GHC 2019-04-15

12 comments.

, https://git.io/fjmhr in jarun/nnn
I don't know, Hangouts? Skype?

> On Apr 15, 2019, at 11:40 PM, Mischievous Meerkat <[email protected]> wrote:
> 
> too slow on gitter
> 
> —
> You are receiving this because you were mentioned.
> Reply to this email directly, view it on GitHub, or mute the thread.
> 


, https://git.io/fjmho in jarun/nnn
Hmm, did gitter stopped sending emails all together? ...

, https://git.io/fjmhK in jarun/nnn
Yeah, just saying.

, https://git.io/fjmh6 in jarun/nnn
https://circleci.com/open-source/:

> CircleCI offers 4 Linux containers with no build limits ($2400 annual value) to every open source project in an organization. Simply keeping your repo public will enable this for you.

, https://git.io/fjmhi in jarun/nnn
Sure, but aren't Linux builds unlimited?

, https://git.io/fjmhP in SNH48Live/KVM48
> further encryptions

Lovely. Except even if there are further encryptions, the key has to be embedded in the app, so it's still technically possible to crack (and makes "need to crack new API" an even more apt description), just harder. However, if that's the case, it would probably be even less feasible on iOS. Android apps, though, are known to be easier to decompile, so with this preliminary info, I'd say the task falls squarely on the shoulders of Android developers.

, https://git.io/fjmpC in jarun/nnn
Cool. What about separating different compilers into a build matrix? It could be cleaner, and one failed job wouldn't block all subsequent tests.

, https://git.io/fjmA6 in SNH48Live/KVM48
yaml.load deprecated
====================

```
/usr/local/lib/python3.7/site-packages/kvm48/config.py:241: YAMLLoadWarning: calling yaml.load() without Loader=... is deprecated, as the default Loader is unsafe. Please read https://msg.pyyaml.org/load for full details.
  obj = yaml.load(fp.read())
```

A small deprecation issue, will deal with this when the elephant in the room is taken care of.

, https://git.io/fjmFZ in SNH48Live/KVM48
Okay, the host is `pocketapi.48.cn`, and the API is unknown. I don't have an Android phone to root, and I can't beat cert pinning on iOS. So, until someone breaks cert pinning (probably on Android) and documents the API for me, consider this project dead. Thanks.

Edit: I might want to try Frida though. It's been on my things to try list for quite a few years, but I never got around to it.

, https://git.io/fjmFn in SNH48Live/KVM48
Side remark: v6 of the stupid app still doesn't support iOS's password autofill. Wonder which awful cross-platform framework they're using for this.

, https://git.io/fjmFc in SNH48Live/KVM48
Also, plive.48.cn straight up refuses to connect outside Mainland China, it seems.

With Alibaba Cloud or China Telecom's DNS servers, plive.48.cn resolves to 115.231.96.215 (AS4134 CHINANET-BACKBONE), where port 80 is closed (not ideal, but fair enough) and port 443 open.

With 1.1.1.1 or 8.8.8.8, plive.48.cn resolves to 47.244.116.101 (AS45102 Alibaba (US) Technology Co., Ltd. -- appears to be a Hong Kong node), where both port 80 and port 443 are closed.

, https://git.io/fjmFC in SNH48Live/KVM48
Broken with Koudai48 v6 -- need to crack new API
================================================

Seems broken as of today, which coincides with the stupid v6 update of the app (and which totally bricks v5 -- who the fuck practice that kind of brain dead brutal upgrade?)

```
APIException: POST https://plive.48.cn/livesystem/api/live/v1/memberLivePage {"type":0,"memberId":0,"groupId":10,"lastTime":1555344000000,"limit":100}: JSONDecodeError: Expecting value: line 1 column 1 (char 0)
```

In fact, the error isn't so much that the body is empty; the status code is actually 404, so my error handling code is apparently ineffective.

I'm acknowledging the issue but don't have time to investigate at the moment.