psinject follow-up from "Require test scripts for (most) CW codes in LALApps"

The following discussion from !1745 (merged) should be addressed:

• @david-keitel started a discussion: (+3 comments)

  I don't think I can open a thread on a file/dir in general, so putting it here inline: what does the move back have to do with the testing regime? I don't really mind where this code lives, except that "this code depend on FrameL directly, and so would have to stay in LALApps" would seem to me an argument to keep it separate from pulsar: because if all the other codes could eventually be moved into lalpulsar, wouldn't this stay as an odd orphan?

From: @karl-wette

It's somewhat of an unrelated change. I had to decide whether to try to write a test for lalapps_psinject, or else leave it exempt like the others. Ultimately I decided to move it back to src/hwinjection/ as well because:

1. lalapps_psinject isn't really a "CW code" in the sense of being (like everything in src/pulsar/) a CW search code, or a tool closely related to analysing data for CW signals. The generation of CW signals for the hardward injections is (as I understand) done by lalapps_Makefakedata_v4. What lalapps_psinject does is really the "hardware injection" part of interfacing with ... however it works 😄 This seems sufficiently different from our other CW codes that it should be in its own category after all.
2. While you and I are nominally the code owners for src/hwinjection/, that's really because there's no current maintainer. I don't personally feel I can maintain this code, given that I really have no idea how it works, or how to test it. I feel more comfortable being maintainer for codes in src/pulsar/, as I have a reasonable idea how most of them work, and could hack together a test if I really had do. That makes me think it should also be in its own category.

I'm not sure where the ultimate best home for this code is, whether in LALSuite or more closely integrated with operations software. It seems to require GDS, so perhaps it should live there? Perhaps more importantly, it needs a maintainer who understands how it works.

From @david-keitel :

I agree with all of that. Moving it elsewhere sounds preferrable, especially if it doesn't profit from being inside LALSuite's CI coverage anyway because it doesn't have a test...

From @evan-goetz :

Perhaps Keith Riles is in the best position to maintain this code in LALSuite? I don't know if there is a better spot for it at the moment and he is the one that is most familiar with the CW hardware injection system along with Dave Barker and Keith Thorne.