@Mark Takata (Adobe) Is there a final word on whether the flag to allow unscoped variable access will still be supported in ACF 2025? Related, is there are target release date for ACF 2025?

Dave, do you have reason to think there was debate on that point? The technote for that March update made it clear that it would NOT be supported beyond cf2023. Or was there perhaps some public discussion (here or elsewhere) from Adobe suggesting they might revisit that decision? To be clear, it would only be ignored rather than "fail". With such jvm -D args, they merely set an env var for the jvm (or cf, running atop it), which can either be regarded by those or not. In other words, we don't need to REMOVE it (if they'd ignore it), though of course we might as well once they do. I realize your question is to Adobe. Just wanted to add this context if it may help others.

Thanks for jumping in Charlie. I was under the impression that a final decision about that hadn't been made, but I could be mistaken. I never thought we'd have to remove that jvm arg, just that at some point it wouldn't be honored any more, so unscoped variable access would crash. My question was whether that was actually going to happen, or more to the point, when will it happen. I totally get that it's a good idea, but for some of us with multiple big apps built over many years by many people, recoding all unscoped variable access seems like a big lift, and we have other pressing work.

Understood on all points, though let's clarify (for everyone following along, not just you) that it's NOT (not, not) that "all unscoped variables" must be addressed. It's only those which would be resolved by cf finding the variable in any of the several scopes listed in the March update technote--which are those scopes that could be manipulated from outside the request, such as form, url, cookie, cgi, etc. Also, I hope folks caught the news of the "patch" Adobe offered which helps SPECIFICALLY IDENTIFY cases where this is happening, as it causes cf to log (to a new unscoped.log) whenever a request hits code that resolves an unscoped variable in one of those scopes. It logs the cf template, the given variable, and the given scope (though not the line number), to help devs find and fix code that would be affected of that workaround was removed. BTW, the app-level workaround will also remain in cf2025. More on all this in my blog posts on the matter in March and July (highlighting the patch/logging), at carehart.org/blog.

So I gather you've confirmed that "the app-level workaround will also remain in cf2025", yes? This is new and welcome news.

The app level feature is not new but from cf2016 (it's just taken on renewed attention). As such, I've seen nothing suggesting THAT was ever even considered to be taken away.

Hmmm, maybe I wasn't parsing what you said carefully enough. Do you (or anyone) know definitively if the jvm arg to allow unscoped variable access will still be honored by CF 2025? Am I right that that's not "the app-level workaround" you were talking about?

First, i realize you are wanting to hear from Adobe with regard to the jvm arg. We'll let theirs be your "final answer". (Or we'll see when CF2025 comes out, or perhaps the pre-release--though I realize some may not "trust" what it shows, as a "final answer".) But as for our last 3 replies here, your first asked, "So I gather you've confirmed that "the app-level workaround will also _remain in cf2025", yes?_" I replied to that, saying that I expect that it (the ability to set searchimpicitscopes at the app level) WILL indeed remain available in cf2025. Do I know? No. I explained why I EXPECT it to. But then your response to that asks again about the jvm arg, and asks if this is or is not the "app-level workaround" I was talking about. It is not. The jvm arg affects the entire instance. The app-level arg affects only the app in which it's set. Maybe that's clear to you, but I'm also just trying to "parse what you've said". :-) Can you confirm if you there is still any confusion? I do realize your main focus is the JVM arg. I was pointing out the app-level setting as an alternative that can help you avoid the "big lift" you feared would come if/when the jvm arg stops being honored.

Thanks Charlie, apologies if I was being dense, I'm pretty certain I understand the situation now, as best we know it.

Indeed, "as best we know it". 🙂 Let's see how the eight-ball shakes out in the next couple of months with the release of CF2025 (or sooner, if someone from Adobe offers a definitive answer in advance).