ${jpf-core}/build/examples

jpf-core.sourcepath=\
  ${jpf-core}/src/examples

jpf-core.test_classpath=\
  ${jpf-core}/build/tests

jpf-core.peer_packages = gov.nasa.jpf.jvm,<model>,<default>



# the default jpf-core properties file with keys for which we need mandatory defaults

########################### 0. global part ###############################

# instances that are both search and VM listeners
#listener = ..

# do we want JPF exceptions to print their stack traces (that's only for
# debugging)
jpf.print_exception_stack = true


# this is where we can specify additional classpath entries that are
# not in the system property class.path (e.g. when running JPF from
# an environment that uses it's own loaders, like Eclipse plugins etc.)
#jpf.native_classpath = ..


########################### 1. Search part ###############################
search.class = gov.nasa.jpf.search.DFSearch


# This flag indicates whether state matching will only be done when a state
# is revisited at a lower depth. By default this is false. If it is set to
# true and no error is found in a limited-depth search then it is guaranteed
# not to have such error below that depth. Note that for traditional
# depth limited search this does not hold
search.match_depth = false

# This flag indicates whether JPF should produce more than one error
# or stop at the first one
search.multiple_errors = false

# the minimum free memory bounds. If we fall below that threshold, we
# stop the search
search.min_free = 1M

# name of the file in which we store error paths. If not set, we don't store
#search.error_path = error.xml

# the standard properties we want to check for
search.properties=\
gov.nasa.jpf.jvm.NotDeadlockedProperty,\
gov.nasa.jpf.jvm.NoUncaughtExceptionsProperty


# various heuristic parameters

# This number specifies the maximum number of states to keep on the queue
# during a heuristic search. By default it is set to -1
search.heuristic.queue_limit = -1

# This flag indicates whether branches with counts less than branch-start
# are to be ranked according to how many times they have been taken.
# It is set to true by default. If it is set to false, they are all valued
# the same
search.heuristic.branch.count_early = true

# This number determines at what point branches are heuristically valued as worse
# than non-branching transitions. By default this value is 1.
branch_start = 1


# This number if greater than 0 is returned as the heuristic value for
# non-branching transitions. By default it is set to -1 in which case the
# value of branch-start is returned instead
search.heuristic.branch.no_branch_return = -1

# exclusive search listeners
# search.listener = 


############################### 2. VM part ###############################


# this is an ordered list of packages from which we try to locate native peers.
# "<model>" means JPF tries the same package like the model class
# "<default>" means no package at all
# (this is going to be extended by jpf.properties files)
#peer_packages = <model>,<default>


vm.class = gov.nasa.jpf.jvm.JVM

# class used to hash/store states (if not set, states are not matched)
vm.storage.class = gov.nasa.jpf.jvm.JenkinsStateSet

# class used to maintain the backtrack stack
vm.backtracker.class = gov.nasa.jpf.jvm.DefaultBacktracker

# serializer to be used by state set (vm.storage.class)
vm.serializer.class = gov.nasa.jpf.jvm.serialize.CFSerializer
#vm.serializer.class = gov.nasa.jpf.jvm.serialize.AdaptiveSerializer
#vm.serializer.class = gov.nasa.jpf.jvm.serialize.FilteringSerializer

# the class that models static fields and classes
vm.static.class = gov.nasa.jpf.jvm.StaticArea

# the class that models the heap
#vm.heap.class = gov.nasa.jpf.jvm.DynamicArea
vm.heap.class = gov.nasa.jpf.jvm.SparseClusterArrayHeap

# the class representing the list of all threads
vm.threadlist.class = gov.nasa.jpf.jvm.ThreadList

# restorer to be used by backtracker such as DefaultBacktracker UNLESS a
# serializer that is also a restorer (such as CollapsingSerializer) is used.
# I.e. this is only read if serializer is not used or it's not a StateRestorer
vm.restorer.class = .jvm.DefaultMementoRestorer

# instruction factory
vm.insn_factory.class = gov.nasa.jpf.jvm.bytecode.InstructionFactory

# fields factory
vm.fields_factory.class = gov.nasa.jpf.jvm.DefaultFieldsFactory

# class used to set method and field attributes at class load time
vm.attributor.class = gov.nasa.jpf.jvm.DefaultAttributor

# pattern list for assertion enabled/disabled classes
#vm.enable_assertions = *
#vm.disable_assertions=

# do we support the Verify.ignorePath() API (to imperatively ignore paths in modeled/instrumented programs)?
vm.verify.ignore_path = true

# do we want on-the-fly partial order reduction?
vm.por = true

# do we treat field access of objects that are reachable from different threads as
# transaction boundaries (in case there is more tha one thread)?
vm.por.field_boundaries = true

# access to fields with these patterns we never treat as transition boundaries
#vm.por.field_boundaries.never = java.*,javax.*,sun.*
vm.por.field_boundaries.never = 

# access to fields with these patterns we break on (if object is shared and
# vm.por.field_boundaries is set - note it's only meaningful if sync_detection
# is set because otherwse we would break anyways)
vm.por.field_boundaries.break =

# do we try to deduce if a field is supposed to be protected by a lock? In this
# case, we stop to treat this field as a boundary step, but only if we see the lock
vm.por.sync_detection = true

# used to detect if a field is lock protected - only required if we use on-the-fly POR
# with sync-detection
vm.por.fli_factory.class = gov.nasa.jpf.jvm.StatisticFieldLockInfoFactory

# do we assume final fields to be race-safe (not really true, esp. for
# instance fields with references leaking from a ctor, but a good starting point)
vm.por.skip_finals=true

# ClassLoaders synchronize the loading of a class.
# Thus, static final fields can never be included in a race condition since only 1 thread is allowed to access the class while it is loading.
# Defaulted to false to maintain backward compatibility in JPF
vm.por.skip_static_finals=false

# When an object's constructor has returned, then the final fields can not be changed.
# Thus, instance final fields can never be included in a race condition (unless the this reference is leaked from a constructor)
# Defaulted to false to maintain backward compatibility in JPF
vm.por.skip_constructed_finals=false

# do we ignore explicitly set Thread.UncaughtHandlers
vm.ignore_uncaught_handler=false

# do we treat returned Thread.UncaughtHandler.uncaughtException() calls as normal thread termination
vm.pass_uncaught_handler=true

# do we reclaim unused memory (run garbage collection)
vm.gc = true

# threshold after which number of allocations to perform a garbage collection
# (even within the same transition, to avoid lots of short living objects)
# -1 means never
vm.max_alloc_gc = -1 

# do we run finalizers on collected objects (only makes sense with garbage collection)
vm.finalize = false

# this is a preemption boundary specifying the max number of instructions after which we
# break the current transition if there are other runnable threads
vm.max_transition_length = 5000

# are thread ids of terminated threads with recycled thread objects reused when creating new
# threads. This is required for programs that sequentially create many short living threads
vm.reuse_tid = true

# do we want to halt execution on each throw of an exception that matches
# one of the given prefixes, w/o looking for exception handlers?
# (useful for empty handler blocks, over-permissive catches or generally
# misused exceptions)
#vm.halt_on_throw = any

# class that is used to create scheduling relevant choice generators.
# this will replace the scheduler
vm.scheduler_factory.class = gov.nasa.jpf.jvm.DefaultSchedulerFactory

# print output as it is generated during the search (for all paths)
vm.tree_output = true

# collect output inside the stored path (to create program trace outout)
vm.path_output = false

# do we want to store the whole path no matter if we report them
# MP: change this before adding a listener
vm.store_steps=false

# untracked property
vm.untracked = true

# from where do we initialize the system properties
#  selected - keys specified as vm.system.properties, values from host
#  file - Java properties file (key=value pairs)
#  host - all system properties from underlying host VM
vm.sysprop.source = selected

# pathname of property file with system properties
#vm.sysprop.file =

# list of key names to load from host VM
#vm.sysprop.keys =

# class we use to model execution time
vm.time.class = gov.nasa.jpf.jvm.SystemTime

# if this is set to true, we throw an exception if we encounter any orphan native peer methods
vm.no_orphan_methods = false

############################### 3. CG part ###############################

#"def" the default value of randomize choices is not randomized
#"path" randomizes choices using preset user-specified seed
#"random" randomizes choices using a different seed at each trial
cg.randomize_choices = def

# if this is set, we create choice generators even if there is only a single
# choice. This is to ensure state storage/matching at all locations where a
# choice generator *could* be created. The default should be to turn it off though,
# since this can produce a lot of additional states (esp. with threads)
cg.break_single_choice = false


# default BooleanChoiceGenerator sequence: do we start with 'false'
cg.boolean.false_first = true

# do we want java.util.Random. nextXX() enumerate choices, or just return a single value?
# (isn't implemented for all types yet)
cg.enumerate_random=false

# maximum number of processors returned by Runtime.availableProcessors(). If this is
# greater than 1, the call represents a ChoiceGenerator
cg.max_processors=1

# creates a CG upon Thread.start, i.e. breaks the starting transition. Note this is
# required for data race detection (which depends on detecting access of shared objects)
cg.threads.break_start=true

# if this option is true, CG is not created when yield() (Thread API)
# is called
cg.threads.break_yield=true

# if this option is true, CG is not created when sleep() (Thread API)
# is called
cg.threads.break_sleep=true

# set if we shold also break on array instructions, e.g. to detect races
# for array elements. This is off by default because it can cause serious
# state explosion
cg.threads.break_arrays=false

# do we support atomic sections. If set to false, Verify.beginAtomic()/endAtomic()
# will not do anything
cg.enable_atomic=true

############################### 3. Report Part ###############################
log.handler.class=gov.nasa.jpf.util.LogHandler

# Windows seem to have a different default
log.level=warning

report.class=gov.nasa.jpf.report.Reporter
report.publisher=console

report.console.class=gov.nasa.jpf.report.ConsolePublisher

# this prints out repository information if the 'jpf' topic is set (for debugging)
#jpf.report.show_repository=true

#property violation reporting is configured separately
report.console.start=jpf,sut

report.console.transition=
report.console.constraint=constraint,snapshot

report.console.property_violation=error,snapshot
report.console.show_steps=true
report.console.show_method=true
report.console.show_code=false

report.console.finished=result,statistics

#jpf.report.console.show_steps=true
#jpf.report.console.show_method=true
#jpf.report.console.show_code=true

report.xml.class=gov.nasa.jpf.report.XMLPublisher

report.html.class=gov.nasa.jpf.report.HTMLPublisher
report.html.start=jpf,sut,platform,user,dtg,config
report.html.constraint=constraint
report.html.property_violation=
report.html.finished=result,statistics,error,snapshot,output


############################### 4. Listener part #############################

# imperative list of listeners
#listener=
#listener+=,.listener.SimpleDot  # MP: enabled vm.store_steps for using listeners

listener.autoload=\
  gov.nasa.jpf.NonNull,\
  gov.nasa.jpf.Const
    

listener.gov.nasa.jpf.NonNull=gov.nasa.jpf.tools.NonNullChecker
listener.gov.nasa.jpf.Const=gov.nasa.jpf.tools.ConstChecker


### PreciseRaceDetector

# we don't check for races in standard libraries
race.exclude=java.*,javax.*


############################### 5. test part #############################

test.report.console.finished=result