gh-138775: fix handle python -m base64 stdin correct with EOF signal

[yihong0618]

This make python -m base64 the same behavior as ast or json

• Issue: python -m base64 do not handle stdin correct like ast.py or json.tool #138775

[aisk]

I’m concerned that the current implementation reads the input in chunks, this allows the function to process the input as a stream via a pipe and avoids using too much memory with large inputs:

cpython/Lib/base64.py

Lines 531 to 537 in 4978bfc

	def encode(input, output):
	"""Encode a file; input and output are binary files."""
	while s := input.read(MAXBINSIZE):
	while len(s) < MAXBINSIZE and (ns := input.read(MAXBINSIZE-len(s))):
	s += ns
	line = binascii.b2a_base64(s)
	output.write(line)

With this change, however, all input is read into memory, which breaks the chunked behavior.

[yihong0618]

I’m concerned that the current implementation reads the input in chunks, this allows the function to process the input as a stream via a pipe and avoids using too much memory with large inputs. :

cpython/Lib/base64.py

Lines 531 to 537 in 4978bfc

	def encode(input, output):
	"""Encode a file; input and output are binary files."""
	while s := input.read(MAXBINSIZE):
	while len(s) < MAXBINSIZE and (ns := input.read(MAXBINSIZE-len(s))):
	s += ns
	line = binascii.b2a_base64(s)
	output.write(line)

With this change, however, all input is read into memory, which breaks the chunked behavior.

Thanks for the comments

I will checked if isatty to check
if
echo 1234 | ./python.exe -m base64
or
./python.exe -m base64
and do not break the old method

and in stdin old behavior is wrong, so I think is fine.
will check isatty if can check or not.

[yihong0618]

keep the old behavior and use isatty to check
as doc https://docs.python.org/3/library/sys.html#sys.stdin

[yihong0618]

Thank you very much for the review again, I will try to learn how to express the change more clearly.
and do better next time, thanks again

[picnixz]

I will try to learn how to express the change more clearly.
and do better next time, thanks again

Don't worry. Considering English isn't your first language, you can just leave formulation matters to others.

[gpshead]

I went with this PR after playing around with the BufferedReader connected to a terminal - it really is difficult to tell when you have an EOF in that scenario vs what the base64.encode() code wants to do when reading fixed sized blocks at a time in order to produce the fixed sized output lines. subsequent .read() calls after a short read or even after a 0 length read (the normal EOF indicator) still block and could later return more data (ugh, terminals!). So doing the isatty check does seem most practical here given that the io stack.

[miss-islington-app]

Thanks @yihong0618 for the PR, and @gpshead for merging it 🌮🎉.. I'm working now to backport this PR to: 3.13, 3.14.
🐍🍒⛏🤖

[bedevere-app]

GH-141432 is a backport of this pull request to the 3.14 branch.

[bedevere-app]

GH-141433 is a backport of this pull request to the 3.13 branch.