"Impersonal" is not the same as "passive", Romans said "itur" but not *"eor", *"itimur", and such, only third person in the impersonal sense.
I did not imply that impersonal is the same as passive, and I would appreciate it if you did not respond to me in reference to things I did not say or imply. The passive makes the object into the subject; the impersonal passive
deletes the subject altogether. If you pay close attention to the discussion, you'll notice that the discussion centers around the past participle, and the contention so far had been that, since the verb is intransitive, it cannot have
passive forms. The fact is that it can, the ones used in impersonal constructions, which I listed. The only implication here is that those forms that aren't used in impersonal constructions don't exist. As to past participles, these forms should in principle exist for all the verbs (except when they don't).
The future passive infinitive "caritum ir
i" is archaic, in classical Latin the preferred form is the perifrastic passive, as attested by Cicero and Ovidius:
cum enim mihi carendum sit conviviis, "since I must be deprived of social gatherings" (nihil novum sub sole, I can add)
“
virque mihi dempto fine carendus abest,”
Ov. H. 1, 50. "I must be deprived of a husband..."
Firstly, I did not mean the future passive infinitive. My goal was to list a construction using the supine, and the verb
īre fit the bill. The construction is active (
see § 659): *
vidēmur multīs rēbus caritum īre = vidēmur caritūrum (later
caritūrī esse) "we seem to be in for a lack of many things". Note that again, I do not claim that this expression with
carēre is attested or at all idiomatic. However, since
carēre is intransitive, a semantically
passive constructrion with
īrī is impossible here (but again, there are transitive attestations). I don't think I've ever seen an impersonal
īrī-construction -
this is the closest I've found, IMO still personal "Varro will approve of it".
Secondly, these are different periphrastic constructions. The
īrī-forms is what the actual grammarians give as future passive infinitives, but they appear with any frequency only in Cicero and Terence. My personal opinion is that they belonged to the older, less literary stratum of the language. The
īrī-forms have no deontic meaning, but express simple futurity; the gerund(ive)s have the deontic meaning inherent to them:
portandum esse "that it must, it has to be carried", or as an adjective "that is to be carried".
What the future "passive" (not in the grammatical sense) meaning was normally expressed with are: 1) the present inf.
spērō saccum portārī; 2) the construction
fore ut/futūrum ut saccus portētur. It's only in later Latin (from Tertullian onwards) that
portandum esse develops into a bona-fide alternative for the future infinitive. There were several further alternative constructions including
spērō portārī posse and later the future-in-the-past
saccus portārī habēbat (also from Tertullian onwards, I don't think it appears for the actual future), as well as
spērō quod/quia portētur.
References: mainly Pinkster H. (2015). the Oxford Latin Syntax Vol. 1.