Language is a funny thing, and the Hakka speaking peoples have had a history of migrations (as have the Cantonese, though fewer people speak of such things - their clan records tell this). It said many of the features of Hakka are due to having passed through those areas on their tracks down southwards over the centuries.
One of these features is said to be where the front high vowels in syllables ending in -ng have come down to us as -n finals, and also many of such characters have colloquial and literary readings.
For instance "root" can be pronounce /kin/, /kEn/ and /kaN/ where /N/ represents the velar nasal ng sound.
Some people pronounce the festival QingMing as /ts'in min/, my wife pronounces it as /tsiaN miaN/.
It is said this change is reflected in Gan speaking areas too, and some hypothesise that Hakka and Gan share a common ancestral language form.
Not only sound changes, but Hakka have words which are also shared with dialects. Spider /la k'ia/ appears in some Min dialects, for instance. Hakka speakers are said to have passed through western Fujian before arriving in current day Meixian. Western Fujian is dotted with Hakka speaking areas like RongDing and it's tulou buildings.
Fujian Hakka dialects differs from Meixian. ChangTing dialect for instance lacks the Ngip Sang tone. Despite this, they are Hakka dialects, as they share other factors like grammar and vocabulary.
In a paper given to me by a researcher into Hakka and other languages, he postulates that Hakka, Gan and Yue dialects like those of the Siyi (sei yap - si yip) may be related. As I've already mentioned, the Yue speakers also migrated, and they may have arrived in the southern Guangdong areas much earlier than the Hakka, but I don't believe some of the Cantonese is as old as 1800 years old, but it might be much later I feel. Anyway, it's complicated, and I'm not a linguist.