drivers: current_sense_amplifier: fix zero-current-voltage binding

[Finomnis]

Fixes #99453.

It seems the new microvolt/microamp rework (#95588) simply forgot to implement zero current voltage offsets.

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

[Finomnis]

@JordanYates how did multiple reviewers miss that? Things like this make me kind of uneasy about the code quality of Zephyr...

I mean, to visually compare the new implementation with the previous implementation is the least a reviewer should do, and imo it is quite hard to miss...

[JordanYates]

@JordanYates how did multiple reviewers miss that?

Presumably because the problem was not with lines of code removed from the PR, the difference can only be caught it you compare line by line the swapped out function. For my part, the logic was written before the offset binding PR was opened.

Things like this make me kind of uneasy about the code quality of Zephyr...

A niche feature of a niche driver broke on one release, its not ideal but it happens. If you can find an embedded framework that supports as much as Zephyr does with better testing coverage, I'm sure we'd all like to hear about it.

I mean, to visually compare the new implementation with the previous implementation is the least a reviewer should do, and imo it is quite hard to miss...

People have limited time, I absolutely do not blame any reviewer for missing this. The great part about Zephyr is that anyone can contribute. Since the problem was obvious to you, it would be great if you could help us proactively review PRs.

[Finomnis]

@JordanYates Fair.

I should probably switch to the release candidates earlier to raise bugs before the official release.

It's just frustrating because this already happened to me in v4.2 where we switch to it with our project and immediately stuff breaks. Last time was #93519, again because somebody "swore they already covered that cornercase that but it must have been lost during merges", which again, makes me a little uneasy about the project reliability ... It's been a while since I've used an actual release and not some commit hash in a vx.x-branch.