scroll

Scroll is a tailing pager.

• Unlike tail, it has interactive features like paging and incremental regex search.

• Unlike less, it continuously tries to keep reading more bytes from the input file.

• Unlike less and tail, it loads the whole file into memory before starting, which makes it perform worse on files which are already large.

On my machine, I start to notice perceptible latency:

• Changing search query in a file larger than a few dozen MB
• Opening a file larger than a few hundred MB

I never notice latency in scrolling or tailing, even in files up to 15 GB. I didn't test any files larger than that because they take too long to open.