Currently std.file.read (http://dlang.org/phobos/std_file.html#.read) may be called with a file name, or a file name and a number indicating maximum bytes read.
A new overload should allow calls like read("file", 10, 100), i.e. read 100 bytes from "file" starting at offset 10.
The Unix primitive http://linux.die.net/man/2/pread does positioned read faster than seek + read.
Comment #1 by robert.schadek — 2024-12-01T16:24:48Z