Conda environment activation does not work for R terminals

[robertwinkler-research]

Type: Bug

This does not work for R environments. A common and very nice workflow before was managing R installations with different conda environments.

Steps to reproduce:
When conda environment is selected with python: select interpreter, on opening R console.

Outcome: conda activate <selected_env> is run, which fails because we are in R. Suggested solutions in the related issue #25291 did not work. Specifying "python.useEnvironmentsExtension": true and "python-envs.terminal.autoActivationType": "shellStartup" leads to the same.

Expected outcome: Silent activation of conda environment for R terminal. Please somehow restore the before behaviour.

Environment Information

Extension Version: 1.1.12031011
Python Extension Version: 2025.11.2025072201

Registered Environment Managers (3):
  - ms-python.python:system (Global)
  - ms-python.python:venv (venv)
  - ms-python.python:conda (Conda)

Total Available Environments: 10
Environment Details:
  1. Python 3.9.23 (3.9.23.final.0) - /bin/python
  2. Python 3.9.23 (3.9.23.final.0) - /usr/bin/python
  3. epigenetics-atlas (3.13.3) (3.13.3) - /home/RWinkler/.conda/envs/epigenetics-atlas
  4. chipbinner_env (3.12.11) (3.12.11) - /home/RWinkler/.conda/envs/chipbinner_env
  5. epiatl-downstr (3.12.10) (3.12.10) - /home/RWinkler/.conda/envs/epiatl-downstr
  6. epicompare_env (3.12.10) (3.12.10) - /home/RWinkler/.conda/envs/epicompare_env
  7. macs3_peak_calling (3.12.10) (3.12.10) - /home/RWinkler/.conda/envs/macs3_peak_calling
  8. mcp_env (3.12.10) (3.12.10) - /home/RWinkler/.conda/envs/mcp_env
  9. epigenetics_peak_calling (3.11.12) (3.11.12) - /home/RWinkler/.conda/envs/epigenetics_peak_calling
  10. epic2_peak_calling (3.9.23) (3.9.23) - /home/RWinkler/.conda/envs/epic2_peak_calling

Python Projects (1):
  1. /home/archive/LR_RWinkler/epigenetics_atlas
     Environment: epiatl-downstr (3.12.10)

Extension Settings:
  Default Environment Manager: ms-python.python:conda
  Default Package Manager: ms-python.python:conda
  Terminal Auto Activation: command

Extension version: 1.0.0
VS Code version: Code 1.102.2 (c306e94f98122556ca081f527b466015e1bc37b0, 2025-07-22T12:15:48.520Z)
OS version: Windows_NT x64 10.0.26100
Modes:
Remote OS version: Linux x64 5.14.0-578.el9.x86_64

System Info

Item	Value
CPUs	AMD Ryzen 7 PRO 6850U with Radeon Graphics (16 x 2695)
GPU Status	2d_canvas: enabled canvas_oop_rasterization: enabled_on direct_rendering_display_compositor: disabled_off_ok gpu_compositing: enabled multiple_raster_threads: enabled_on opengl: enabled_on rasterization: enabled raw_draw: disabled_off_ok skia_graphite: disabled_off video_decode: enabled video_encode: enabled vulkan: disabled_off webgl: enabled webgl2: enabled webgpu: enabled webnn: disabled_off
Load (avg)	undefined
Memory (System)	30.76GB (9.57GB free)
Process Argv	--crash-reporter-id 185a55ad-3945-4d53-a2b2-7311cd0331bb
Screen Reader	no
VM	0%

Item	Value
Remote	SSH: 192.168.155.62
OS	Linux x64 5.14.0-578.el9.x86_64
CPUs	AMD EPYC 7763 64-Core Processor (256 x 2450)
Memory (System)	1007.17GB (954.98GB free)
VM	0%

A/B Experiments

vsliv368cf:30146710
vswsl492cf:30256860
binariesv615:30325510
h48ei257:31000450
nativeloc1:31344060
dwcopilot:31170013
6074i472:31201624
dwoutputs:31242946
9064b325:31222308
copilot_t_ci:31333650
e5gg6876:31282496
pythoneinst12:31285622
c7cif404:31314491
pythonpulldiag:31343502
996jf627:31283433
pythonrdcb7:31342333
usemplatestapi:31297334
0aa6g176:31307128
747dc170:31275177
aj953862:31281341
generatesymbolt:31295002
convertfstringf:31295003
9d2cg352:31346308
usemarketplace:31343026
nesew2to5:31336538
agentclaude:31350858
4f60g487:31327383
nes-diff-11:31337487
replacestringexc:31350595
nes-set-on:31351930
6abeh943:31336334
yijiwantestdri0626-c:31336931
0927b901:31350571
f76d9909:31348711
45650338:31354155
0cj2b977:31352657
a02b4192-color:31352671

[eleanorjboyd]

Hi! Sorry you are running into this problem. Is this a new bug you are experiencing that used to work before? With "python-envs.terminal.autoActivationType": "command" do you see anything injected into the shell on startup?

[robertwinkler-research]

Hi! Thanks for looking at this. This is a new bug since maybe a day before I opened the issue, I am guessing that's when new python environments extension started to be used? Before, R terminals would open in activated environment without explicit activation after terminal opened, silently.
With "python-envs.terminal.autoActivationType": "command", this is the terminal output upon opening R terminal. Let me know what more information would be useful. Are you able to reproduce?

conda activate my_env
R version 4.5.1 (2025-06-13) -- "Great Square Root"
Platform: x86_64-redhat-linux-gnu (64-bit)

r$> conda activate my_env
Error: unexpected symbol in "conda activate"

r$>

[robertwinkler-research]

Trying again now, with shellStartup, there is no explicit activation but also no activation of the environment at all with R, output becomes the below, but environment is not activated.

R version 4.5.1 (2025-06-13) -- "Great Square Root"
Platform: x86_64-redhat-linux-gnu (64-bit)

r$>

[anthonykim1]

@robertwinkler-research What happens if you uninstall Python environment extension? I still don't remember Python extension supporting R terminals.

[robertwinkler-research]

Deactivating the python environment extension leads to the same behaviour.
I will say I do not really understand how environment management and activation works or used to work in vscode, but previously, implicit activation worked for R terminals and I would love that to work again. The R terminal would open within activated conda environment.
In my very naive mind upon opening an R terminal the environment could be activated before running R/radian, similarly to what bash terminal does when shellStartup option is set? Again, I don't know how any of this works, sorry. I also understand supporting R is not a goal of the python environments extension, but previously this worked and it would be a shame to regress.
The extensions current default behaviour of running conda activate in the opened R terminal leading to error can't be what is desired in any case.

[eleanorjboyd]

@anthonykim1 could you follow up here