Add a MemoryStream buffer to ReadStream first

[taooceros]

fix #598

[taooceros]

@Garulf https://ci.appveyor.com/api/buildjobs/ylh2uokiws4ee2hw/artifacts/Output%2FPackages%2FFlow-Launcher-v1.8.0.exe

[jjw24]

@taooceros could you let me know what's tested please

[taooceros]

@taooceros could you let me know what's tested please

I have tested the scenario descripted in #598 (a very large json output). It works fine with very large json (1500000 items).

The change is really simple. I just add a memorystream layer to read the console output from the process allowing it to exit, which causes the hang (not a lot refactor as I say in #598 due to consideration of possible future change).

Probably a regular memorystream will be fine because we generally won't need to deal with that much item, but I just found the recycle one

[Garulf]

Testing in a moment, thank you so much for looking into this!

[Garulf]

@taooceros Works perfectly! All values are being returned and Flow Launcher no longer gets stuck in a loading state.

[taooceros]

@taooceros Works perfectly! All values are being returned and Flow Launcher no longer gets stuck in a loading state.

Perfect! Thank you so much for reporting the bug!

[Garulf]

Not sure whats causing this error but it might be releated:

03:09:09.2594-04:00 - ERROR - JsonRPCPlugin.ExecuteAsync - Exception for filename <C:\Users\William\AppData\Local\Programs\Python\Python39\pythonw.exe> with argument <>
EXCEPTION OCCURS: System.Threading.Tasks.TaskCanceledException: A task was canceled.
   at async Task System.IO.Stream.CopyToAsyncInternal(Stream destination, int bufferSize, CancellationToken cancellationToken)
   at async Task<Stream> Flow.Launcher.Core.Plugin.JsonRPCPlugin.ExecuteAsync(ProcessStartInfo startInfo, CancellationToken token) x 2 in C:/projects/flow-launcher/Flow.Launcher.Core/Plugin/JsonRPCPlugin.cs:line 288

Flow Launcher seems to work regardless

[taooceros]

This exception shouldn't be logged. Weird. Will take a look later.

[jjw24]

@taooceros let me know when this is ready, assuming you are still looking into the log above.

[taooceros]

@jjw24 Good to review. I add another manual process.kill() to kill the process once the token is expired.

[taooceros]

Hold on, I found an issue, let me fix it.

[taooceros]

Done