NormAngle gets stuck on certain compilers when the angle input is infinity

[kmchen12]

double NormAngle(double angle)
{
	while (angle < -PI) angle += PI2;
	while (angle >= PI) angle -= PI2;
	return angle;
}

When compiling on Windows,
I frequently encounter an issue with Satellite No. 59 where the calculation:
a = 1 / (2 / r - v2 / (CGCS2000_SQRT_GM * CGCS2000_SQRT_GM)); results in a negative number.

This causes an inf value to be passed into Alm->w = NormAngle(w);, which hangs the program.
Interestingly, this issue does not occur on WSL using the exact same settings."

[globsky]

This equation is used when trying to recover BDS GEO satellites almanac from ephemeris. The GEO ephemeris uses self defined coordinate instead of ECEF, so the position and velocity in ECEF coordinate should be calculated first and then use these values to calculate orbit parameters. The equation you mentioned here is used to calculate the major semi-axis and should not be an negative value. Could you provide the ephemeris of SV59 you use? Better with the version of the C compiler on Windows.

[kmchen12]

I am using BRDC00IGS_R_20260100000_01D_MN.rnx, and the compiler is

  -- Building for: Visual Studio 17 2022
  -- Selecting Windows SDK version 10.0.22621.0 to target Windows 10.0.26100.
  -- The C compiler identification is MSVC 19.44.35213.0
  -- The CXX compiler identification is MSVC 19.44.35213.0
  -- Detecting C compiler ABI info

[globsky]

I tried with the the same RINEX file and Visual Studio 2022 and it doesn't shows any problem. I assume you set the scenario time to 2026/01/01 00:00:00, so the ephemeris of SV59 been used at that time is:
svid 59 ';' unsigned char
toe 345600 int
week 1043 int
M0 1.1145945518600000 double
delta_n -1.0351145452500000e-08 double
delta_n_dot 0.0000000000000000 double
ecc 0.00017518887762000001 double
sqrtA 6493.3740520499996 double
axis_dot 0.0000000000000000 double
omega0 -3.0946436859499999 double
i0 0.040945440033399998 double
w -1.7901278088899999 double
omega_dot 1.1379045411499999e-08 double
idot -9.1146653766799996e-10 double
cuc 1.4104880392599999e-05 double
cus 1.5016645193099999e-05 double
crc -456.18750000000000 double
crs 424.57812500000000 double
cic 1.5506520867300000e-07 double
cis 5.3085386753099997e-08 double
af0 -1.5133991837500001e-07 double
af1 2.1316282072799998e-14 double
af2 0.0000000000000000 double
satellite position and velocity (in inertial coordinate) calculated at time 345600 is:
x -32325249.589090198 double
y 27035365.678121176 double
z 1282420.3896694093 double
vx -1974.4141661812885 double
vy -2354.9696685678878 double
vz -107.40928138889453 double
and with variables r=42160139.473234057 and v2=9455730.1932205781 then results
a 42165900.726037666 double
I do not know at which step you got incorrect value.