I have a pretty good size web app developed in 2017R3 and finished over the last year or so in 2019R1.1 I am considering moving it up to the current version but am still not sure. I know this means going to Web 2.0, but even then not sure how much trouble I will be in!
A few questions:
Is Web 2.0 stable yet? I saw some nasty things about connection losses that do not come back again.
What other issues remain that are show stoppers?
Assuming there are issues, how far out does anyone know (or think) they might be before they are fixed?
Would it be better to wait for another 6 months or more before moving in this direction?
I’ve been building and deploying Web 2.0 projects since before day one. There are still issues to be sorted with the WebListbox, or if you use some third-party products, and a couple of other places, but it’s come a long way.
The key whenever I start moving a project from Web 1.0 to Web 2.0, for me, is psychological. I don’t think of it as “upgrading” or even “converting”. I know what the app needs to do, and I build it in Web 2.0 as though it were a new project as the two are dissimilar enough that the upgrade/convert mindset can feel frustrating due to the way some things are implemented (like the removal of some events to make the connection more robust).
In all, I almost always come out the other side of moving an app to Web 2.0 feeling better about the project at a base level.
Is Web 2.0 stable yet?
I would say yes, though there are things you’ll miss from Web 1.0 that require additional work if you absolutely depend on them.
I saw some nasty things about connection losses that do not come back again.
I’ve previously posted code to handle page reloading when the connection has been lost, and Sessions are handled a bit differently in Web 2.0. You can read more about that here: https://forum.xojo.com/t/session-interrupt-message/64489
What other issues remain that are show stoppers?
This will be subjective not only depending on the developer, but the project. I’d suggest you browse through Feedback and look for things that you might encounter which are unresolved.
Would it be better to wait for another 6 months or more before moving in this direction?
As I said before, I dove in early, but it’s nothing for me to throw together some JavaScript if I need to, or I may not run in to some issues because I use my GraffitiSuite components.
I spent nearly 3 weeks porting to 2020/21 from 2019 last December. I got it compiling. I reverted back to 2019. For a while I was maintaining both sources (e.g changes to one were made in the other), but unfortunately, that got impossible.
I test the app on new builds to see how it goes. WebContainerControls are still too slow and resizing is just a mess. WebListBoxes seem messy.
I ended up making my own CustomWebTextFields to accept the enter key. I had about 2k of them.
At this point, I am not even sure I can go to 2021 because I’ve lost track of all the changes I made to 2019 code.
So for me, it is absolutely scary. I have been using RB/Xojo since 1998(?). I’m not one for feelings, but this one really hurt my feelings and I legitimately stress about it every day.
I use a ton of WebContainerControls and WebListBoxes.
Those might be show stoppers - or at least wait for a while until they are more ready. Do you know if they are actively being worked on for improvement?
That is not an “upgrade”, you have to do a complete REWRITE of the app.
A LOT!!!
NO
As you have readed, it is still feature incomplete… Many events not implemented, there are no Contro sets, some of the components have many bugs and are not working as expected like the table, and others dont work reliable like the uploader that corrupts files.
At the current speed… A couple of years.
If you dont have show stoppers in web 1, stay there!!! Not wort it the work of a rewrite, including the little learning curve of the API 2 language changes.
Yeah, yeah, you can build a web ab with LOTS of work arounds and spending money on third party components. But only you can say if is worth your time, effort and money for almost no tangible improvement.
Yes, they are working on improving web 2.0, it is only a really slow process, and as allways, each release has new features but also new bugs, so never stand still waiting for a release, and never use a new release for production. You can encounter really nasty bugs with conde that worked before.
I made the leap around 4 months ago and have just transitioned what I would call a fairly complex project over to 2.0. A few little surprises along the way but using some third party controls (Anthony’s amazing Graffitisuite and MBS) made my project look 10000% better than where we were with 1.0. I server on a closed local network so can’t comment on web deployment, and there are a few little quirks that have popped up if you read my forum posts, but in the most part a great solution.
I would encourage you to wait. I see incremental (yet slow) optimizations and fixes with each release, but I would not convert just yet. It’s a painful task for large projects and not a process you will want to do twice. I still think that the slowness of WebContainerControls is absolutely unacceptable.
During the last months of 2020 and the first couple of months of 2021 I developed a Web 2.0 app. Being an early adopter at the time I found many bugs and had to work around them, with the help of Norman Palardy, Hal Gumbert (both sadly banned from this forum) and @Tim_Parnell I was able to get the app working properly and delivered. At times it got very frustrating but in the end the app worked as expected, I did use Anthony’s Graffitisuite and that helped a lot.
With the last releases Web 2.0 has become more usable out of the box but there are things that are missing or not working properly (weblistbox I’m looking at you). Currently I’m working on a new Web 2.0 project and things have been smoother than my first experience, it’s far from perfect but usable (depending on your needs of course).
I don’t excuse Xojo for the bugs and the workarounds we still have to use to get things working and seriously hope they can manage to improve Web 2.0 as it has great potential. As a one man show I really appreciate the way Xojo makes things easier for me to develop Web apps as I don’t have to mess with a backend language and HTML/CSS/JS for the front end (although if one is working with web there is no escaping having to use JS/CSS for some stuff). Xojo has a fantastic opportunity with Web 2.0 to make it a great tool. At this point it is useable but not without some level of pain and swearing involved.
So moving to Web 2.0 is doable, I did. I’m really hoping Xojo will fix the bugs and implement missing features that could make this a great tool.
But to answer your question more directly: It is scary but doable.
This forum is you best friend in moving to Web 2.0. Ask direct questions and you’ll get direct answers.
I have pushed Web 2.0 very hard and found many workarounds, but also many bugs with no workarounds. Slowly most of those bugs have been fixed, but there are still some occasional quicksand bogs.
On a positive note, I am loving the fact that my code is much cleaner and clearer across desktop, web and mobile with API 2.0. I would definitely jump into Web 2.0 now, as the final bugs will hopefully be gone by your release date, or else you can use MBS or GraffitiSuite to spakfill your final problem areas.
While Web 2.0 show much promise, and Xojo is working diligently to kill of some of bugs, There are some real insidious bugs that make working with a database very very dangerous for the integrity of your data.
These appear to be fixed in the next release
If you use the lost focus event for a Webtextfiled, your users input may be ignored. No indication error will be given (it might even take the first 1/2 of the word and ignore the second half) – this bag has been fixed and the current Beta
If you use a weblist box, there are cases when if you click on row 3 you will get the row tag for row 4. – xojo is aware of this bug and they’re fixing it diligently. I think I saw that it was marked fixed but it has not been released in the most recent beta build yet
None of these bugs can be trapped for.
If you have business critical data, you can’t use the current release.
Thanks again everyone! I appreciate the candor! @Jay Menna - my app is database intensive. So your warning is really a show stopper. Thanks for the info!
It’s a double edged sword. I hope Web 2.0 is kept in the highest priority because there’s no other RAD tools that compare. Not having control sets and a buggy listbox (The most important control IMO) feels like it was released way too soon.
