Can we use hashing also when comparing pihole.toml and pihole.toml.temp (and the other files). This might be faster than comparing line by line.

We'd anyway have to read line-by-line to be able to skip the first N lines to skip the timestamp in the header (or all files will always be different!). If reading in bulk, we have to iterate until we have found enough \n and then offset the SHA computation. Either or, comparing line-by-line is much faster because iterating in C comes at no costs (unless we are talking about billions of lines) and a direct memory comparison is an instruction the CPU can do at once, the entire SHA computation is, on the other hand, a pretty massive operation:

(source)

The core of the SHA algorithm is called the compression function and consists of 4 cycles of 20 steps each (=80 steps). The loops are very similar in structure except that they use a different primitive logic function.

Each block is taken as an input parameter by all 4 loops along with a constant K and the values of the eight registers. At the end of the computation new values will be obtained for A, B, C, D, E, F, G, and H which will be used for the computation of the next block until the final block H is obtained.

edit FIPS PUB 180-4 has all the details, look at Scn. 6.2 in particular.

but manually editing pihole.toml does not log the writing and does not backup the log

I think this is because the recent bugfix #1775 was not included here. I just merged development-v6