camera_app: reorder camera definition parameters for GCSs

[rmackay9]

This PR makes two relatively minor changes:

1. GCSs list parameters in the order they appear in the XML. This changes the order to put the live controls first, then thermal, streams, recording, photo and image, overlays, tracking and survey.
2. sitl: fix stale CAM_MODE assertion in camera definition test

My AI has tested this but I have not yet. I think this is an improvement but happy to discuss if another order might be better.

Here's the new order for the MT11

#	Before	After
1	CAM_MODE	CAM_MODE
2	SURV_LENS	CAM_ZOOM
3	SURV_PATTERN	CAM_OPT_ZOOM
4	SURV_OVERLAP	CAM_AUTOFOCUS
5	SURV_FORE	CAM_LENS
6	SURV_AFT	CAM_SOURCE
7	SURV_BURST	CAM_THERM_GAIN
8	SURV_BURST_MS	CAM_PALETTE
9	SURV_DWELL_MS	RAW_STREAM_FPS
10	SURV_SETTLE_MS	RAW_RECORD_FPS
11	SURV_ERROR_CD	VIDEO_MAIN_RES
12	SURV_RATE_CD	VIDEO_MAIN_CODEC
13	SURV_TERR_MS	VIDEO_SUB_RES
14	SURV_TERR_M	VIDEO_SUB_CODEC
15	SURV_TERR_TOL	REC_RESOLUTION
16	SURV_MIN_SPEED	REC_AUTOSTART
17	SURV_TURN_DEG	PHOTO_SCOPE
18	CAM_ZOOM	IMG_BRIGHTNESS
19	CAM_OPT_ZOOM	IMG_SATURATION
20	CAM_AUTOFOCUS	IMG_CONTRAST
21	CAM_LENS	IMG_EXPOSURE
22	CAM_SOURCE	IMG_ISO
23	CAM_THERM_GAIN	IMG_SHUTTER
24	CAM_PALETTE	IMG_METERING
25	PHOTO_SCOPE	IMG_WHITE_BAL
26	RAW_STREAM_FPS	OSD_CROSS
27	RAW_RECORD_FPS	OSD_RECORD
28	OSD_CROSS	OSD_THERMAL_FOV
29	OSD_RECORD	MAV_POS_TARGET
30	OSD_THERMAL_FOV	TRACK_METHOD
31	LOG_DISARMED	LOG_DISARMED
32	TRACK_METHOD	SURV_LENS
33	MAV_POS_TARGET	SURV_PATTERN
34	REC_AUTOSTART	SURV_OVERLAP
35	REC_RESOLUTION	SURV_FORE
36	VIDEO_MAIN_RES	SURV_AFT
37	VIDEO_MAIN_CODEC	SURV_BURST
38	VIDEO_SUB_RES	SURV_BURST_MS
39	VIDEO_SUB_CODEC	SURV_DWELL_MS
40	IMG_BRIGHTNESS	SURV_SETTLE_MS
41	IMG_SATURATION	SURV_ERROR_CD
42	IMG_CONTRAST	SURV_RATE_CD
43	IMG_EXPOSURE	SURV_TERR_MS
44	IMG_ISO	SURV_TERR_M
45	IMG_SHUTTER	SURV_TERR_TOL
46	IMG_METERING	SURV_MIN_SPEED
47	IMG_WHITE_BAL	SURV_TURN_DEG

[AP-Review]

Automated review note — AI-generated (Claude+Codex), validated against the live diff. Please sanity-check before acting.
Verdict: ACCEPT

Reviewed at head d8f940c754.
Full report: https://firmware.ardupilot.org/Tools/APReview/DevCallReviews/PRReviews/ardupilot/ap_cameragimbal/53/1.html#prAP_CameraGimbal-53

Thanks, this looks good to me.

I checked that the reorder only changes order. I exported the camera-definition XML for all five targets at base and at head: every target keeps the same parameter set, with identical per-parameter XML, and all names are within 16 characters. Merging the SURV_* guards into APCAM_HAVE_PHOTO && APCAM_HAVE_THERMAL (https://github.com/ardupilot/ap_cameragimbal/pull/53/files#diff-4026e5b010f9e7daf7c40fac6c3f24b6114af3a10b43e9437069225f6b023e9bR100) is equivalent to the old nesting. Because cam_definition_version is a CRC of the XML, it changes for every target, so GCS caches will refresh. make camera-definition-test passes.

The CAM_MODE assertion fix (https://github.com/ardupilot/ap_cameragimbal/pull/53/files#diff-9df17b86452451a64062886777d468b05ec4ea9dc416b9e42e3312be185566f2R147) is correct. The old assertion already failed on zr10 under SITL, and the new one matches the target flags and the host test.

One minor point, not blocking: CI only runs sitl/test_camera_definition.py with --targets mt11 --zoom-only, which skips protocol_checks. That is why the stale assertion went unnoticed. It may be worth widening that job at some point.

Not checked: hardware, and whether QGC or Mission Planner show parameters in XML order.