Bug #13952: String#succ not updating code range

• Status: Closed
• Assignee: -
• Target version: -
• ruby -v: ruby 2.4.2p198 (2017-09-14 revision 59899) [x86_64-linux]
• Backport: 2.3: DONE, 2.4: DONE

[nirvdrum (Kevin Menard)]

I'm seeing some strange behavior with String#succ and updating code ranges. I haven't yet traced the code to see what the culprit is, but I'm reproducing my findings here so they don't get lost (and maybe someone has a better idea of what's going on.)

This sequence of calls produces the expected output.

x = "\xFF".force_encoding("binary")
y = x.succ
z = String.new
z << 0x01 << 0x00

puts "x ASCII-only?: #{x.ascii_only?}"
puts "y ASCII-only?: #{y.ascii_only?}"
puts "z ASCII-only?: #{z.ascii_only?}"
puts "y Encoding: #{y.encoding}"
puts "y Bytes: #{y.bytes}"
puts "z Encoding: #{z.encoding}"
puts "z Bytes: #{z.bytes}"

The output is:

x ASCII-only?: false
y ASCII-only?: true
z ASCII-only?: true
y Encoding: ASCII-8BIT
y Bytes: [1, 0]
z Encoding: ASCII-8BIT
z Bytes: [1, 0]

However, by inserting a call that would force x to calculate its code range prior to the String#succ call, we get a different set of results:

x = "\xFF".force_encoding("binary")
x.ascii_only?
y = x.succ
z = String.new
z << 0x01 << 0x00

puts "x ASCII-only?: #{x.ascii_only?}"
puts "y ASCII-only?: #{y.ascii_only?}"
puts "z ASCII-only?: #{z.ascii_only?}"
puts "y Encoding: #{y.encoding}"
puts "y Bytes: #{y.bytes}"
puts "z Encoding: #{z.encoding}"
puts "z Bytes: #{z.bytes}"

Now we see that y isn't considered to be ASCII-only, even though it has the exact same encoding and byte sequence as z (and as y in the previous call sequence that did work):

x ASCII-only?: false
y ASCII-only?: false
z ASCII-only?: true
y Encoding: ASCII-8BIT
y Bytes: [1, 0]
z Encoding: ASCII-8BIT
z Bytes: [1, 0]

Having not looked at it, it looks like the code range isn't updated and we only get the correct result if CR_UNKNOWN hasn't been replaced by some other call that needs the code range.

[nobu (Nobuyoshi Nakada)]

• Status changed from Open to Closed

Applied in changeset trunk|r60066.

---

string.c: fix ASCII-only on succ

• string.c (str_succ): clear coderange cache when no alpha-numeric
  character case, carried part may become ASCII-only.
  [ruby-core:83062] [Bug #13952]

[nagachika (Tomoyuki Chikanaga)]

• Backport changed from 2.3: UNKNOWN, 2.4: UNKNOWN to 2.3: REQUIRED, 2.4: REQUIRED

[nagachika (Tomoyuki Chikanaga)]

• Backport changed from 2.3: REQUIRED, 2.4: REQUIRED to 2.3: REQUIRED, 2.4: DONE

ruby_2_4 r61455 merged revision(s) 60066.

[usa (Usaku NAKAMURA)]

• Backport changed from 2.3: REQUIRED, 2.4: DONE to 2.3: DONE, 2.4: DONE

ruby_2_3 r62139 merged revision(s) 60066.