I read long ago that verbs containing more than two syllabs and whose syllab before る containing the sound e or i is written in hiragana are all ichidan. Is it an actual rule or just a useful trick with possible exceptions ?
Is it true that a polysyllabic verb ending in iru or eru whose e/i syllab is written in hiragana ichidan?
In general, yes.
The underlying rule for kana conventions is that conjugable endings are written in kana. For example, for the verb kokorozasu ("to intend to do something"), only the -su on the end conjugates, so the kanji-fied spelling is 志す with that conjugable す hanging out as hiragana. For the verb hairu ("to enter"), only the -ru on the end conjugates, so the kanji-fied spelling is 入る with that conjugable る hanging out as hiragana.
However, for modern ichidan verbs, the -i- or -e- before the る doesn't change either. So why is that showing up in the kana after the kanji?
Historically, nearly all of the 一段活用動詞【いちだんかつようどうし】 (monograde-conjugation verbs) evolved from earlier 二段【にだん】 (bigrade) forms. So modern ichidan verb 変える "to change" evolved regularly from earlier classical 変う (well, strictly speaking, 変ふ -- the ending fu changed regularly to u, but the ふ still appears in old kana spellings, and monolingual Japanese dictionaries may still show it this way). The 変う form for the classical verb is the dictionary form (終止形【しゅうしけい】 or "terminal form"). The attributive form (連体形【れんたいけい】) ended in える, and the attributive of the bigrade verbs basically merged with the terminal to become the modern dictionary form.
Since that える ending is, historically, the conjugated ending, this is still spelled in kana. This also helps differentiate the modern ichidan verbs from regular godan verbs, more clearly distinguishing 変える ("to change", ichidan) from 帰る ("to return", godan).
So too with the other modern ichidan verbs. For instance, させる derives from classical さす, される from classical さる, られる from classical らる, 広【ひろ】める from classical 広【ひろ】む, and so on and so forth.
Another element that incentivizes fuller kana spellings is that some verbs have ichidan variants that only differ visibly by that -i- or -e- verb ending. One clear example is 見【み】る versus 見【み】える. If the spelling of the -e- portion were subsumed into the kanji (such as the kokoroza- portion at the start of this post, subsumed into the 志 kanji), then we would not be able to tell the difference between these two verbs, and reading would be (even more) complicated.
Put it all together, and the end result is that modern ichidan verbs have the -i- or -e- verb endings spelled out in kana, even though these parts no longer change in the modern language.
Caveat: As with many things in language, there may be exceptions to this. I can't think of any at the moment, but I'm also not a native speaker. Plus, manga are famous / notorious for creative spellings that may stray quite far from accepted norms, so when reading manga, take any odd spellings with a grain of salt. :)