veksha / cuda_codeium Goto Github PK

this bug is similar to bug in cuda_hotspots: veksha/cuda_hotspots#22
because we use ed.cmd(cmds.cmd_FileNew) and we assume that we now can use ed var that will point to new tab editor, but this is not always the case on macOS.

fixed. (workaround)
@Alexey-T, can you update cuda_codeium in addon manager, please?

to [info]:
os=win,linux,macos

Content:

2023.05.14
+ add: command "Chat..."

2023.05.10
initial version

The Chat works, but "Get completions" doesn't get any item.
"Fibonacci" gets in PHP a list of numbers, works.
"create array with 5 strings" gets 0 items.
Tested on Windows 11 x64. Any Tip?

Hi.
Environment: Ubuntu (KDE Neon if it matters), arch amd64.
Cudatext version: latest - 1.194.0
Plugin version: both from Cudatext manager (which takes from sourceforge and I guess may be a bit outdated) and installation from git (like 30 mins before, so seems to be actual plugin version as well)

Logged in succesfully.
Faced error by failing to get any completions, but haven't faced any error messages.
Next try was with chat - with this I've got next message:

ERROR: GetChatMessage failed. Error: Failed to parse: http://127.0.0.1:None/exa.language_server_pb.LanguageServerService/GetChatMessage

Which seems like a trouble with gathering port.

I've checked init.py for plugin and added few prints to find_port function.
The result is next:
files = os.listdir(self.manager_dir) (Line 217) has empty list, while self.manager_dir contains path like /tmp/codeium_swscy6t7 and this folder is empty, so it seems to be the reason of None port mentioned in error message above.

This thing also causes infinite loop on my device (in all tries to use plugin), which leads to freezes with closing.
I haven't tried to fix it by hardcoding port, so wondering why is there such kind of choosing port 🤔

It will help to find typos in the question.

Python 3.8.10
Loaded session: "history session.json", 0ms, 0 file(s)
Startup: 300ms, plugins: 0ms ()
Init: cuda_codeium
Init: cudatext
Traceback (most recent call last):
  File "/home/user/.config/cudatext/py/cuda_codeium/__init__.py", line 380, in Ask
    self.request_GetChatMessage(question)
  File "/home/user/.config/cudatext/py/cuda_codeium/__init__.py", line 393, in request_GetChatMessage
    from . import proto_pb2
  File "/home/user/.config/cudatext/py/cuda_codeium/proto_pb2.py", line 5, in <module>
    from google.protobuf.internal import builder as _builder
ImportError: cannot import name 'builder' from 'google.protobuf.internal' (/usr/lib/python3/dist-packages/google/protobuf/internal/__init__.py)
ERROR: Exception in CudaText for Ask: ImportError: cannot import name 'builder' from 'google.protobuf.internal' (/usr/lib/python3/dist-packages/google/protobuf/internal/__init__.py)

• got new token
• called the chat
• entered 'what year is now?'

Ubuntu 20.04

can they be published to normal .gif files and placed to repo?

lexer fits:

@Alexey-T, added subj. you can publish new version.