Preserve selected data server in movie frames

[Esteban82]

Done with Claude Sonnet 5

Fix #8209.

When running gmt movie, the selected GMT_DATA_SERVER was not consistently preserved inside the frame sessions. As a result, gmt --show-dataserver could report a different server in the movie script than the one configured in the parent session.

This PR ensures that the active data server setting is propagated correctly to movie frame sessions.

Fix: swapped the order at all 6 locations where scripts are generated (preflight, postflight, intro title-frame loop x2, main script x2), so cp -f gmt.conf . now runs before gmt begin.

Test

The issue can be reproduced with:

gmt set GMT_DATA_SERVER oceania
gmt --show-dataserver
cat << 'EOF' > main.sh
gmt begin
	gmt --show-dataserver
	gmt basemap -R-114/-103/35/40 -JM10c -Baf
gmt end
EOF
gmt movie main.sh -Chd -Nanim02 -T1 -Zs -Ml,png

Expected result: the same data server name should be printed twice (once in the terminal and once from inside the movie script). For example:

http://oceania.generic-mapping-tools.org
http://oceania.generic-mapping-tools.org

[Esteban82]

I see this errors for the movie tests. Is it safe to merge it?

Start 738: test/movie/trajectory2.sh
67/75 Test #738: test/movie/trajectory2.sh .............***Failed    0.88 sec
Set GMT_SESSION_NAME = 83801
/usr/bin/gm compare: image difference exceeds limit (0.00886259 > 0.003).
test/movie/trajectory2.ps: RMS Error = 0.0089 [FAIL]
memtrack errors: 0
exit status: 1

      Start 739: test/movie/trajectory3.sh
68/75 Test #739: test/movie/trajectory3.sh .............***Failed    0.91 sec
Set GMT_SESSION_NAME = 83877
/usr/bin/gm compare: image difference exceeds limit (0.00737297 > 0.003).
test/movie/trajectory3.ps: RMS Error = 0.0074 [FAIL]
memtrack errors: 0
exit status: 1

      Start 740: test/movie/trajectory5.sh
69/75 Test #740: test/movie/trajectory5.sh .............***Failed    0.91 sec
Set GMT_SESSION_NAME = 83953
/usr/bin/gm compare: image difference exceeds limit (0.0101027 > 0.003).
test/movie/trajectory5.ps: RMS Error = 0.0101 [FAIL]
memtrack errors: 0
exit status: 1

[joa-quim]

Hmm, need to look in what those differences consist.

[Esteban82]

The errors in the tests are because -Ar<dpu> falls back to the session's PROJ_LENGTH_UNIT, which now depends on the conf/begin order this PR swaps. Pinned trajectory2/3/5.sh to -Arc to match the existing baseline.

Maybe would be better to update the baselines, but I don't know very well how to do it.

[Esteban82]

Now the tests for trajectory2/3/5.sh passed. I think I can merged it.