Matt
Administrators
-
Last visited
-
Posts
230 -
Reputation
283
Everything posted by Matt
-
It’s about to be released. ☠️
Beautiful 🥰
Yep, although v5 is a big upgrade, it’s still just an upgrade so all translated strings remain translated.
Can we have dog videos please? Not a cat fan tbh.
Already reported and fixed for the next release.
They are an open source product using the open source version. They wrote that blog one month after the version of CKEditor that requires attribution on every editor. If they have not added that, then they risk compliance issues - and CKEditor has a very active compliance department. But as far as I can tell, they don’t have a source mode editing function.
We thought about it early on but most browsers have tools to do this, so we didn’t see the point.
Yes, it’s supported.
I’ll just add a few thoughts and a little history on top of the excellent technical details Matt added above. When we first approached v5 and the need for a new editor given that CKEditor 4 was end of life, no longer open source and only supported with a commercial license, we naturally went with CKEditor 5. What happened with CKEditor 5?Well, it wasn’t great. CKEditor 5, like TipTap, and all other modern editors had a very different way of parsing content. Instead of treating everything as strings, it wants to create nodes for each tag, which meant that HTML editing would be problematic. Matt spend a few months working on it and we got it into a finished state but there were a lot of issues. The first was that there was no easy way to add custom buttons without recompiling the JS, which made it impossible for a distributed piece of software. Secondly, the compiled JS was huge and it was taking 3-4 seconds for the editor to open. And finally, I became very uncomfortable with the direction of travel CKEditor was taking. They made it clear with recent blog posts that they only wanted open source projects to use CKEditor without a commercial license. Their open source version switched license types and forced a “Powered by CKEditor” icon on every editor at the bottom of the editing window. To remove this, and to use it in a commercial product you’d need a commercial license. I consider this fair. After all, the CKEditor developers should get paid for their work in paid software and clearly their current model of providing an open source product with an optional license wasn’t generated enough revenue. We contacted them and enquired about a license that we could distribute, that did not use their cloud deployment/CDN and did not use any of their advanced plugins and we were quoted $35,000/year with basic support of a maximum of 24 support tickets a year. So we were left with a bloated and slow implementation that could not allow for custom buttons that cost $35,000 every year with the risk of that price increasing year on year. Even though we had it finished and working in v5, we knew we had to make a tough decision. Enter TipTapWe did a lot of research into various editors. From fully functional editors such as TinyMCE to more raw editors such as Quill and just about everything in between. We wanted the best balance of ease of development, good support and speed. TipTap fit all those requirements when others did not. We bought an enterprise license which gives us access to their developers via Slack which has been really helpful. The cost of this license is about a third of CKEditor 5. So what about HTML editingLike CKEditor 5, and all modern editors, doesn’t think of input as merely a string to parse into blocks, but it looks at the document (the editor contents) as a tree of nodes. Each node is handled in a specific way, so you need to ensure all possible HTML tags are handled via extensions. This is great because it’s faster, more secure and arguably easier to write extensions to handle custom buttons and plugins but it also means the editor has no concept of ‘source mode’. There is no source, it’s an internal array of items. It is a common request but there is no clean solution at this time and none of the suggested workarounds actually work reliably. However, we also felt it was important to take a step back and, as Matt said above, not just give up and add in a text area instead of an editor when you want ‘HTML editing’ because for one, most people do not not to compose a complicated post in a text area manually entering HTML tags. They want to edit the already created HTML in the editor, which isn’t easy. Clicking edit would destroy the contents as TipTap has no way to create a document tree for random HTML pasted in. Part of this solution was to embrace the ease of which TipTap can add extensions. It is the reason we created the editor permissions framework to allow you (the admin) to have specific functionality that others do not get. It’s important to remember that v5.0.0 is the very start of the future for us. We have lots of plans to add more extensions, such as tables in minor point releases but we have to draw a line and say “this is fine for the first version”. If you are using source mode heavily and pasting in JS and tracking code, then let us know the specifics. I bet there’s an easy way to do it in v5 that doesn’t rely on your hand editing posts which is time consuming. It’s also likely that you won’t upgrade to v5 right away and watch how it develops as v4 will still be supported with bug fixes. Hopefully you can see that TipTap is not just a second choice because CKEditor was expensive, it’s actually far superior in every way but also we didn’t just decide to remove HTML editing on a whim. Let’s work together to build proper tools and not just give up and throw in a rough workaround.
Yes, and yes. We haven’t done a huge amount of v4 upgrade testing, so maybe don’t do that on your live site.
I addressed this question mark over how many alphas we’ve released here: https://preview.invisionalpha5.com/topic/202-alpha-16-released-beta-next-week/ But either way, I don’t think basing your community’s future over the release cadence of our next generation community platform is the best metric.
Welcome to the final Alpha release of Invision Community 5! It’s the final alpha…Yes, that time has finally come. Two alphas after the last time I said this. But this time I mean it! If you want to know what was changed, the release notes are here. We’ve now moved the Bug Tracker to our support forums ahead of the beta, so please report any issues there. The hold up was due to an optimisation branch to reduce what was a silly amount of queries per page. We wanted this to go into an alpha as there’s a lot of changes, and we didn’t want the first beta to be obviously broken. It doesn’t set the right vibe. 🫠 Why so many alphas?I’ve seen this question pop up a lot, and it’s a valid question. We’re on Alpha 16. That’s a pretty big number. We started releasing alphas 14 weeks ago, so we’ve averaged just over 1 release a week. Does this mean that it’s been a really bad process and something has gone terribly wrong? Absolutely not! Version 5 has a brand new UI, brand new CSS, loads of new and rewritten JS, loads of new features, a new editor, a new page builder, and so on. That’s a lot to not only test to make sure there are no bugs, but also in terms of feedback from testers. We’ve made a lot of changes over the past 14 weeks based on that feedback. We also decided to make sure we kept releasing alphas as often as we could. We didn’t want to wait 3-4 weeks before each release. When we move to beta, it is to signal the end of new features, large refactors and any potential u-turns or revisions to existing new features. We want beta to just be about bugs, bugs, bugs and not introducing new ones by adding surprise new features. In terms of stability, I am completely confident that you could use this alpha on your production community and be totally fine. I know this because we’re using it on a few private sites internally. You may consider an alpha to be a raw and largely untested piece of software, but with Invision Community 5 it is not the case. It’s very polished and very well tested already so we’re hoping for a very short beta testing phase. It’s more akin to an early ‘release candidate’ than rough and ready product to be beaten into shape. We could have released 6 alphas and 8 betas, the labels aren’t that important. Now is the time to move into beta though as all the heavy lifting and major testing has been completed. I’ll do a wrap up blog post on our main site to review a few newish features that we haven’t announced yet (but you’ve been using here for ages now). Thanks for your patience. 👏
Welcome to Beta.. uh.. Alpha 15! We were planning on Beta 1 this week, but we’re working on a few branches that aren’t quite ready, including a lovely optimisation branch to make sure v5 is ready to go but there were a few bugs that were being reported multiple times, including weird Chrome/Edge bugs so, welcome to Alpha 15! Out now in all good stores CloudLand®
Joking aside, XenForo is fine software. Choosing a platform is a very personal decision based on what you need for your community. What is your community based around?
Guess what? Alpha 14 is on it’s way through CloudLand® If you’re curious as to what has been fixed, check out the release notes. Will there be another alpha? Or will it be beta? Find out next week! Or maybe in an hour when I release another alpha as this one broke something important.
Thanks all! We knew that we wanted a longer alpha testing cycle because we figured that the more stable we get it now, the quicker the later phases go. We also had a lot of new stuff to bed in, which took a few months but when we do get to the public alpha, I’m pretty confident that beyond self-hosting installation issues which will take time to get through, we’ll have a stable bit of software. I’m also pretty sure that 5.0.0 Alpha 13 is more stable than 4.0.0 Final. 😅 We also wanted to test a lot with some of our private sites to make sure that the Pages changes weren’t going to affect the ability to do all the things you need to do, and it’d kind of suck to do a u-turn a few betas in whereas the alpha stage is the perfect time to evaluate this kind of thing.
Yes, our current thinking is to release beta for both cloud customers and self hosted at the same time and just bathe in the chaos.
We are almost done with the alpha stage, and we’ll move into beta very soon. How many more alpha releases do you think we’ll have?
