Jongwook Choi comments

Results 564 comments of


                                            Jongwook Choi

Output flooded with EventLoop related messages

Questions: - Does that happen also on pynvim 0.4.3 or only on pynvim 0.5.0? - What is your python version? - Can you make a simple reproduction by any chance?...

Output flooded with EventLoop related messages

Without a reasonable reproduction, bugfix might take a bit long. Please let us know when you've got some free cycles. My guess is that fd4247ced2b536c82fe76c52a6a6042eebb31ad4 is the relevant difference against...

Output flooded with EventLoop related messages

Hmm if https://github.com/neovim/pynvim/pull/556#issuecomment-1851031322 doesn't fix this, I'll need a repro to help you troubleshoot this issue. At least any traceback information will be needed. Can you try enabling neovim logging?...

Output flooded with EventLoop related messages

Minimal repro: ```python vim = pynvim.attach('child', argv=['nvim', '--embed', '--clean', '-i', 'NONE']) vim.close() vim = pynvim.attach('child', argv=['nvim', '--embed', '--clean', '-i', 'NONE']) vim.close() # Result Loop that handles pid 1048 is closed...

refactor(lsp): add typing and docs for LSP handlers

Rebased to the current master, fixing conflicts with other concurrent LSP-related works (e.g. #27376 #27420 #27427 #27443 #27753). There are a few minor TODOs remaining for this PR to be...

Feature: Run in a terminal, but send output to quickfix?

That's quite ugly -- but quite nice workaround. I think it would be quite non-trivial to stream the output from the terminal (but technically possible, see https://github.com/tpope/vim-dispatch), but first-class and...

Feature: Run in a terminal, but send output to quickfix?

``` :AsyncRun -mode=term -pos=TAB -program=capture ls -la ``` Why does it print some verbose messages like the following? ``` (1 of 1): [ls] (1 of 29): [ls] Press ENTER or...

multi-gpu inference produces broken output

@dji-transpire Can you check the versions for CUDA, CUDNN, CUBLAS, NVIDIA driver versions, or any relevant SDKs -- were they the same? The model (generation) of GPU are already different...

multi-gpu inference produces broken output

Note: A workaround for this bug is to use the CMake flag `-DLLAMA_CUDA_PEER_MAX_BATCH_SIZE=0` when building llama.cpp, as done in jmorganca/ollama#1261. Or more simply: ``` make LLAMA_CUBLAS=1 LLAMA_CUDA_PEER_MAX_BATCH_SIZE=0 ``` This will...

[MAC, intel chip] [app crash] crashs while downloading model or hiting record button

``` [1] 77385 illegal hardware instruction /Applications/Buzz.app/Contents/MacOS/Buzz ``` This is probably due to #6, but now this should be possible.