mirror of
https://github.com/discourse/discourse-ai.git
synced 2025-03-09 11:48:47 +00:00
af4f871096
Previous to this change, on truncation we could return invalid UTF-8 strings to caller This also allows tools to read up to 30 megs vs the old 4 megs.
45 lines
1.1 KiB
Ruby
45 lines
1.1 KiB
Ruby
# frozen_string_literal: true
|
|
|
|
RSpec.describe DiscourseAi::AiBot::Tools::Tool do
|
|
let :tool_class do
|
|
described_class
|
|
end
|
|
|
|
let :corrupt_string do
|
|
"\xC3\x28\xA0\xA1\xE2\x28\xA1\xE2\x82\x28\xF0\x28\x8C\xBC"
|
|
end
|
|
|
|
describe "#read_response_body" do
|
|
class FakeResponse
|
|
def initialize(chunk)
|
|
@chunk = chunk
|
|
end
|
|
|
|
def read_body
|
|
yield @chunk while true
|
|
end
|
|
end
|
|
|
|
it "never returns a corrupt string" do
|
|
response = FakeResponse.new(corrupt_string)
|
|
result = tool_class.read_response_body(response, max_length: 100.bytes)
|
|
|
|
expect(result.encoding).to eq(Encoding::UTF_8)
|
|
expect(result.valid_encoding?).to eq(true)
|
|
|
|
# scrubbing removes 7 chars
|
|
expect(result.length).to eq(93)
|
|
end
|
|
|
|
it "returns correctly truncated strings" do
|
|
response = FakeResponse.new("abc")
|
|
result = tool_class.read_response_body(response, max_length: 10.bytes)
|
|
|
|
expect(result.encoding).to eq(Encoding::UTF_8)
|
|
expect(result.valid_encoding?).to eq(true)
|
|
|
|
expect(result).to eq("abcabcabca")
|
|
end
|
|
end
|
|
end
|