Bug #20416: IO#read doesn't change buffer encoding unless `maxlen = nil`

• Status: Closed
• Assignee: -
• Target version: -
• ruby -v: 3.3.0
• Backport: 3.0: UNKNOWN, 3.1: UNKNOWN, 3.2: UNKNOWN, 3.3: UNKNOWN

[andrykonchin (Andrew Konchin)]

IO#read(nil, buffer) doesn't change buffer's encoding, but it does when maxlen parameter is specified:

buffer = "".encode(Encoding::ISO_8859_1)
p File.open("test.rb").read(1000, buffer).encoding # => #<Encoding:ISO-8859-1>

buffer = "".encode(Encoding::ISO_8859_1)
p File.open("test.rb").read(nil, buffer).encoding  # => #<Encoding:UTF-8>

I would expect changing buffer encoding in all the cases.

[byroot (Jean Boussier)]

Related: https://github.com/ruby/spec/pull/1145

[Eregon (Benoit Daloze)]

• Description updated (diff)

[byroot (Jean Boussier)]

I git blamed this behavior all the way down to 2007, by @matz (Yukihiro Matsumoto): https://github.com/ruby/ruby/commit/0ca7036682dad1caea4e7a4fe8fd908d94c878e8 and it seems deliberate

* io.c (read_all): should associate default external encoding.
* io.c (io_read): should NOT associate default external encoding.

[nobu (Nobuyoshi Nakada)]

The subject should be "IO#read doesn't change buffer encoding unless maxlen = nil"?

Since IO#read with maxlen is byte-oriented operation, there is not an encoding to be set from that IO.

[byroot (Jean Boussier)]

Yeah, I don't think it's a bug. @nobu (Nobuyoshi Nakada) should we close?

[andrykonchin (Andrew Konchin)]

• Subject changed from IO#read doesn't preserve buffer encoding if `maxlen = nil` to IO#read doesn't change buffer encoding unless `maxlen = nil`

[mame (Yusuke Endoh)]

• Status changed from Open to Closed

Discussed at the dev meeting, and @matz (Yukihiro Matsumoto) agreed with @nobu (Nobuyoshi Nakada). Closing