Octo review start fails with large pull requests

[cj]

Issue Description

bug report

Describe what happened (or what feature you want)

When opening large pull requests :Octo review start fails with:

Error executing vim.schedule lua callback: Vim:E474: Trailing characters: [{"additions":11,"blob_url":"https://github.com/******/*****/blob/81b56f1be49c867d3274c60ce6a93626524cf496/packages%2Fui%2Ficons%2FDribble.tsx","changes":11,"contents_url":"https://api.github.com/repos/s
napnhd/snap-system/contents/packages%2Fui%2Ficons%2FDribble.tsx?ref=81b56f1be49c867d3274c60ce6a93626524cf496","deletions":0,"filename":"packages/ui/icons/Dribble.tsx","patch":"@@ -0,0 +1,11 @@\n+export function DribbleIcon() {\n+  return (\n+    \u003csvg className=\"h-6 w-6\" fill=\
"currentColor\" viewBox=\"0 0 24 24\"\u003e\n+      \u003cpath\n+        fillRule=\"evenodd\"\n+        d=\"M12 2C6.48 2 2 6.48 2 12s4.48 10 10 10c5.51 0 10-4.48 10-10S17.51 2 12 2zm6.605 4.61a8.502 8.502 0 011.93 5.314c-.281-.054-3.101-.629-5.943-.271-.065-.141-.12-.293-.184-.445a25
.416 25.416 0 00-.564-1.236c3.145-1.28 4.577-3.124 4.761-3.362zM12 3.475c2.17 0 4.154.813 5.662 2.148-.152.216-1.443 1.941-4.48 3.08-1.399-2.57-2.95-4.675-3.189-5A8.687 8.687 0 0112 3.475zm-3.633.803a53.896 53.896 0 013
stack traceback:
        [C]: in function 'json_decode'
        ...ack/packer/opt/octo.nvim/lua/octo/model/pull-request.lua:86: in function 'cb'
        ...nvim/site/pack/packer/opt/octo.nvim/lua/octo/gh/init.lua:92: in function ''
        vim/_editor.lua: in function <vim/_editor.lua:0>

Describe what you expected to happen

It to start the pr review.

How to reproduce it (as minimally and precisely as possible)

1. Create pr with over 100 file changes
2. Run :Octo review start

Tell us your environment

NVIM v0.7.0 Build type: Release LuaJIT 2.1.0-beta3 Compiled by linuxbrew@30f25fa6036f

Features: +acl +iconv +tui See ":help feature-compile"

system vimrc file: "$VIM/sysinit.vim" fall-back for $VIM: " /home/linuxbrew/.linuxbrew/Cellar/neovim/0.7.0/share/nvim"

Anything else we need to know?

gh.nvim had a similar issue https://github.com/ldelossa/gh.nvim/issues/8

[pwntester]

Thanks for reporting this, I can think of some workarounds piping the gh output into jq but I think its better to wait until gh implements array merging natively. This is actively being implemented so it should probably be released soon.

ref: https://github.com/cli/cli/pull/5652

[jloesch30]

Do we have a temporary workaround for this at the moment? @pwntester I think you hinted at one and it looks like the cli PR you mentioned is gaining activity again

[JohnCosta27]

Also experiencing this issue on large PRs

[robsonpeixoto]

I'm my tests the json_decode does not accept big strings. Maybe will be necessary implement a paginator instead of use the gh pagination feature