Bug #22342: IO#gets can ignore the limit argument in some edge cases

• Status: Closed
• Assignee: -
• Target version: -
• Backport: 3.3: UNKNOWN, 3.4: DONE, 4.0: REQUIRED

[rhenium (Kazuki Yamaguchi)]

io.gets(rs, limit) ignores the limit value and reads until rs is found or EOF, if the limit-th byte of the stream matches rs's last byte and:

1. The potential start of the record separator, the limit-rs.bytesize-th byte of the stream is not at a character boundary

str = "\x83\\xxx".force_encoding("Windows-31J") # "ソxxx".encode("Windows-31J")
IO.pipe("Windows-31J"){|r,w|w.write(str); w.close_write; p r.gets("\\", 2); p r.read}

# Expected:
"\x{835C}"
"xxx"

# Actual:
"\x{835C}xxx"
""

2. Or, limit < rs.bytesize

IO.pipe{|r,w|w.write("a:bc"); w.close_write; p r.gets(":::", 2); p r.read}

# Expected:
"a:"
"bc"

# Actual:
"a:bc"
""

This behavior appears to date back to Ruby 1.9.x.

[rhenium (Kazuki Yamaguchi)]

I've opened a GitHub PR for this: https://github.com/ruby/ruby/pull/18961

[rhenium (Kazuki Yamaguchi)]

• Status changed from Open to Closed

Applied in changeset git|227006291d35b2255b3f7a888d75102c129b0472.

---

Fix IO#gets ignoring the limit argument

IO#gets first searches for separator candidates by looking for the last
byte of the record separator, and then checks whether it is an actual
match.

If a candidate is found at limit-th byte of the stream but is rejected
because it does not start at a character boundary, the limit check is
incorrectly skipped.

[Bug #22342]

[nagachika (Tomoyuki Chikanaga)]

• Backport changed from 3.3: UNKNOWN, 3.4: REQUIRED, 4.0: REQUIRED to 3.3: UNKNOWN, 3.4: DONE, 4.0: REQUIRED

ruby_3_4 0a33ab92c08c8d1dbba0b3c97781d16b40ac53a8 merged revision(s) 227006291d35b2255b3f7a888d75102c129b0472.