Release Output module feature

[adrianarce-elemwave]

No description provided.

[lmdiazangulo]

I think there is an issue with the sampling period being ignored for movies. Could you review that? We should add some specific tests to ensure this works properly.

My conversation with copilot:

User: This branch is changing the way outputs are provided. I think that samplingPeriod is being ignored for movies. Is it true?

GitHub Copilot: I’ll trace samplingPeriod from input parsing through movie scheduling and compare it with the nearby output implementation. The key check is whether movie writes are gated by that value or only by the global timestep loop.

The parser and movie module both contain the concept, so the likely control point is the new update_movie_probe_output path rather than JSON parsing. I’m narrowing in on the parsed domain fields, the output type’s timing state, and the exact write guard now; that will distinguish “parsed but unused” from “applied under a different name.”

The evidence is now fairly direct: readVolProbe copies the JSON sampling period into VolProbe_t%tstep, and preprocessing copies it into the observation schedule. But the new movie updater increments nTime and saves on every solver call, with no visible check against this%domain%tstep; I’m checking preprocessing and the save helpers once to confirm whether scheduling is supposed to happen upstream.

Yes. In this branch, samplingPeriod is parsed and copied into the movie’s domain%tstep, but it is not used when sampling.

• JSON parsing stores it in VolProbe_t%tstep in smbjson.F90:1793-1795.
• Output preprocessing carries it into domain%tstep in output.F90:550-565.
• However, update_movie_probe_output increments nTime and saves data on every solver timestep in movieProbeOutput.F90:256-272, without checking domain%tstep, tstart, or tstop.

The movie save helpers then record the current solver time directly, for example movieProbeOutput.F90:682-687.

So a movie currently samples at the simulation timestep, effectively ignoring samplingPeriod. The existing movie test does not catch this because it only runs one timestep.

[lmdiazangulo]

The ngspice submodule should point to the one in the dev branch. I believe there is no need to change it in with this PR.