Upgit is creating a large history.log file in the same directory as the binary despite passing --no-log / -n. Also, 544KB of data takes 3 minutes to upload.

[futuremotiondev]

This is the Powershell script I'm using right now:

https://gist.github.com/futuremotiondev/01f46fc608d20cae6ebd8da6b4fdaca3

I passed 183 SVG files to upgit (total file size of 544KB), and it took about ~3 minutes for the operation to complete. Here is the history.log file that was created during this process:

history.log

I don't know why it's generating this log when I'm explicitly passing --no-log / -n. The log file itself is 51KB!

Maybe there is a bottleneck in speed because of the constant disk IO as upgit writes to this log file.

I switched my [replacements] configuration setting to:

[replacements]
"raw.githubusercontent.com" = "cdn.jsdelivr.net/gh"
"/main" = "@main"

Despite this change, the upload process is still incredibly slow. There is no way that transferring 544KB of data to GitHub should take three minutes. That's insane.

Can you please look in to optimizing the upload process? Off the top of my head:

1. Support parallel uploads
2. Allow end users to completely disable history.log creation
3. Double check that your upload routine / function is not being bottle necked by some kind of function call?

I don't know. I really love the idea of upgit but the performance is really a let down.

[futuremotiondev]

Attaching the files I used for the upload for you to reproduce if this helps

TestUploadSet.zip

[pluveto]

I‘ll fix this in next version