Dear Copilot, when I say it works, don't fix it.

[Koladweep]

Type: Bug

I gave feedback, not a prompt. It should understand when to code, and more importantly, when not to code.

Extension version: 0.20.3 VS Code version: Code 1.93.1 (38c31bc77e0dd6ae88a4e9cc93428cc27a56ba40, 2024-09-11T17:20:05.685Z) OS version: Windows_NT x64 10.0.20348 Modes:

System Info

|Item|Value| |---|---| |CPUs|Intel(R) Core(TM) i5-2430M CPU @ 2.40GHz (4 x 2395)| |GPU Status|2d_canvas: enabled
canvas_oop_rasterization: disabled_off
direct_rendering_display_compositor: disabled_off_ok
gpu_compositing: enabled
multiple_raster_threads: enabled_on
opengl: enabled_on
rasterization: unavailable_off
raw_draw: disabled_off_ok
skia_graphite: disabled_off
video_decode: enabled
video_encode: unavailable_off
vulkan: disabled_off
webgl: enabled
webgl2: unavailable_off
webgpu: enabled
webnn: disabled_off| |Load (avg)|undefined| |Memory (System)|7.94GB (2.52GB free)| |Process Argv|--crash-reporter-id 2f5541c3-8040-4b81-85fd-0bfc23b778ea| |Screen Reader|yes| |VM|67%|

A/B Experiments

``` vsliv368cf:30146710 vspor879:30202332 vspor708:30202333 vspor363:30204092 vscod805cf:30301675 binariesv615:30325510 vsaa593:30376534 py29gd2263:31024239 c4g48928:30535728 azure-dev_surveyone:30548225 962ge761:30959799 pythongtdpath:30769146 welcomedialogc:30910334 pythonnoceb:30805159 asynctok:30898717 pythonmypyd1:30879173 h48ei257:31000450 pythontbext0:30879054 accentitlementst:30995554 dsvsc016:30899300 dsvsc017:30899301 dsvsc018:30899302 cppperfnew:31000557 dsvsc020:30976470 pythonait:31006305 dsvsc021:30996838 da93g388:31013173 a69g1124:31058053 dvdeprecation:31068756 dwnewjupyter:31046869 impr_priority:31102340 nativerepl1:31139838 refactort:31108082 pythonrstrctxt:31112756 flightc:31134773 wkspc-onlycs-t:31132770 wkspc-ranged-t:31125599 ei213698:31121563 ```

[digitarald]

@Koladweep do you have an example where it provided code but should not have?

[Koladweep]

I'm sorry i cleared out my history before coming across your message. I can describe the issue. The issue arises during conversation after Copilot has already generated a piece of code. You read it and A) either try it or B) ask for some clarification [not code], it often launches into another full-blown code before the previous one was even tried or reviewed - sometimes throttling/truncating midway. This ends up in a lot of waste generated. More like early day issues with ChatGPT.

PS. [unrelated] Plus it doesn't ask before switching context to the selected tab. The option to select focus in each response might be helpful.