😲 Emoji characters in keys & values in v4 are lost/corrupted

[elasticdotventures]

$ yq -V
yq version 4.7.0

The way yq v4 handles emoji is odd, inconsistent, unpredictable (which did not occur on earlier yq 2x versions which had other limitations)

yq should (imho) pass utf8/emoji through unmolested. yq works properly with pinyin (chinese mandarin) fonts but ideograms are so much more powerful and universal it'd be nice to use them with.

For example let's say emojifile.yaml with contents:

---
"bash.🔨/init.10级.🥾.b00t.sh": ""
"bash.🔨/init.20级.🐧.linux.sh": ""
"bash.🔨/init.22级.🐙.git.sh": ""
"bash.🔨/init.30级.🐳.层.docker.sh": ""
"bash.🔨/init.32级.💠.层.hashicorp.sh": ""
"bash.🔨/init.40级.🐍.语.python.sh": ""
"bash.🔨/init.40级.🚀.语.node.sh": ""
"bash.🔨/init.42级.🦄.语.typescript.sh": ""
"bash.🔨/init.43级.🥷.语.vue.sh": ""
"bash.🔨/init.44级.☕.语.java.sh": ""
"bash.🔨/init.44级.🏇.语.go.sh": ""
"bash.🔨/init.50级.👾.云☁️.gcp.sh": ""
"bash.🔨/init.50级.🤖.云☁️.azure.sh": ""
"bash.🔨/init.50级.🦉.云☁️.aws.sh": ""
"bash.🔨/init.60级.🎙️💙.应用.vscode.sh": ""
"bash.🔨/init.70级.☎️.msg.sh": ""
"bash.🔨/init.70级.🎬.video.sh": ""
"bash.🔨/init.70级.📱.mobile.sh": ""
"bash.🔨/init.70级.🕹️.gamesim.sh": ""
"bash.🔨/init.70级.🤑.ecommerce.sh": ""
"bash.🔨/init.70级.🥯.crypto.sh": ""
"bash.🔨/init.70级.🧠.ai.sh": ""
"bash.🔨/init.80级.🐱‍💻.esp32.sh": ""

then

$ cat emojifile.yaml | yq eval

will produce (on my ubuntu system)

"bash.\/init.10级.\.b00t.sh": ""
"bash.\/init.20级.\.linux.sh": ""
"bash.\/init.22级.\.git.sh": ""
"bash.\/init.30级.\.层.docker.sh": ""
"bash.\/init.32级.\.层.hashicorp.sh": ""
"bash.\/init.40级.\.语.python.sh": ""
"bash.\/init.40级.\.语.node.sh": ""
"bash.\/init.42级.\.语.typescript.sh": ""
"bash.\/init.43级.\.语.vue.sh": ""
"bash.\/init.44级.☕.语.java.sh": ""
"bash.\/init.44级.\.语.go.sh": ""
"bash.\/init.50级.\.云☁️.gcp.sh": ""
"bash.\/init.50级.\.云☁️.azure.sh": ""
"bash.\/init.50级.\.云☁️.aws.sh": ""
"bash.\/init.60级.\️\.应用.vscode.sh": ""
"bash.\/init.70级.☎️.msg.sh": ""
"bash.\/init.70级.\.video.sh": ""
"bash.\/init.70级.\.mobile.sh": ""
"bash.\/init.70级.\️.gamesim.sh": ""
"bash.\/init.70级.\.ecommerce.sh": ""
"bash.\/init.70级.\.crypto.sh": ""
"bash.\/init.70级.\.ai.sh": ""
"bash.\/init.80级.\‍\.esp32.sh": ""

This is for b00t framework.

[elasticdotventures]

cat emojifile.yaml | yq eval -M

"bash.\U0001F528/init.10级.\U0001F97E.b00t.sh": ""
"bash.\U0001F528/init.20级.\U0001F427.linux.sh": ""
"bash.\U0001F528/init.22级.\U0001F419.git.sh": ""
"bash.\U0001F528/init.30级.\U0001F433.层.docker.sh": ""
"bash.\U0001F528/init.32级.\U0001F4A0.层.hashicorp.sh": ""
"bash.\U0001F528/init.40级.\U0001F40D.语.python.sh": ""
"bash.\U0001F528/init.40级.\U0001F680.语.node.sh": ""
"bash.\U0001F528/init.42级.\U0001F984.语.typescript.sh": ""
"bash.\U0001F528/init.43级.\U0001F977.语.vue.sh": ""
"bash.\U0001F528/init.44级.☕.语.java.sh": ""
"bash.\U0001F528/init.44级.\U0001F3C7.语.go.sh": ""
"bash.\U0001F528/init.50级.\U0001F47E.云☁️.gcp.sh": ""
"bash.\U0001F528/init.50级.\U0001F916.云☁️.azure.sh": ""
"bash.\U0001F528/init.50级.\U0001F989.云☁️.aws.sh": ""
"bash.\U0001F528/init.60级.\U0001F399️\U0001F499.应用.vscode.sh": ""
"bash.\U0001F528/init.70级.☎️.msg.sh": ""
"bash.\U0001F528/init.70级.\U0001F3AC.video.sh": ""
"bash.\U0001F528/init.70级.\U0001F4F1.mobile.sh": ""
"bash.\U0001F528/init.70级.\U0001F579️.gamesim.sh": ""
"bash.\U0001F528/init.70级.\U0001F911.ecommerce.sh": ""
"bash.\U0001F528/init.70级.\U0001F96F.crypto.sh": ""
"bash.\U0001F528/init.70级.\U0001F9E0.ai.sh": ""
"bash.\U0001F528/init.80级.\U0001F431‍\U0001F4BB.esp32.sh": ""

[elasticdotventures]

BUT -j (json) apparently works

$ cat emojifile.yaml | yq eval -j

{
  "bash.🔨/init.10级.🥾.b00t.sh": "",
  "bash.🔨/init.20级.🐧.linux.sh": "",
  "bash.🔨/init.22级.🐙.git.sh": "",
  "bash.🔨/init.30级.🐳.层.docker.sh": "",
  "bash.🔨/init.32级.💠.层.hashicorp.sh": "",
  "bash.🔨/init.40级.🐍.语.python.sh": "",
  "bash.🔨/init.40级.🚀.语.node.sh": "",
  "bash.🔨/init.42级.🦄.语.typescript.sh": "",
  "bash.🔨/init.43级.🥷.语.vue.sh": "",
  "bash.🔨/init.44级.☕.语.java.sh": "",
  "bash.🔨/init.44级.🏇.语.go.sh": "",
  "bash.🔨/init.50级.👾.云☁️.gcp.sh": "",
  "bash.🔨/init.50级.🤖.云☁️.azure.sh": "",
  "bash.🔨/init.50级.🦉.云☁️.aws.sh": "",
  "bash.🔨/init.60级.🎙️💙.应用.vscode.sh": "",
  "bash.🔨/init.70级.☎️.msg.sh": "",
  "bash.🔨/init.70级.🎬.video.sh": "",
  "bash.🔨/init.70级.📱.mobile.sh": "",
  "bash.🔨/init.70级.🕹️.gamesim.sh": "",
  "bash.🔨/init.70级.🤑.ecommerce.sh": "",
  "bash.🔨/init.70级.🥯.crypto.sh": "",
  "bash.🔨/init.70级.🧠.ai.sh": "",
  "bash.🔨/init.80级.🐱‍💻.esp32.sh": ""
}

[elasticdotventures]

Just confirmed same behavior on yq 4.8.0

[elasticdotventures]

Just confirmed that the "other" yq project works properly with Emoji.
https://github.com/kislyuk/yq

When I said "earlier" versions worked, that was incorrect.
I didn't realize I'd switched repos.

[mikefarah]

Digging a little into this - and as far as I can tell it's an issue with go-yaml, the underlying yaml parser :(

go-yaml/yaml#279

Not sure if I'll be able to work around it

[mikefarah]

Raised a new issue here: go-yaml/yaml#737

[mikefarah]

Note that '-j' works because the issue is with the yaml Encoder and the json encoder works fine.

[zhangguanzhang]

if use shell, could used this command

tr -cd '\11\12\15\40-\176' < 1.yml  > new.yml

[surrim]

$ cat emojifile.yaml | yq eval

will produce (on my ubuntu system)

...

I found a small workaround. You can wrap it with echo -e, then it will work:

echo -e "$(cat emojifile.yaml | yq eval)"