Hoc is always scanned long in classical poetry, because it is the same as *hocc (from *hocce < *hodce). I give only a couple of examples, but you can check by yourself using http://www.pedecerto.eu/ricerca/forma. It is the same for the other words formed with the intensifying particle -ce, like istuc (from *istucce < *istudce) or illuc (from *illucce < *illudce)
Lucil. sat. 1032: hoc etiam accipe quod dico: nam pertinet ad
Verg. ecl. 1,539: Quod genus hoc hominum? quaeue hunc tam barbara
Concerning other words which present the same gemination phenomenon, there are:
words from -rr, like cor (from *cord), which is scanned long by archaic authors only (e.g. Plaut. Pers. 802 [quaternary cataleptic anapest] cor uritur, caput ne ardescat, but already in Lucil. sat. 516 is scanned short: vera putant, credunt signis cor inesse in aenis), or ter (from *terr, cfr. the compound terruncius);
words from -ss (from the assimilation of the dental conosonant of the stem and the -s of the first person nominative singular desinence), like miles, compos, dives, eques, hospes, impos et similia, or in es (second person singular of sum, from *ess), cfr. e.g. Plaut. Aul. 528 (iambic senarius): miles impransus astat, aes censet dari; Plaut. Cas. 817 (iambic senarius): sospes iter incipe hoc, uti viro tuo; but already in Ennius these syllables are scanned short (e.g. ann. 269: spernitur orator bonus, horridus miles amatur).
Vd. M. Leumann, Lateinische Laut- und Formenlehre, München 1977, pp. 220-221; S. Boldrini, La prosodia e la metrica dei Romani, Roma 1992, pp. 50-51.