What actually happens when you print()

lesson 8-2 · ~11 min · 26/28

Following one print() all the way down

You now know every layer. Here is print("hi"), bottom to top no more mysteries:

Python: print formats its arguments and calls sys.stdout.write("hi\n").
Buffer (lesson 7-2): the text lands in a user-space buffer. To a terminal, stdout is line-buffered, so the \n triggers a flush.
System call (lesson 8-1): the flush issues write(1, b"hi\n", 3). That 1 is a file descriptor, a small number the kernel gives your process for each open thing. 0 is stdin, 1 is stdout, 2 is stderr.
Kernel: the CPU switches to kernel mode, the kernel routes the bytes to whatever descriptor 1 is connected to, a terminal, a file, or a pipe to another program.
Terminal: another user-space process receives the bytes, decodes them as UTF-8 (lesson 2-3), and draws glyphs.

Your program never touched the screen. It handed 3 bytes across the boundary.

Code exercise · python

Run this. os.write is a thin wrapper around the raw write system call, no Python buffering at all. File descriptor 1 is stdout, and it needs raw bytes, not a string.

Code exercise · python

Run this. print is a convenience layer over sys.stdout.write. The write method also returns how many characters it wrote, which mirrors what the underlying syscall reports.

Code exercise · python

Your turn. Without using print or sys.stdout, write the text "hello from fd 1" plus a newline to standard output using os.write and file descriptor 1. Remember it must be bytes.

Quiz

You run python3 app.py > out.txt in the shell. app.py calls print("hi"). Where do the bytes go, and did app.py need any code changes?

Quiz

print() to a terminal flushes on every newline, but print() redirected to a file flushes only when a large buffer fills. Connecting this to lesson 7-2, why is that a sensible default?