Argument error when trying to parse unicode in JSON about elixir-json HOT 11 CLOSED

Thanks for the bug report. I'll look into this ASAP.

Sent from mobile

On Sep 30, 2014, at 7:07 PM, Nathan Zorn [email protected] wrote:

The parser seems to have issues with some unicode strings.

iex(8)> JSON.decode("{"name": "Fahad Al-Esmail \u00f8"}")
{:ok, %{"name" => "Fahad Al-Esmail ø"}}
iex(9)> JSON.decode("{"name": "Fahad Al-Esmail \ud83c"}")
** (ArgumentError) argument error
lib/json/parser/bitstring/string.ex:85: JSON.Parser.Bitstring.String.parse_escaped_unicode_codepoint/3
lib/json/parser/bitstring/string.ex:65: JSON.Parser.Bitstring.String.parse_string_contents/2
lib/json/parser/bitstring/object.ex:50: JSON.Parser.Bitstring.Object.parse_object_value/3
lib/json/decoder.ex:30: JSON.Decoder.BitString.decode/1

—
Reply to this email directly or view it on GitHub.

from elixir-json.

So, turns out my issue was an elixir/erlang upgrade. I didn't rebuild and got this error. After rebuilding everything it works now.

from elixir-json.

Cool, thanks for the update! ❤️

from elixir-json.

Actually... it turns out it had nothing to do with the elixir/erlang upgrade. (I work with @thepug ). If you could re-open this issue and look into it @cblage , we'd be much obliged!

from elixir-json.

Heya! Sure, I'll try to investigate this ASAP. Probably won't be able to do much until late Saturday tho

from elixir-json.

Here's a definitive failure from the console where Jiffy handles it fine:

iex(30)> s = "{\"foo\":\"I love emoji! \\ud83c\\uddfa\\ud83c\\uddf8 So do you!\"}"                                                                                                 "{\"foo\":\"I love emoji! \\ud83c\\uddfa\\ud83c\\uddf8 So do you!\"}"
iex(31)> :jiffy.decode(s)
{[{"foo", "I love emoji! 🇺🇸 So do you!"}]}
iex(32)> JSON.decode(s)                                                                                                                                                            ** (ArgumentError) argument error
    lib/json/parser/bitstring/string.ex:85: JSON.Parser.Bitstring.String.parse_escaped_unicode_codepoint/3
    lib/json/parser/bitstring/string.ex:65: JSON.Parser.Bitstring.String.parse_string_contents/2
    lib/json/parser/bitstring/object.ex:50: JSON.Parser.Bitstring.Object.parse_object_value/3
    lib/json/decoder.ex:30: JSON.Decoder.BitString.decode/1

from elixir-json.

In particular, it is 4-byte Emoji Unicode that seems to be the problem. Those 4-bytes of Unicode are a US flag emoji for example:

http://www.iemoji.com/view/emoji/176/places/regional-indicator-symbol-letters-us

from elixir-json.

Thanks for the debugging @belucid, this will help a lot :)

from elixir-json.

WIP for fix: #30

from elixir-json.

@belucid I've been completely swamped recently, if anyone wants to contribute with a fix, please feel free. I'm doing my best to get this fixed soonish 😶

from elixir-json.

So far as I've seen nobody has worked on this, am I right. I need this fixed for my project; I think I will be working on this today.

from elixir-json.