memory safety: prevent duplicate close and invalid pattern reads

[KJ7LNW]

Description

Copied plot cards shared a raw FILE*, causing multiple destructors to close the same stream. Average-only radiation patterns also emitted per-angle rows from arrays that mode intentionally leaves uninitialized.

This change gives copied plot cards shared ownership of their stream and suppresses per-angle output when the RP card requests average-only reporting, while preserving the heading and average summary.

Type of Change

• Bug fix (non-breaking change that fixes an issue)

Implementation Details

• Store plot streams in std::shared_ptr<FILE> with fclose as the deleter.
• Access the raw stream only at C standard I/O boundaries.
• Name the RP average-only mode and omit its unpopulated per-angle rows.

Testing

• CMake Debug build completed.
• CTest passed necpp_unit and necpp_smoke_hertzian_dipole (2/2).
• All three regression decks exited normally under GDB after the fix.

20-12.nec: plot-current stream ownership

• Before: Copied plot cards retained the same raw stream, so result destruction called fclose more than once on it.
• After: The deck exits normally, writes both plot files, and the final shared owner closes each stream once.

CM Currents and radiation patterns
CM PL demonstration 1
CE
GW 1 11 0 -.2375 0 0 .2375 0 .001
GE
FR 0 1 0 0 299.7925 1
EX 0 1 6 00 1 0
PL 1 2 0 m20-12-currents.dat
XQ
PL 3 2 0 3 m20-12-rp0.dat
RP 0 1 361 1000 90 0 1.00000 1.00000
EN

20-13.nec: near-field stream ownership

• Before: The near-field result retained a copied raw stream, so teardown called fclose after another plot-card copy had closed it.
• After: The deck exits normally, writes near-field plot output, and the final shared owner closes the stream once.

CM PL demonstration 1
CM Near fields
CE
GW 1 11 0 -.2375 0 0 .2375 0 .001
GE
FR 0 1 0 0 299.7925 1
EX 0 1 6 00 1 0
PL 2 1 4 0 m20-13-NF.dat
NE 0 11 1 1 -5.0 -5.0 0.0 1.0 1.0 1.0
NH 0 11 1 1 -5.0 -5.0 0.0 1.0 1.0 1.0
EN

15-9a.nec: average-only radiation pattern

• Before: A=2 left per-angle arrays unpopulated, but report generation still iterated those rows and read an uninitialized polarization-sense index.
• After: The deck exits normally, emits no per-angle rows, and retains the average power and solid-angle summary.

CM 1 wl above Average Ground
CM Copper #12 Dipole
CE
GW 1 41 0 -100 420 0 100 420 0.0404331
GS 0 0 .02540
GE 1 -1 0
GN 2 0 0 0 13 .005
EX 0 1 21 0 1 0
LD 5 1 1 41 5.8001E7
FR 0 1 0 0 28.5 1
RP 0 19 73 1002 -90 0 5 5
EN

Expected average-only summary:

AVERAGE POWER GAIN: 1.4811E+00
SOLID ANGLE USED IN AVERAGING: (2.0000)*PI STERADIANS