v1.0.21 of WInnForum test harness for SAS UUT testing

This is a maintenance release to update v1.0.20 which was released on 20 November 2018. The following issues were fixed:

Issue	Title	Description
#861	Relax the checks on xPR configs (still matches spec)	Previously, we required that the extra (non-utilized) protected entity fields be absent in an xPR config; now we permit them as long as they do not actually contain any entities. For example, an EPR config may now also contain the field 'gwpzRecords' so long as there are no actual GWPZ records configured. This change is to fulfill a request by ITS which will allow them to leverage their existing MCP config generation code to create valid xPR configs. This PR does not contain any substantive changes. To quickly double-check that your custom configs remain valid after this change -- which should be the case for all previously-valid configs -- put a return statement on line 225 to bypass actual execution of MCP (i.e. the test will exit immediately after checking the config). I have already done this with the WinnForum default configs, the latest ITS draft study guide, and Google's internal tests.

v1.0.20 of WInnForum test harness for SAS UUT testing

This is a maintenance release to update v1.0.19 which was released on 13 November 2018. The following issues were fixed:

Issue	Title	Description
#860	small fix of move list when CBSD position is at same location of the protected point	fix move list when CBSD position is at same location of the protected point, the calculation of pathloss return a value(0), so for this case we form an array(with same length as reliabilities) of interferencex of the same value.
#859	update if PPA cbsd_reference_id is empty array in sas_test_harness.py generatePpaRecords()	FAD.2 generates default.config with empty "cbsdReferenceId": []. According to WINNF-TS-0096 v1.3.0 SAS-SAS Protocol spec, section 8.5.1 PPA Information Object definition mentions that cbsdReferenceId field is a "List of one or more CBSD Reference IDs in the cluster list of the PPA". So ppa_record['ppaInfo']['cbsdReferenceId'] should also be updated if it's an empty array.
#858	Revert "PCR - add buffering to remove gaps between census tracts ( fixes #814 )"	Reverts #852. As decided in the sas-test-dev group, this PR reverts the PR handling the issue #814. See email thread in sas_test_dev mail list.
#857	for FAD_1, consider lat and lng equal if same to 6 decimal places ( Fixes #822 )	Fixes #822 Adhering to SAS-CBSD specification WINNF-TS-0016, Version V1.2.2, 1 October 2018 which says: "Latitude of the CBSD antenna location in degrees relative to the WGS 84 datum [n.11]. The allowed range is from -90.000000 to +90.000000. Positive values represent latitudes north of the equator; negative values south of the equator. Values are specified using 6 digits to the right of the decimal point." This should not affect the latitude or longitude precision elsewhere.
#856	Remove obsolete census file.	This file is obsolete, was here to support default config in PCR and WDB while the SAS-Data database was being built. And should have been removed once that one became active. Because it stayed there, it now creates confusion such as expressed in issue #814, since it was generated by an old process with some defects.
#855	Tiny fix to treat not handledFSS Scenario	Fix tiny bug for MCP test when the config contains multi FSS at least one of them is for FPR.3 scenario(i.e FSS with TTC false and out of band). spec : [WINNF.FT.S.FPR.3] Multiple CBSDs from Multiple SASs Inside and Outside the Neighborhood of an FSS Station for “FSS Scenario 2” with TT&C Flag = OFF
#854	DPA: improve logs for enabling post processing eval	Some logs were introduced in DPA CheckInterference for better post analysis of result in #830. This PR improves this logging according to various report received since then: (1) output of CSV logs put into harness/testcases/output/; (2) output the neighbor list in addition of the keep list, to allow full post processing analysis (for example redo random movelist with ref model and anlayse agg interference variation); (3) fix issues with SAS UUT keep list to be the keep list per channel and not all SAS active grants in whole US for all channels. This means we run the active grants through a neighbor filter, just used for that; (4) main test harness script can now be run in debug mode by adding option: --log_level=debug
#852	PCR - add buffering to remove gaps between census tracts ( fixes #814 )	Adds buffering to remove gaps between census tracts for PCR. Fixes #814. Based on work by @chandanahs

v1.0.19 of WInnForum test harness for SAS UUT testing

This is a maintenance release to update v1.0.18 which was released on 6 November 2018. The following issues were fixed:

Issue	Title	Description
#847	Fix mem Leak on ITM	Affects the calculations with array or reliabilities (like DPA), and average computed value. The mem leak is expected to be about 16kB for a DPA CBSD-Point link, and 800 bytes for a CBSD-point link in other protection type. It is thus quite bug for DPA and affect negatively any test of the test harness
#845	fix MCP injection of gwbl records

v1.0.18 of WInnForum test harness for SAS UUT testing

This is a maintenance release to update v1.0.17 which was released on 30 October 2018. The following issues were fixed:

Issue	Title	Description
#844	Peer fccid injection for IPR testcases	Added peer fccid injection for IPR testcases. This is related to Whitelist fccIds by injecting before FAD Issue #810
#843	Wait until the very end to fail tests which check aggregate interference to DPAs	This allows us to collect the maximum amount of information for later analysis.
#842	Improve portability of JSON schema checking code	Fixes #840.
#841	Add TCP keepalive to all test harness requests	As suggested by Sony, this change prevents the connection from dropping which would cause a test failure.
#839	Filter FSS with high frequency below 3700 MHz for purge list calculation	Do not apply FSS purge list when FSS high frequency is below 3700 MHz
#837	DPA: Add new modes for DPA margin.	As discussed in latest sas-test-dev meeting, two modes are added: + 'target (margin_db)': using the absolute target + margin_db + 'linear (margin_db)': using a linear margin (in mW) computed from the equivalent margin_db at the threshold. Sucess if interference below ref_model_interference_mw + margin_mw This corresponds respectively to option 1 and option 2. To configure these modes, simply replace the ''movelistMargin" tag in the test config, for example: "movelistMargin" : "linear (1.5)" This PR also fixed a few bug in the _PrintStatistics() routine: - was failing if an input threshold is integer (normally not the case in test harness, but possibly) - was failing in one condition (peer sas with no grants in neighborhood) as the test was using the wrong A_DPA.

v1.0.17 of WInnForum test harness for SAS UUT testing

This is a maintenance release to update v1.0.16 which was released on 25 October 2018. The following issues were fixed:

Issue	Title	Description
#836	Fix error creating PAL record for FIPS codes with leading zeroes	Fixes issue #834.
#835	Batch conditional reg data	Without this change, the injection of conditional registration data does not follow the batching methodology we've established for other APIs.

v1.0.16 of WInnForum test harness for SAS UUT testing

This is a maintenance release to update v1.0.15 which was released on 16 October 2018. The following issues were fixed:

Issue	Title	Description
#832	Fix DPA logging for "no grants in neighborhood" case	The output of calcAggregatedInterference() is effectively a scalar in the case where there are no CBSDs in the neighborhood, and this case was not accounted for in the previous version of the logging code. This fix addresses that issue, and more gracefully catches errors in the new logging code so that the test case may proceed even if the logging fails (though an error will be logged, of course).
#831	Miscellaneous tool and doc provided to ITS by mail.	For the benefit of all, some little tools and docs provided to ITS are in the repository: (1) doc on the DPA setup in MCP/IPR configs, using the dpa builder; (2) the catA outdoor HAAT tool to check maximum allowed height.
#830	Add more logging of A_DPA, A_DPA_ref, and keep lists to DPA manager	There are no functional changes in this PR. Instead, it adds logging for A_DPA and A_DPA_ref for every DPA point at every azimuth. It also adds logging of statistics on their difference.
#829	Update DPAs KML from NTIA	Update DPAs KMLs: E-DPA to v1.07, P-DPA to v1.06. As provided by Nick from NTIA. The difference with previous DPAs are: (1) using the default value (instead of NaN) for catA distance (150km), and OOB CatB (25km) and OOB CatA (0km, ie deactivated); (2) portal DPA "Barking Sands" has now its CatB distance set properly; (3) CatA distance are set to not exceed CatB distances for West9: 144km, Alameda: 128km, Bremerton-Everett: 128km. Also put the PDF file describing the DPA KML format.

v1.0.15 of WInnForum test harness for SAS UUT testing

This is a maintenance release to update v1.0.14 which was released on 9 October 2018. The following issues were fixed:

Issue	Title	Description
#827	DPA: Fix azimuth range to include the max_azimuth	As reported by Red Technologies. Includes the max_azimuth in the list of azimuth to check. Make sure it does not go beyond the max azimuth and different of first azimuth (case of 0-360 min/max).
#826	Add additional debugging information to some of the protocol tests.	There are no functional changes in this PR. Instead, it proposes to collect information on multiple failures before halting the test, allowing for faster debugging in the presence of multiple errors. It also logs additional information which may be useful during debugging.

v1.0.14 of WInnForum test harness for SAS UUT testing

This is a maintenance release to update v1.0.13 which was released on 2 October 2018. The following issues were fixed:

Issue	Title	Description
821	EHata: Cleanup Average terrain code for the NTIA compatibility mode.	Note: no effect on the code used by test harness, this is a cleanup. The compatibility mode (option WinnforumExtension=False) replicating ITS implementation has been upgraded to latest version found in NTIA reference version (as was done in NTIA/ehata@80954b1). This has no effect on test harness which used WinnforumExtensions=True, but removes confusion when comparing the Winnforum version versus the NTIA version (like was reported recently by ITS). In some sense, it also documents precisely what variation we have now with the NTIA ref code: just the use of ceil() in this average terrain calculation (in addition of the clipping of the effective height in ExtendedHata.cpp which differs as per the spec).
820	EHata: Cleanup of the float -> double switch.	Previously to avoid having the code base differing too much from the original ITS reference implementation, we defined a global #define float double in the ehata.h header file. Obviously this is very hack and can fail depending on the order of inclusion of ehata.h, or if t is forgotten in one of the cpp file. The proper way is now implemented by directly changing all float occurence with double. This is in line with the change made by ITS on their reference implementation at: NTIA/ehata@df2b580
819	EHata: Fix Fine Rolling Hill correction.	When the rolling hill is below 10m, the correction should be zero. Otherwise it does not follow the EHata spec on one hand, plus it introduces some randomness and discontinuity in the correction of 2dB or 4dB when the terrain is very flat near the mobile (ie a microscopic change of altitude can cause up to 4dB variation in pathloss). This change is make to reflect the proposed change in the ITS original code base, which was integrated recently at: NTIA/ehata@b54b71f Further documentation on the root issue. See Ticket on NTIA GitHub: NTIA/ehata#7 Issue #333
817	DRG.6 remove registered CBSDs from second preload conditional data	Update DRG 6 so the test doesn't send preload conditional data for registrated CBSD(not successfully deregistered CBSDs)

v1.0.13 of WInnForum test harness for SAS UUT testing

This is a maintenance release to update v1.0.12 which was released on 25 September 2018. The following issues were fixed:

Issue	Title	Description
813	Update FDB 5 and 6 expiration dates to 2028-09-23.	Currently the FDB.5 and FDB.6 default GWBLs have expired. This PR moves the expiration date into the future by a decade (2028-09-23). By that point this testcase should well and truly be deprecated, and if not the test will also print a log to make it easier on anyone trying to debug why this test could be failing.

v1.0.12 of WInnForum test harness for SAS UUT testing

This is a maintenance release to update v1.0.11 which was released on 18 September 2018. The following issues were fixed:

Issue	Title	Description
812	Introduce a flag to turn on/off early-exit in PAT.	The current test code causes the PAT test to fail as soon as the failure threshold is reached. While this can save time if you want only a pass/fail result for a SAS, it makes debugging difficult: the debugger typically wants to get the maximum amount of data from each test run, and in that case the early-exit feature is counterproductive. ITS recently noted the drawbacks of our overly-eager optimization, so this PR introduces a flag which lets them toggle between the two modes (with the default being to disable early-exit).
811	Update MCP testcase to ignore failed registrations	ITS has requested this change to account for differences in SAS behavior during registration. The change allows the test harness to ignore registration failures and continue test execution, whereas previously a registration failure would immediately cause a test case failure. Registration failures will be logged so that unexpected or widespread registration failures can be detected. This change is consistent with the test spec.
810	Whitelist fccIds by injecting before FAD	FccIds in the SAS TH are unknown to the SUUT and this could cause problems for some SUUTs. These changes will inject the fccIds when the SAS TH is loaded with a FAD .
809	Fix time zone issue in FDB.8	With the existing code, current_time correctly uses PDT but scheduled_cpas_start_time and scheduled_cpas_end_time uses LMT, which ultimately causes the test harness to wait too long for CPAS to complete. This change corrects the timezone discrepancy.