Completion Scripts Should Follow Standard Naming Conventions

[ssanumand]

The generated ZSH completion script in zsh_completions.go uses camelCase for local variables (e.g., lastParam, lastChar, flagPrefix, requestComp, noSpace, keepOrder, hasActiveHelp) rather than the standard snake_case convention used in ZSH scripts.

While functionally correct, this deviates from idiomatic ZSH and creates inconsistency with how function names are already handled (e.g., __%[1]s_debug, _%[1]s). Using snake_case for variables would make the generated scripts more consistent with other ZSH completion scripts and more familiar to the ZSH community.

I'm happy to submit a PR for this change if it would be welcome. Opening this issue first to confirm whether the current naming was an intentional choice.

PS: I'd also be happy to do the same for other shells.

[marckhouzam]

Thank you for the suggestion. Personally I don’t feel it’s worth the effort. The generated script looks very different than normal zsh completion scripts already, users don’t need and should not modify it anyway.

[ssanumand]

Thanks for considering it, @marckhouzam! I completely understand the pragmatic view, users shouldn't need to modify generated scripts, and the effort-to-benefit ratio matters.

I'd suggest the change primarily benefits maintainability rather than user modification. When developers debug completion issues or contribute improvements to the generator itself, idiomatic variable naming reduces cognitive friction when reading the generated output alongside other ZSH code. The implementation is straightforward (essentially find-replace in the template), and it aligns the generated code with the naming pattern already used for functions.

That said, I respect that you know the codebase and community needs better than I do. If you feel the current approach works well, I'm happy to close this and move on. Just wanted to clarify the reasoning in case it changes the calculus.

Either way, thanks for maintaining such a valuable project!