Merge lFormula/glFormula via shared mkFormula() helper

[Copilot]

lFormula() and glFormula() shared ~90% identical implementation, noted by DRY/TODO comments in the code. This PR extracts the shared body into a private mkFormula() helper, eliminating the duplication.

Changes

• New mkFormula() internal helper — contains all shared logic: model-frame evaluation, reTrms construction, identifiability checks, predvars extraction, and X matrix building/scaling. Behavioral differences are parameterized:

  • allow.n — FALSE for lmer, TRUE for glmer
  • check_zero_rows — lmer-only "0 (non-NA) cases" guard
  • check_na_Zt — lmer-only NA-in-Z check
  • set_varnames_fixed — lmer-only varnames.fixed attribute

• Additional cleanup from PR feedback — moved the remaining duplicated preprocessing from both public functions into mkFormula():

  • fr.form. and fr.form construction is now centralized
  • shared weights/offset environment assignment loop is now centralized
  • preserved original missing()-based argument detection behavior in caller frame

• lFormula() — now keeps only lmer-specific setup (arg checking, family redirect, match.call() construction), delegates preprocessing/build steps to mkFormula(), appends REML

• glFormula() — now keeps only glmer-specific setup (family handling, mustart/etastart model-frame args), delegates preprocessing/build steps to mkFormula(), sets start attribute on fr post-call, appends family

• Removed all DRY!, TODO: merge, and ## FIXME: does start= do anything? comments that prompted this change

Public signatures of both functions are unchanged.

Original prompt

---

This section details on the original issue you should resolve

<issue_title>lFormula() and glFormula() clean-up</issue_title>
<issue_description>The following comments in the code:

lme4/R/modular.R

Lines 712 to 713 in 0a7f766

	## TODO: remove any arguments that aren't actually used by glFormula (same for lFormula)
	## TODO(?): lFormula() and glFormula() are very similar: merge or use common baseFun()

suggest that we should merge lFormula() and glFormula(). Also, should remove unnecessary arguments. </issue_description>

<agent_instructions>Try to merge lformula and glformula() without breaking anything :-)</agent_instructions>

Comments on the Issue (you are @copilot in this section)

• Fixes lFormula() and glFormula() clean-up #3

---

✨ Let Copilot coding agent set things up for you — coding agent works faster and does higher quality work when set up for your repo.

[annahuynhly]

@copilot can you observe the issue above? I.e., the fact that fr.form and fr.form. are computed separately in glmer and lmer.

The following code is also repeated (which I did not mention earlier):

 for (i in c("weights", "offset")) {
        if (!eval(bquote(missing(x=.(i)))))
            assign(i,get(i,parent.frame()),environment(fr.form))
    }

[Copilot]

@copilot can you observe the issue above? I.e., the fact that fr.form and fr.form. are computed separately in glmer and lmer.

The following code is also repeated (which I did...

Done in 28e5651. I moved the shared fr.form./fr.form construction and the shared weights/offset environment assignment loop into mkFormula(), so lFormula() and glFormula() no longer duplicate that block.