* Frame Log DRG now manage their own jobs (no longer add a phantom restart job)
* Actually removed frame log drg restart jobs this time
* Refactored to avoid duplicate code
* Refactor
* Fixed DR constructors
* Fixed some syntax stuff centos was throwing a tantrum about
* More centos whining
* Centos...
* Centos...
* Clean up
* Added a check before calling MM delete_var in ExternalApplication destructor;
Made sure that the MulticastGroup is not initialized before initializing it in VariableServerListenThread.cpp;
* Updated to call multicast group initialization to be consistent for the unit test.
* Removed unnecessary command c str pointer.
* Removed unnecessary command c str pointer.
* Added optional remote sheel config file for the remote startup command.
* Updated to cd to the sim directory after sourcing a config file in case the config file changed pwd.
* Speed up Trick::ScheduledJobQueue::push
* Make comparator a static function
* Use upper_bound instead
* Use explicit types
* Update comment
* Fix formatting
* Added TRICK_LDFLAGS to the rule for linking shared library so user can use TRICK_LDFLAGS for linking shared library if necessary.
* Removed the new line added by accident.
* Repurposed -OO so all sim run outputs including S_sie.resource are saved to the specified folder. The files are placed in the sub-dir either RUN_xxx or DP_Product of the specified folder.
* Updated for repurposed -OO.
* Added quoting code markdown for such as in order to show <name> as is.
* Added quoting code for missing items.
* Updated to raise an error when the --read-only-sim flag is used without the -O or -OO. Also made updates to the related document accordingly.
* Fixed the error message to be more clear.
* Proposed fix for deadlock on shutdown
* Terminate C style comment.
* Added needed libs for applicable tests and updated the logic for when allowing/disallowing connections.
* Need to load additional trick libs for applicable tests for Linux.
---------
Co-authored-by: Hong Chen <hong.chen-1@nasa.gov>
* Implemented use of the python GIL API to replace the less portable mutex solution.
* Replaced nullptr with NULL
---------
Co-authored-by: Thomas Brain <thomas.a.brain@nasa.gov>
* Added additional resource usage info for sim run.
* Fixed so voluntary context switches output uses corresponding ru_nvcsw instead of ru_nivcsw due to copy and paste error.
* Removed added resource usage data for page faults and block operations due to inconsistency on different OS.
* Updated to have voluntary and involuntary usage data for both initilization and run for sim shutdown run summary.
* Repurpose the "Sim/Real Time" field on the sim control panel #1626
Added calculations in the realtime sync monitor job to calculate the
average sim/realtime ratio for the past 100 frames. This is saved to
a new variable, actual_run_ratio. Changed the sim control panel to
display the actual_run_ratio value.
* Repurpose the "Sim/Real Time" field on the sim control panel #1626
how does this compile locally but not in CI? Didn't include cmath.
* Updated to log applicable time in seconds instead of tics, remove scale attribute in dp product XMLs, and reset clock before realtime clock sync to avoid logging epoch time for frame logging.
* Removed "_seconds" from applicable time variable names.
* Add type, dispersion, min_value, max_value and other relevant internal
members of MonteCarloVariable* classes to the output of
MonteCarlo_Meta_data_output. Motivation is for users wanting to post-process
dispersion parameters used during generation of runs
* Protect against invalid memory access when length of values is zero in
MonteCarloVariableRandomStringSet::generate_assignment(). Add a new
verif sim warning case to cover these new lines
* Update new verif data for SIM_mc_generation to support these changes
Closes#1574
Co-authored-by: Dan Jordan <daniel.d.jordan@nasa.gov>
* Added system CPU time and initialization system CPU time and added user wording for existing CPU time to distinguish it from system CPU time used for trick run summary.
* Updated the order of a couple of times of shutdown messages printed on screen and sim/cpu time ratio to both user and system cpu time.
* Minor change for lining up the shutdown message.
* Added input file SHA1 info being printed on screen when running a trick sim with -d option (basically verifying the input)
* Added additional checks to exclude trace info printout for file in /opt or .trick/ or not a real file such as <xxx> as the filename.
* In DPV_textbuffer.cpp, close the opened file on read failure, so we don't have a resource leak. #1561
* Disable 2 tests in VariableServerSessionThread_test.cc: exit_if_handle_message_fails, and exit_if_write_fails.
* Possible speedup for requesting sie file #1555
When creating the sie file there is a marker string that Trick searches
for where it writes the run time allocated memory. The loop that was
searching for the string was iterating one character at a time and reading
from file each time. This was slow. Changed the loop to read 1Mb into
memory at a time and search for the string. Handled the case where the
marker string could straddle the 1Mb boundary by copying a small portion
of the previous part of the file for the next search. For my one test
point sie file generation dropped from 90+ seconds to 3.5 seconds.
* Possible speedup for requesting sie file #1555
Math was wrong on some offset values. Fixed them now.
* Possible speedup for requesting sie file #1555
The mac won't allow me to declare a "char buff[1000001];". It compiles
but throws an exception when run. Changed it to "char *buff = new char[1000001];"
* Added SIM RAM usage info to trick run summary.
* Fixed to use #if for determing which OS.
* Removed the system print out statement when MonitorHealthStatusTask is finished.
* Changed "M" to "MB" for RAM usage info on run summery.
* Fixed SIGABRT for checkpointing change variables
* Expanded the Data Recording Checkpoint test sim
- Added a case for checkpointing change variables
* Updated the test case for checkpointing change vars
- Changed the header, input, and log files to get better data representation
* Expanded the Checkpoint test sim
- Added a test case covering when the only tracked variable is the change var.
* Variable Server byteswapping crashes the sim #1513
The code to byteswap a variable server buffer has a comment saying there is a bug. The comment is
correct. The original code would swap the parameter and anything else that followed that parameter
in the structure. Crashes everywhere. Created a new routine that byteswaps a single parameter.
Strangely we didn't have such a routine until now. Did some testing of doubles, floats, ints,
shorts, and chars and all were swapped correctly.
* Variable Server byteswapping crashes the sim #1513
enabling binary byteswap test.
* Add unittests to cover uncovered code in MemoryManager ADEF_ and REF_ parsers.
* Remove the second production of the 'user_defined_type' target
because it is superfluous, and impossible to reach. This is
because the NAME token is defined as pattern : '[_a-zA-Z][_a-zA-Z0-9:]*'.
This pattern will always match a string of colon separated names.
Note that the NAME itoken is used for colon separated type-name as well as
colon separated variable names.
* Add a test to MM_declare_var_unittest that covers the HEX token in adef_parser.l.
* Add a unittest for MemoryManager_JSON_Intf.cpp.
* Tweak MM_JSON_Intf.cc to get better code-coverage.