download nodejs via http_proxy

[lookis]

In my case, my corporation has a bad connection to AWS server. please follow http_proxy and https_proxy, thanks

and this issue may need to transfer to libherokubuildpack repo which I can't see.

[schneems]

Hello! Sorry for the delay in response. I'm one of the language maintainers. I primarily work on the Ruby buildpack, but I also contribute to this one.

I'm curious about this statement:

libherokubuildpack repo which I can't see.

Can you give me some more details there? The code should be public can you view https://github.com/heroku/libcnb.rs. If you can't see it, can you give me some details on what you do see or what might be limiting it?

In my case, my corporation has a bad connection to AWS server.

I'm curious about how this is showing up in your builds. Are builds timing out or is something else going wrong? Can you give a specific error message or failure details?

I ask, because perhaps adding a timeout or some retry logic might be beneficial to all. It would be helpful for me to better understand the specifics of how your "bad connection" manifests and causes problems when running pack.

please follow http_proxy and https_proxy, thanks

This sounds like you're asking us to allow using a proxy for downloading binaries from S3. Is that correct? I'm curious what this specific request would solve for you. It sounds like you might proxy requests to S3 through some kind of a cache where you have more control over the connection quality.

Can you give me more details on how you might want that to work? How would you like to tell the buildpack what proxy to use? Any details are helpful: do you have an example or pattern from another project you think we should emulate?

[lookis]

It has been a long time and I have already give up my project. and here is some more information for this issue:

I'm curious about how this is showing up in your builds. Are builds timing out or is something else going wrong? Can you give a specific error message or failure details?

for security reason, my corporation block every connection to outside of LANs, and provides http proxy server to use, so timeout and retry will not work in this situation.

This sounds like you're asking us to allow using a proxy for downloading binaries from S3. Is that correct?

Yes, this is correct, I need to use proxy to bypass the blocking

[schneems]

Thanks for the feedback! I'm sorry you had to give up the project. If you want to try in the future, one possible way could be mapping that URL locally to a different destination. There might be some automated way to do it, on a mac you can change your /etc/hosts manually.

To give you an idea what I mean: I have a small utility that I don't use anymore, but it's design is to decrease my social media use. Where I map urls like "reddit.com" to my localhost so when I want to go into focus mode I don't get distracted https://github.com/schneems/dotfiles/blob/6740d3dcb8a768404fca292a41eb777877f0afd7/bash/functions/focus.sh.

I'm curious, does your company also use Heroku? If so, are you having to go through a proxy as for that as well?

[lookis]

mapping the URL is not a option because SSL will fail. my corporation doesn't use Heroku service, just want to create DevOps pipeline to
build nodejs projects with buildpacks