I've learned that ru verbs should drop る and replace it with た.
But 入る becomes 入った and I'm not sure why.
The same goes for
- 食べる → 食べた
- しゃべる → しゃべた
Are these irregulars or am I misunderstanding something?
入る hairu is a consonant-stem verb, i.e. hair·u.
Besides a handful of exceptions, there are two type of verbs
where I have used an interpunct · to separate the stem from the ending (and the hyphen - means there is more coming before it).
These types follow different inflection patterns.
Of course there are no pure consonants in kana, so written in kana you cannot show where the border between stem and ending is — for example a verb ending in ～く may inflect to ～か (as in 聞く → 聞かない), but written as above, the stem ending in -k does not change (kik·u → kik·anai). (This is why it is called the stem. And the stem of vowel-stem/consonant-stem verbs end in a vowel/consonant, respectively.)
Now, seeing a verb written in kana, there is no way of telling whether a verb ending in -eru / -iru is a vowel-stem verb or a consonant stem verb, i.e. whether it is -e·ru / -i·ru or -er·u / -ir·u.
The following vowel-stem–consonant-stem verb pairs show that this difference simply has to be learned:
In particular, their past tenses are different
Now, 入る is hair·u (i.e. a consonant-stem verb), so its past tense follows the same pattern as 切る, giving 入った.
However, the way the verb is written in kanji gives some way of guessing whether it is vowel-stem or consonant-stem. As a rule of thumb (for verbs of three or more morae), if the -i or -e is part of the kanji, then a verb is more likely to be consonant-stem.
In particular, since はいる is a three mora verb, written 入る rather than *入いる, it is more likely to be consonant-stem. The rule also correctly identifies 食べる as a vowel-stem verb (→ 食べた) and 喋る【しゃべる】 as a consonant-stem verb (→ 喋った).
The name "ru verb" is misleading.
This does not mean that every verb ending in -ru conjugates the same way.
Certain verbs end in -ru, and that final -ru disappears in entirety when conjugating. These are called 一段活用 (ichidan katsuyō, literally "monograde conjugation" or "one-step conjugation") verbs in Japanese grammars, because there is only one vowel on the end of the verb stem (either -i or -e). These are often labeled as "type 2" verbs in English texts about Japanese grammar.
One everyday example is 食べる (taberu, "to eat"), where that final -ru just disappears entirely in conjugated forms:
... etc. As you can see here, the -ru vanishes entirely in the conjugated forms.
Other verbs end in -ru or -ku or -su, etc., and that final piece doesn't quite vanish entirely -- the consonant remains, but the vowel changes. These are called 五段活用 (godan katsuyō, literally "quintigrade conjugation" or "five-step conjugation") verbs in Japanese grammars, because there are five different possible vowels on the end of the verb stem (-a, -i, -u, -e, -o). These are often labeled as "type 1" verbs in English texts about Japanese grammar.
One everyday example with -ru is 入る (hairu, "to enter, to go into"), where that final -ru doesn't disappear entirely, and only the vowel changes:
As you can see here, the -r- sticks around, and the vowel changes.
Things ultimately get a bit more complicated than this (such as the conjunctive forms and plain past forms, where you often get geminate [i.e. doubled] consonants instead of the final consonant, as in haitta), but that's the basic framework. There's good information on Wikipedia, at https://en.wikipedia.org/wiki/Japanese_verb_conjugation.
The rule for '-ru' (ichidan) verbs and '-u' (godan) verbs is 'anything ending in '-iru' or '-eru' can be a '-ru' verb. But 入る and しゃべる are not; they are '-u' verbs that conjugate as you would expect.