MODX you are not testing enough! MODX 2.3 ThemePackerComponent

MODX I love you, and like the fact that we are evolving into something bigger and better fast!
But... We are moving too fast!

I updated some client sites to MODX 2.3 and find a lot of CRITICAL bugs:

• images in backend broken
• long TV names break layout
• password reset not working
• scripts return a 1 if there is no actual return statement
• many extra's don't work

MODX 2.3 is simply NOT ready to be LIVE!
https://github.com/modxcms/revolution/issues?labels=priority-1-urgent&page=1&state=open
https://github.com/modxcms/revolution/issues?labels=priority-1-urgent&page=1&state=closed
Why did you launch? This is very very bad for MODX I think and will creates a lot of frustration.

It is great to see how fast issues are being fixed, but MODX 2.3.0 is not good to go right now.


Same with the whole ThemeForest / ThemePackagerComponent thing,
It has SO MANY bug's and any MODX theme created with it is simply broken!
https://github.com/modxcms/ThemePackagerComponent/issues?state=open
No uninstalling, updating destrois the site and many more bugs....

Again why launch this thing if it is not tested?


We really need to test more and think before we go live with new stuff....

Given that 2.3 was in the making for what, 2 years, I really don't think 2.3 has been moving too fast. In fact, I'd argue it's been too slow.

There are always issues that only show their ugly faces after a release and people really start working with things in production rather than quickly testing things in a controlled dev environment. A brand new design and a lot of improvements under the hood are simply going to result in bugs to be found as people start using it. No amount of testing in dev environments can beat actually using a product from day to day.

Also, the issues you mention have all been fixed in the source and are due to be released as 2.3.1 very soon (this week, possibly today). Or is that too fast?

As for many extras don't work? None of the extras I personally maintain have so far needed an update to fix it in 2.3; the only things I've released updates for were design tweaks to make them look better in 2.3. 2.3 is for 99% a backwards compatible release with 2.2. There are a couple extras that used the 1% that did change and as far as I can tell, the developers affected have been busy releasing updates for that.

There's no denying that some bigger extras have had issues (I'd say primarily Articles and Babel judging from the discussions) which doesn't really say anything about the core. Extras as big as those need someone to lead it, and it seems both of those extras have been largely abandoned and don't have people actively working on it.

As stated... This was a bit of a rant (after some frustration with live sites of clients)

I might have had bad luck, all my client sites that I updated used Articles and Babel, and all of them wanted to edit images and could not after updating...
Also on all sites I had the number 1 appear suddenly. A small fix but clients do not like it.

Of course you are right and I am really happy with 2.3, but it is just frustrating to see big issues popping up so fast on live sites...

I have to thank Mark for tipping me off to stay away from the 2.3.0 release earlier. I'm on some dev builds that seem to be working well (July 2 build I think)

Let's remember any new release should not be considered production ready, unless you're prepared to roll back if needed (should be anyway). Even OSX 10.8, 10.9 etc.. all have bugs some worse than others after the first release, most people I know wait until the .1 release before upgrading. I like to live on the bleeding edge but I know the risks.

I would agree that this does look bad for MODx, with all the new attention with this release.. let's get 2.3.1 out the door! (done right of course)

I (and most of the people) like the 2.3 release
It's normal to have a lot of bugs, but that's why the forums and the issues at github are. That's why all we are testing and reporting the bugs.
I think everyone MUST make a backup before he/she decide to update to 2.3 version.
If there are a lot of problems after the update then you can go back to 2.2.x version until the issues are fixed.
This is the best way to make MODx better

We've had the modmore site on the latest nightlies for some time, updating regularly with no issues to speak off. Unfortunately building the final 2.3.0 released caused some issues which surfaced the days after the release. I don't think anything could have been done differently to prevent that, but indeed if you worry about production sites it's always safer to wait for a .1.

Personally I like building from git directly so I can always get the very latest bleeding edge

Ok... I calmed down a little and no big harm done....

I know 2.3 has been a huge effort and it is great work Thanks to all and I will continue to do my part.
Glad I had a change to rant a bit, feel a lot better now!

Menno, I hear your frustration. Of course, no one wants to release buggy software and no one wants an upgrade to cause any headaches. For that, I feel badly and sympathize with you.

This is my personal opinion. I do not believe time was the issue with this release. Waiting longer before releasing would not have ensured the valid bugs/issues you encountered would have been caught. There are just not enough people testing development builds on production sites or clones of production sites. Likely, all of these bugs would have been found if 2.3 were tested against real-world sites. Most testing is done in laboratory conditions where we test to make sure that the changes to the core work as expected and do not break functionality in generally empty builds. The only way to ensure releases are as bug free as possible is for development builds to be tested on real sites by as many people as possible. Perhaps our failing is not ensuring enough people knew about it and asking for help from the community.

That being said, we did our best to suggest people check the list of Extras and it's compatibility before upgrading to 2.3.

2.3.1 is expected in the next few days. We have resolved all of the bugs you mention. As we look toward 2.4, we need to get more involvement from people to help test each build and especially after the feature freeze is announced.

On the Extras compatibility. We could probably do a better job of getting the word out about pre-release testing. However, we cannot wait for a release because an Extra has not been updated by the author or abandoned by the author. MODX the company, can't realistically or sustainably adopt all Extras no matter how widely used they are. Though in the case of Articles, we would have stepped up had others like Mark and JP not done so. This is why it's so important Extras aren't built relying on undocumented quirks or bypassing the parser or other native functionality. In theory, following Semantic Versioning there should be no reason that such Extras would fail on a minor version upgrade.

In 2.3 there were bugs introduced that broke extras that has nothing to do with poor Extra architecture. The phpThumb library that is called in Revo has a bug, as does the return value issue.

I think everyone who uses MODX would like to see each release go more smoothly. Lets turn this into a discussion about how to make the release of 2.4 and beyond, better and more predictable.

Thanks Jay,

Totally get it. Great info.

Huh I just used your social sharing info for my share buttons, thanks for that! Plus, I am absolutely pro-ranting!