As might you already noticed, @OrangeFoxChat now has topics. I activated them as part of the experiment, let's see how it goes.
💩1
Last time I am working on a new Sophie update. The main purpose of it is to cover all of the problematic places with the current Sophie:
1. Typing system - the python's duck-typing (🦆) is pretty much okay and mostly reliable, but by default, it is not anyhow checked, with Sophie which caused constant NoneType issues.
2. Parsing of text - that's the most fundamental and the most issue-rated aspect of every Telegram bot. In Sophie's case, we are having problems parsing the argument text all the time.
3. Tests - Sophie was always tested manually before releasing, and even if every case was tested during the development, that doesn't mean it wasn't broken while developing.
4. Formatting messages - apart from it being pretty easy, this is the luckiest to break, as you aren't going to test it in every case.
5. Logic, noodle and obsolete code - Sophie exists for over 3 years, and even though there was fundamental refactoring (2.0 version), was incredibly difficult to change. The current version contains 2 libraries to work with a Telegram API, and there's no way to remove one of them.
I have already published my own libraries to resolve the second and third problems:
- https://gitlab.com/SophieBot/ass
- https://gitlab.com/SophieBot/stf
Currently, there are a bunch of unresolved issues, but we're moving fast-forward!
By the way, to this day, Sophie has been used in around 10 000 chats with more than 3 million known users. And had about 5 new chats every 24 hours.
1. Typing system - the python's duck-typing (🦆) is pretty much okay and mostly reliable, but by default, it is not anyhow checked, with Sophie which caused constant NoneType issues.
2. Parsing of text - that's the most fundamental and the most issue-rated aspect of every Telegram bot. In Sophie's case, we are having problems parsing the argument text all the time.
3. Tests - Sophie was always tested manually before releasing, and even if every case was tested during the development, that doesn't mean it wasn't broken while developing.
4. Formatting messages - apart from it being pretty easy, this is the luckiest to break, as you aren't going to test it in every case.
5. Logic, noodle and obsolete code - Sophie exists for over 3 years, and even though there was fundamental refactoring (2.0 version), was incredibly difficult to change. The current version contains 2 libraries to work with a Telegram API, and there's no way to remove one of them.
I have already published my own libraries to resolve the second and third problems:
- https://gitlab.com/SophieBot/ass
- https://gitlab.com/SophieBot/stf
Currently, there are a bunch of unresolved issues, but we're moving fast-forward!
By the way, to this day, Sophie has been used in around 10 000 chats with more than 3 million known users. And had about 5 new chats every 24 hours.
💩1🤝1
Okay, so today was a pretty news-big day.
Let's start with Ukraine - Ukraine got more than 100 rockets from russia on its energy infrastructure, and many cities and areas are having blackout. You can read it basically in every news channel. My opinion is the same.
Another thing that's happening is Twitter, Elon only after 2 weeks of owning most of the twitter's actions and being a CEO already was able to make a lot of nonsense.
He's saying that Twitter does more than 1200 requests with every timeline refresh for each device. That is nonsense as much as it sounds.
He fired an Twitter's App developer that denied it.
He said that Twitter has more than 1000 "useless microservices" and he's going to turn them off.
If it wasn't done before, what changed when Elon joined?
Contabo is experiencing downtimes for more than 2 hours now. Yesterday it had a 2 hours downtime also. That's the reason why orangefox.download and orangefox.tech is down. Although the good NEWS is that the API is downloads server is UP, and sometimes, when CloudFlare is allows to use "always online" feature, you can download a release.
Apparently, one of their "microservices" was shut down
Let's start with Ukraine - Ukraine got more than 100 rockets from russia on its energy infrastructure, and many cities and areas are having blackout. You can read it basically in every news channel. My opinion is the same.
Another thing that's happening is Twitter, Elon only after 2 weeks of owning most of the twitter's actions and being a CEO already was able to make a lot of nonsense.
He's saying that Twitter does more than 1200 requests with every timeline refresh for each device. That is nonsense as much as it sounds.
He fired an Twitter's App developer that denied it.
He said that Twitter has more than 1000 "useless microservices" and he's going to turn them off.
If it wasn't done before, what changed when Elon joined?
Contabo is experiencing downtimes for more than 2 hours now. Yesterday it had a 2 hours downtime also. That's the reason why orangefox.download and orangefox.tech is down. Although the good NEWS is that the API is downloads server is UP, and sometimes, when CloudFlare is allows to use "always online" feature, you can download a release.
Apparently, one of their "microservices" was shut down
👏1💩1
The other message would be about Telegram. Recently Telegram released an update containing Topics. The feature was implemented incredibly awful.
The topics are just service messages about its creations and every user's message replying to it, so you have a chain of messages, which you need to trace if you want to know which topic the message belongs to (if you don't have a topic enabled).
For end-users create a complete mess of messages and replies on the same unsupported message over and over. Topics also share the same permissions, chats and restrictions between all of them, so it makes it impossible to change users' permission in one of a topic.
For bot developers:
This is a special section, that is a complete nightmare.
1. Enabling topics means all of the messages would be a reply, if you're code handles replies, it will handle the service message over and over, over and over.
2. In topics-enabled chats there's no way to send a message without a reply, so yes, all of the code in your bots that just sends a message to chat is broken with topics enabled!
3. There are no methods, messages or anything to detect when the topic is deleted. The code can't know whatever topic exists or now. The same with its renaming.
The topics are just service messages about its creations and every user's message replying to it, so you have a chain of messages, which you need to trace if you want to know which topic the message belongs to (if you don't have a topic enabled).
For end-users create a complete mess of messages and replies on the same unsupported message over and over. Topics also share the same permissions, chats and restrictions between all of them, so it makes it impossible to change users' permission in one of a topic.
For bot developers:
This is a special section, that is a complete nightmare.
1. Enabling topics means all of the messages would be a reply, if you're code handles replies, it will handle the service message over and over, over and over.
2. In topics-enabled chats there's no way to send a message without a reply, so yes, all of the code in your bots that just sends a message to chat is broken with topics enabled!
3. There are no methods, messages or anything to detect when the topic is deleted. The code can't know whatever topic exists or now. The same with its renaming.
💩1
The issue with Contabo seems to be fixed.
Tomorrow I'll provide a list of what can be done to negotiate such situations in future.
Tomorrow I'll provide a list of what can be done to negotiate such situations in future.
💩1
#YachaBlog
Last time I am working on a new Sophie update. The main purpose of it is to cover all of the problematic places with the current Sophie: 1. Typing system - the python's duck-typing (🦆) is pretty much okay and mostly reliable, but by default, it is not anyhow…
Talking about ASS, I had the idea of parsing arguments like that for more than 3 years and I'm finally happy there's something similar to it we're having now.
Here you can see an example of what ASS is able to do.
Here you can see an example of what ASS is able to do.
💩1
Regarding the bw.owo.family status.
Today it was re-provisioned, as I won't be able to restore the OS it was powered on for several reasons.
The general issue is that I lost root access to it.
That happened when I initially provisioned the VM of Bitwarden, I had to change the default Linux credentials, like I'm always doing, but due to password manager was, obviously, down, I had to note the passwords somewhere else and forgot them.
My usual setups also include SSH keys, which were also there, but due to my trying to log in, I was blocked by the tallow, so I had no way to login into the VM in any possible usual way.
Resetting root password on Linux.
The OS I'm using is openSUSE MicroOS which has an inbuilt btrfs snapshots file system, where every update creates a snapshot which contains changes over the previous one. The /etc is also a 2 directories overlay that contains changes over the last made snapshot. This mounting system works awesome when the system is unable to boot and requires a rollback, but not when you want to change the last snapshot. Why? Usually, you'd boot it with init=/bin/sh, where you had no systemd, and so on no D-Bus, which is required for a snapper, a SUSE tool to manage snapshots. You also cannot change /etc directory directly in the last snapshot, since it'll be overwritten by an overlay.
How it was restored.
By simply copying the files from a /var volume (which is a dedicated partition) and re-installing the OS.
What are the findings?
Do not lose access, obviously. The other thing is making a workaround to login into it.
The new OrangeFox's Virtual Machines now have LDAP access, which will settle the current issue.
The current state.
bw.owo.family instance is on, has all the user's data, and now fixes the issues with the latest Bitwarden apps.
Moreover, now not the OS only, but also the Bitwarden server is now updated automatically, thanks to podman-auto-update.
----
The next issue that was fixed is the unstable work of mail.orangefox.tech. The issue is basically with insufficient RAM, the mail server was moved to a more suitable place. Moreover, it now also does self-update itself.
Today it was re-provisioned, as I won't be able to restore the OS it was powered on for several reasons.
The general issue is that I lost root access to it.
That happened when I initially provisioned the VM of Bitwarden, I had to change the default Linux credentials, like I'm always doing, but due to password manager was, obviously, down, I had to note the passwords somewhere else and forgot them.
My usual setups also include SSH keys, which were also there, but due to my trying to log in, I was blocked by the tallow, so I had no way to login into the VM in any possible usual way.
Resetting root password on Linux.
The OS I'm using is openSUSE MicroOS which has an inbuilt btrfs snapshots file system, where every update creates a snapshot which contains changes over the previous one. The /etc is also a 2 directories overlay that contains changes over the last made snapshot. This mounting system works awesome when the system is unable to boot and requires a rollback, but not when you want to change the last snapshot. Why? Usually, you'd boot it with init=/bin/sh, where you had no systemd, and so on no D-Bus, which is required for a snapper, a SUSE tool to manage snapshots. You also cannot change /etc directory directly in the last snapshot, since it'll be overwritten by an overlay.
How it was restored.
By simply copying the files from a /var volume (which is a dedicated partition) and re-installing the OS.
What are the findings?
Do not lose access, obviously. The other thing is making a workaround to login into it.
The new OrangeFox's Virtual Machines now have LDAP access, which will settle the current issue.
The current state.
bw.owo.family instance is on, has all the user's data, and now fixes the issues with the latest Bitwarden apps.
Moreover, now not the OS only, but also the Bitwarden server is now updated automatically, thanks to podman-auto-update.
----
The next issue that was fixed is the unstable work of mail.orangefox.tech. The issue is basically with insufficient RAM, the mail server was moved to a more suitable place. Moreover, it now also does self-update itself.
💩2
Thanks @unfoxo and @h_2_o0 for their suggestions.
I have already fixed the CloudFlare CDN cache issue with orangefox.download. So the server will have less requests.
I have already fixed the CloudFlare CDN cache issue with orangefox.download. So the server will have less requests.
👍1💩1
I just found that orangefox.download// is not working. So yeah with double slash. Gonna fix that soon.
OrangeFox Recovery Downloads
Download OrangeFox Recovery
Download OrangeFox Recovery for your Android device
💩1