Please wait...
0 files
Subject: Re: Citadel 942 released -- big changes to encryption
And yes, filesystems with built-in snapshotting are an insanely great
blessing.
To be clear, I am talking about block-layers with built-in snapshotting. btrfs, on the other hand, is no bueno from what I am hearing.
Subject: Re: Citadel 942 released -- big changes to encryption
In the future I'd like to have Citadel create its own backups by creating database replicas, and standing up multiple servers in a master/slave configuration.
That's a somewhat distant future plan, though.
Subject: Re: Citadel 942 released -- big changes to encryption
Article: Examining btrfs, Linuxbs perpetually half-finished filesystem
This btrfs filesystem overview highlights some longstanding shortcomings
https://arstechnica.com/gadgets/2021/09/examining-btrfs-linuxs-perpetually-half-finished-filesystem/
Subject: Re: Citadel 942 released -- big changes to encryption
They're obviously trying to make it equivalent to WAFL. NetApp has a few decades head start on them.
Since this room is really the Citadel announcements mailing list, I'll bring it full circle. The absolute best way to do a backup of a Citadel database is to snapshot it and then make a copy of the snapshot. Berkeley DB tends to be friendly to online backups as long as you copy the cdb.* files first and the log.* files second, but for *any* database there is never anything better than a point-in-time snapshot.
Speaking of mailing lists, we're still being blocked by gmail, so many of you are not reading this. If you happen to see this message and you are a gmail user, please do me a favor and find *any* mail from citadel.org in your spam folder, and mark it as "not spam". This will help teach Google that we are not spammers.
Subject: Re: Citadel 942 released -- big changes to encryption
Got mine upgraded but the versions dont match for core and webcit? Server says 942 but webcit still says 941 ( easy install ). Will have to tackle Lets Encrypt bot this weekend.
And sort of related to the last 2 posts: i wonder what the max load we could take with a decent sized box. ( concurrent.. not just mailboxes registered )
Subject: Re: Citadel 942 released -- big changes to encryption
oooo fancy. Will sort of miss the old classic look, but everyone else is moving on to 'modern look'.
needs the "Edit" button!! add that in!
Interested in seeing what the new WebCit will look like? Here's a nice screenshot of the forum view. This is a very early preview, just to give you an idea of the kind of future we're building for everyone's favorite software.Yes, this is very much vaporware right now! It won't be ready for some time yet, because only the forum view is complete. As we develop the mailbox, calendar, contacts, blog, and wiki views, and fit out all of the settings and administrative tools, we'll post more screenshots. This is just to give you an idea of what's in store. WebCit-NG follows a "REST-first, DAV-first" development methodology, and the client-side UI follows the "responsive" development style so it will look great on both large and small screens.In case anyone's wondering, we're using the W3.CSS framework, which is more lightweight than Bootstrap. Much of the iconography is from Font Awesome. And no, we're not using JQuery.
I do believe that is more than just interface stuff. That would also need some back end changes which a bad thing to do when you are trying to create new GUI for an existing app. Its a good way to get yourself into trouble.
needs the "Edit" button!! add that in!
Also remember that this is all REST and DAV behind the scenes, so it's going to be a lot easier to build new interfaces than it was before. Funny story about that -- the original WebCit had the entire UI hard coded in the C program that ran the web server. Eventually that was replaced with something that was driven by a template engine, to try to separate the user interface from the application logic. But the template engine ended up being *more* complex than the C code it replaced. Eventually the entire WebCit framework became so unmaintainable that it became clear it was time for a fresh approach.
Normally I am not a fan of the "throw it all away and start over" approach -- it's almost always a mistake. In this case though, there was so much awful code (much of it mine) and so much cruft to deal with obsolete browsers, that the choice was obvious.
I looked through the commit history for webcit-ng and saw that the first commit was five years ago. Sun Mar 12 00:30:06 2017 to be exact. For a project that I am this excited about, it sure is taking a long time.
Admittedly, there were a lot of side quests. We ended up rewriting most of the protocol handlers in the Citadel Server core, replacing thousands of lines of event-driven code with simple wrappers around libcurl. We removed the legacy CitaNet code, since no one was using it. We completely rewrote the mailing list server, which is now more robust and more secure. Most significantly of all, we moved most of the remaining external files into the database, allowing us to package the Citadel system as a Docker container. I do believe that we now have the most robust and feature-rich communications server on the market, brought to the world as a 100% open source offering, no strings attached.
Those efforts were necessary. During that time, however, WebCit-NG was put on the back burner. It is now priority one, and will remain so until it is functionally complete and replaces WebCit-classic. Our user community deserves it. We will post more screenshots to this forum from time to time, so you can watch it come together. I am excited about this next phase of the Citadel story, because this will not only bring an ultra-modernized user interface, but also enable us to interface with even more third-party software.
As a reminder: if you are reading this via the mailing list and you would like to unsubscribe, go to https://uncensored.citadel.org/listsub to change your subscription preferences.
I have trepidations about the scale of change, being very much in my comfort zone with the current version of Citadel I have running in VM - very stable and requiring little user interaction.
But... I can see that this version will probably light a fire for me to move forward. I think it probably will attract more users, too.
For many years, a lot of people have been turned off by the current web interface, because it isn't clear on whether it wants to be a personal information manager or a community bulletin board. It's too "weird" for some. In the new interface, the different things the Citadel System can do are clearly separated into different sections of the site. For example, a bulletin board will have a lot of rooms under the section called "Forums" and it will look and act like a traditional web forum. When you go to the section called "Mail" it will display a tree of your mail folders and the site will act like a mail system.
Likewise for Calendars, Contacts, Blogs, Wikis, and anything else we add in the future.
The "traditional Citadel look and feel" will be maintained in the text client.
Based on how we're building things right now, it should be easy to extend the system with new "applications" that present user interfaces of their own, in addition to the existing ones.
One might even begin comparing it to things like OwnCloud and NextCloud, except that Citadel is really good at communicating with the rest of the world.
Will be a shame to lose the 'room' label, but i know, we gotta get out of the 80s at some point if we are going to survive.
Mentioning nextcloud, and just a random thought: i wonder if a plugin at some point would be appropriate? Sure, they wouldn't be using the NG interface that you are working hard on, but at least the back end would be used, potentially increasing the user base. The email plugin already will work, unsure about dav to calendars and contacts ( ill have to look on my install ), so might be just a couple of pieces that need to be created for it. ( and i know, even if that were to happen its WAY down the road )
will there be a migration assistant from old citadel to new citadel?
That will not be necessary. Citadel Server is not changing. At some point in the future you will receive the new UI as part of the regular updates.
And yes, the possibility of integrating Citadel with NextCloud is certainly there; moreso now that we are REST/DAV enabling everything.
I think this is the logical way to move forward. The Web interface already is really only superficially in the format of a traditional Citadel - so changing it to a more Forum based presentation doesn't really bother me much - especially if it makes it more accessible to people with no nostalgia for the Citadel experience.
On the other hand, maintaining the Citadel Text Client presentation should satisfy any who have a nostalgic attachment to the "genuine Citadel presentation/experience." It is really a "best of both worlds" compromise.
Already, at this point - the Citadel experience is simply a legacy evolution of the NAME and philosophy of Cit-86 et. al. into a more modern experience.
I mean, modern Linux, modern OS X and modern Windows have very little in common with where they originated. Same for Photoshop or any other experience and UI. Successful platforms usually have nothing in common with their origins other than the name and the original purpose.
Sat Feb 19 2022 14:21:40 EST from IGnatius T FoobarThe new version will become the only available version when it is completed.
For many years, a lot of people have been turned off by the current web interface, because it isn't clear on whether it wants to be a personal information manager or a community bulletin board. It's too "weird" for some. In the new interface, the different things the Citadel System can do are clearly separated into different sections of the site. For example, a bulletin board will have a lot of rooms under the section called "Forums" and it will look and act like a traditional web forum. When you go to the section called "Mail" it will display a tree of your mail folders and the site will act like a mail system.
Likewise for Calendars, Contacts, Blogs, Wikis, and anything else we add in the future.
The "traditional Citadel look and feel" will be maintained in the text client.
Based on how we're building things right now, it should be easy to extend the system with new "applications" that present user interfaces of their own, in addition to the existing ones.
One might even begin comparing it to things like OwnCloud and NextCloud, except that Citadel is really good at communicating with the rest of the world.
I am coming into this late. Its always a difficult thing to decide if you will support classic features and functions or just totally abandon them in future versions. I wish Ubiquiti understood this concept but whatever.
I guess you dont have enough software releases with versions over time and feedback to gauge which features would be missed if they were retired. The hard part as I see it is that you have such a diverse installed user base that use one, some or all of citadel functions. Do you have a survey of whats out there that asks "Hey if the tasks function or the notes function evaporated, would you care?"
That might be a full time 100k a year position. In my idiot opinion, you have:
Enterprise level groupware with active support.
Albeit the support is not provided in real time, Citadel as a downloaded Linux install does more, works better and is more stable than most of the junk commercially available.
I guess a fork is reasonable? One limitation I see is the software is not easily integrated with the ways of the evil empires.