Issues&releases are maintained by the QA team. Devs may be “summoned” to a particular issue when required but most of the work is done by QA. Issues are also automatically synced to JIRA. There’re many changes and re-labeling by “adguard-bot”, this all actually comes from JIRA sync.
Hi, let me please answer from the personal account as at this time there’s no one to answer from the company account.
It’s not that hard to miss considering that the mirrored branch is not the default one on Github so you don’t see when you just open the repo.
Internally we use self-hosted Git so devs also just don’t follow the state of the source code on GH.
🚨@GitHub has flagged several key contributors to ad-blocking filter lists, impacting the popular List-KR and disrupting our workflow. We suspect this is due to automated protections, not malicious activity (1/3)🧵
@LWGShane@GRos3s@AdGuard The wording is not ideal here. The problem is that the card payment provider that we're using does not allow changing email address of an existing subscription and thus it will continue to send emails to the old address.
The problem is that is still requires routing all the traffic through the app which seems like an overkill for just SNI splitting.
The alternative approach would be to do that for a selected list of domains and just “redirect” those domains to a local snippet proxy. But then the question is which domains to redirect? Probably, requires UI to configure and overall, it will be quite a big task if we go this way. That’s what was stopping us from doing that.
I’d honestly prefer having proper CoreLibs filtering ported to iOS, if they at some point allow sideloading in EU, we’ll start working on that.
🚨Last Call for the Ad-Filtering Dev Summit starting TOMORROW!🚨
Don't miss discussions from @ay_meshkov, @gertrudkolb, @pgl, @fanboynz, @avl7771, @LisetteMeij & more on urgent ad-blocking community topics. Secure your spot now! #AFDS2023
https://t.co/4RF8wbkFbB
Well, it's not supposed to be used as a library yet, it does not clean up connections properly at the moment. Maybe one day I'll refurbish it into "libgocurl", then it'll be the time.
Regarding the docs, it may sound strange, but that's how it works :)
1. https://t.co/H8RdVORgTV is used as a relay, a client-facing server. In the case of CF you probably can use any domain as a relay since they are on the same anycast network, what matters is the IP address.
2. Unencrypted (outer) ClientHello uses the domain from the DNS record, it's "https://t.co/6aE8vSzSfO" currently.
3. And finally the encrypted (inner) ClientHello will use the server name that you actually want to connect to.
Curl is undoubtedly an awesome tool, but it's been too often lately when I couldn't do what I needed with it. I am too lame to patch curl so I made my own instead: https://t.co/WFgwwMua6E
It's basically a simplified version of curl written in pure Go, but it also adds some new stuff: encrypted ClientHello support, json output, splitting clienthello, HTTP/3 out-of-the-box.
You didn’t hear about it because they target russian speakers only. A few years ago it’s market share in russia was about 25%, now I think it’s even higher.
Besides that, it’s probably the most serious Chromium fork out there if counting the sheer number of original features and modifications, Opera and Brave aren’t even close.
Sadly, there are also dubious if not outrageous decisions like this one with limiting ad blockers permissions.