Google connections.

[begin-theadventure]

Why does this app connect to google.com and google.com.home?

[vinifmor]

It is just an internet connection check... google.com is widely used in many countries. This only call is done at:
https://github.com/vinifmor/bauh/blob/master/bauh/commons/internet.py

[netstat-peanut]

It is just an internet connection check... google.com is widely used in many countries. This only call is done at: https://github.com/vinifmor/bauh/blob/master/bauh/commons/internet.py

Would it be possible to revisit the use of google.com here? I understand the need for prevalence but it's a controversial company among members of the FOSS community nonetheless. For instance, all G-owned domains (incl. TLD) are blocked using DNSBL on my local network and their netblocks are null routed, meaning those measures must be sidestepped for application installation/updates via bauh.

I took a stab at creating an index of hosts that would resolve in series; proceeding if any 1 name resolves. Unfortunately I'm little more than a copy/paste code monkey and the back-asswards nature of Python makes my head spin.

The 2LD I was using along with justification:

• duckduckgo.com (popular, high-traffic commercial website, albeit less controversial)
• internic.net (diversity of TLD w/ benign, well-established entity)
• w3.org (well-established, benign entity and diversity of TLD using less-likely target of censorship. )
• iec.ch (diversity of TLD w/ benign, well-established non-profit + international TLD)
• mit.edu (diversity of TLD w/ benign, well-established entity. censorship contingent upon prevailing gov)

[vinifmor]

Sure... I'll take a look on those you mentioned, and change the call.

[netstat-peanut]

Sure... I'll take a look on those you mentioned, and change the call.

Thanks @vinifmor!

Would be feasible to wrap gethostbyname around the update/installation routine--essentially leveraging whichever hostnames require resolution to accomplish what was asked (i.e. flathub.org)? I'm sure class InternetChecker originates from just such a procedure elsewhere and I reckon google.com could be replaced with a (perhaps preexisting?) variable drawn from what's being executed in the moment...

I've probably done little more than expose my peanut brain for what it is but I hate to submit issues without some manner of suggestion :)

[vinifmor]

Folks, I've replaced google.com by w3.org, and the changes are available on the staging branch. @netstat-peanut about the internet checking code: it's a general way for the app's core to manage it's operations life cycles and also provide this information for every supported packaging technology.

[netstat-peanut]

thank you @vinifmor !