Closed JoshuaWatt closed 2 years ago
Tactically speaking, this will accomplish what it says on the tin.
I don't see the harm, but what situation motivates this? It doesn't immediately seem to address the internal coming problem inside Garmin, unless you're planning on just setting "no_proxy=*" while still leaving http_proxy set. That seems a little bit indirect; if this is the case, wouldn't just avoiding setting http_proxy in the first place be better?
Currently we are spamming the logs by using the proxy to access the internal servers, so this should at least increase the signal to noise ratio, aside from simply being useful in general
Okay. I see that wget pays attention to this. Are there other environment variables that serve the same role for stuff like Git's built-in HTTP fetcher, which we should add in here too?
Okay. I see that wget pays attention to this. Are there other environment variables that serve the same role for stuff like Git's built-in HTTP fetcher, which we should add in here too?
According to https://about.gitlab.com/blog/2021/01/27/we-need-to-talk-no-proxy/ this should give us the best coverage
Okay. I see that wget pays attention to this. Are there other environment variables that serve the same role for stuff like Git's built-in HTTP fetcher, which we should add in here too?
According to https://about.gitlab.com/blog/2021/01/27/we-need-to-talk-no-proxy/ this should give us the best coverage
TL;DR no_proxy
is supported by pretty much everything, although the semantics slightly vary
Okay. If you already know that does what you want, seems fine to me.
Updates pyrex to pass no_proxy in addition to http_proxy and https_proxy