Proposal for dropping ie11 support in Vue 3 #296
Replies: 45 comments 55 replies
-
|
I think an estimate on 2.7 release date would be very helpful for those who plan to support IE. |
Beta Was this translation helpful? Give feedback.
-
|
No estimates can be made at this point because the scope of features to be backported are not finalized yet. If only targeting a minimal feature set aimed at bringing Composition API into Vue 2 core, then it will probably be around end of Q3 2021. |
Beta Was this translation helpful? Give feedback.
-
|
How can community help with the 2.7 release to make it happen sooner? |
Beta Was this translation helpful? Give feedback.
-
|
There are some initial steps that we need to figure out (i.e. the best way to merge |
Beta Was this translation helpful? Give feedback.
-
|
|
Beta Was this translation helpful? Give feedback.
-
|
Meet the future reject IE |
Beta Was this translation helpful? Give feedback.
-
|
it's true to drop IE! |
Beta Was this translation helpful? Give feedback.
This comment has been hidden.
This comment has been hidden.
-
|
Not having to pollute their codebase for a browser thats on it's last legs? And a browser which doesnt support the Vue 3 reactivity system out of the box anyway |
Beta Was this translation helpful? Give feedback.
-
|
We wrote a Motivations section here |
Beta Was this translation helpful? Give feedback.
-
|
Please read the full RFC. |
Beta Was this translation helpful? Give feedback.
-
|
I think this is a great idea. We're stuck on Vue 2 not because of IE11, but because of waiting on the ecosystem to catch-up. As a stop-gap we've been using I have two questions:
|
Beta Was this translation helpful? Give feedback.
-
|
Merging |
Beta Was this translation helpful? Give feedback.
-
It is feature complete, however porting it directly into Vue 2 core will enable less code (direct access to internals removes some hacks needed), better performance, and making writing Composition API code that works in both Vue 2/3 a lot easier since now
Yes, that's another part of the goal for 2.7 (not explicitly mentioned in this RFC). There can be many minor API alignments (e.g. expose the same
Technically composables should work the same way in Vue 2/3, since Composition API behavior is expected to be consistent. The reactivity language coverage in Vue 3 is a superset of Vue 2's, so any composables that works in Vue 2 should theoretically just work in Vue 3. (the only thing you probably need to change is where you import the methods from). |
Beta Was this translation helpful? Give feedback.
-
|
Thanks for that feedback, very happy to hear, especially that you're thinking about not just this step but the migration step onto Vue 3
The list of limitations in the current vue 2 composition api requires some less than ideal workarounds that wont be required in Vue 3. Will any of these limitations be fixable by access to Vue 2's internals? |
Beta Was this translation helpful? Give feedback.
-
|
I doubt it, most them are due to ES5 language constraints. Maybe some. We'll see as we work on it. |
Beta Was this translation helpful? Give feedback.
-
|
I like that! |
Beta Was this translation helpful? Give feedback.
-
|
We think backporting stuff to Vue 2 will be much easier than supporting IE11 in Vue 3. Worth it for those that still need IE11 support for their Vue apps! |
Beta Was this translation helpful? Give feedback.
-
|
That's fair! So it raises another question: |
Beta Was this translation helpful? Give feedback.
-
|
The overall effort shouldn't be huge, since majority of Composition API for Vue 2 is already implemented and validated in This effort is targeting users who have non-trivial codebases that are stuck on Vue 2 due to various reasons (IE11, dependencies, no bandwidth for upgrading). We are pretty sure there are non-trivial amount of existing apps that fall in this group. Our one-time effort could probably benefit many users in such scenarios, saving them the cost of migration altogether. |
Beta Was this translation helpful? Give feedback.
-
|
Glad to hear it. Last question before I go :) |
Beta Was this translation helpful? Give feedback.
-
|
The work on 2.7 is planned to happen after an effort to finalize recommending Vue 3 as the default (docs, npm latest tags, etc.). So the goal is for 2.7 to serve as (1) an option for those who absolutely need to support IE and (2) a version that provides enough Vue 3 compatibility to ease the migration efforts for those who want to upgrade to Vue 3. |
Beta Was this translation helpful? Give feedback.
-
|
When was the vue 2 end of life. |
Beta Was this translation helpful? Give feedback.
-
|
For bug fixes, 18 months after the release of Vue 2.7. Some security fixes may be release after that. |
Beta Was this translation helpful? Give feedback.
-
|
I'm a developer at the Wikimedia Foundation, which made the decision to adopt Vue as our framework of choice a little over a year ago. Supporting legacy browsers like IE11 has been a requirement for us since forever basically, although this is starting to change. We have several Vue-powered features in production that are still expected to support IE11 for some time. I think this proposal will actually make things easier for us. If a 2.7 release of Vue is able to back-port some of the main benefits of Vue 3 (Composition API, better typings, etc) while retaining IE11 support, there will be less pressure for us to migrate to Vue 3 right away. I think this would allow us to have a smoother migration path to Vue 3 when we are eventually able to leave IE11 behind us completely. Thanks for taking the need to support legacy browsers seriously! It was one of the main advantages of Vue in relation to other frameworks as we were evaluating. |
Beta Was this translation helpful? Give feedback.
-
|
I completely agree with this direction. I work for an enterprise medical company and all our apps have to support IE11 still. If I could use the core features of Vue 3 in Vue 2 and later transition to Vue 3 when we move to Chromium Edge in 2 years that would be perfect. |
Beta Was this translation helpful? Give feedback.
-
|
Will Vue2.7 also support component API and TypeScript in the future? |
Beta Was this translation helpful? Give feedback.
-
|
Yes, that's in the RFC |
Beta Was this translation helpful? Give feedback.
-
|
Drop IE11 support in Vue3, but please provide a fallback for the same (in-built, without a modernizr dependency) just like we do while using |
Beta Was this translation helpful? Give feedback.
-
|
The main issue is that Proxy API, widely used in the Vue 3 reactivity implementation, can't be correctly polyfilled in ES5 (see here). So, to support IE, Vue should support two different reactivity implementations:
To Vue core team, supporting IE means a lot of manual code and test writing (it's not just configuring Babel) and the burden of maintaining two different implementations. To Vue 3 users, supporting IE means you're either actively targetting one subset of Vue reactivity (and using |
Beta Was this translation helpful? Give feedback.
-
|
@leopiccionia Thanks for making this easy to understand. But is creating a feature detection out-of-the-box for any unsupported browser hard? I just wanted to have the availability of some fallback code block where we are at least try to tell the user that their browser is unsupported, instead of just showing a broken/partially-working website on all the unsupported websites. I understand there are a lot of other ways to do this on the client end as well as on SSR as well. But getting something out-of-the-box solves a need for some extra dependency library/manual checks. |
Beta Was this translation helpful? Give feedback.
-
|
It would be nice if the summary at the top of this page could be updated to have a nested list of all the Vue 3 features, and their liklihood of being backported. Documenting this at the top would be very helpful. Otherwise most of these comments will just be asking about specific features. I know that the "Full Rendered Proposal" has a list at the bottom, but no reason to hide the information behind an extra click. |
Beta Was this translation helpful? Give feedback.
-
|
I don't think it is a good idea. IE on desktop has 1.7% usage worldwide and 2.2% in North America. This is millions of users in total, many of which cannot upgrade to a newer web browser due to stupid policies or unsupported configurations. Don't break the web for them! It is too early to drop IE11, let's wait until its global usage on desktop is less than <1%, in time of Vue.js 4. In comparison, jQuery and React still support IE9+ and Angular supports IE11. |
Beta Was this translation helpful? Give feedback.
-
|
The problem is not in decision to drop support for IE11. Problem is that at the beginning after Vue 3 was released, there was promises that support of IE11 will be in Qx of 2020. I guess it's ok it was delayed, but say: "We are dropping IE11 support" out of blue in beginning of Q2 2021 is just late. I started to developing new app in Vue 3 and now I will have to rewrite code back to Vue 2. So once again: It's OK to decide that Vue 3 will not support legacy browsers. However make this decision in time of releasing version, not a half year after. |
Beta Was this translation helpful? Give feedback.
-
|
@kuba108 You would have rewritten code anyway since not all of it would have been compatible with IE11, since Vue 3 would require a different reactivity implementation that works in IE11. |
Beta Was this translation helpful? Give feedback.
-
|
@Akryum If it is so clear now, why it was not clear in September 2020? I thought polyfills will build compatible version with IE and when IE will be gone in about a year from now, I will change polyfill list and voila! It's very unusual to support and plan development of two major versions at the same time. It reminds me Python and the problem with version 2 and 3. Many of my friends were unable to switch from 2 to 3 over 8 years. |
Beta Was this translation helpful? Give feedback.
-
That is a false statement. It was always planned to have a compatibility build with an alternate reactivity system with the same caveats as Vue 2. So not all code made with Vue 3 standard would have worked, that's why we also planned to an hybrid development version to warn for those cases. Some standard 6 years-old JS features are not supported in IE11 and not polyfillable, like Proxy.
We are not though, Vue 2 is already developped. We are continuing supporting it though. The goal is to ease migration of existing projects to Vue 3 and provide an escape hatch for IE11 support.
??? |
Beta Was this translation helpful? Give feedback.
-
|
The surport IE is a must in china if you want to make a enterpirse oriented application...saddly.. |
Beta Was this translation helpful? Give feedback.
-
|
I work for a site with ~ 800 million impressions and ~ 15 million unique clients per month in Austria. We see a clear end for IE support and only provide a basic user experience for it. We recently launched a prominent project without initial IE11 support and had not one hateful mail in our inbox. Since our projects have longer release cycles regarding underlying frameworks and libraries, I don't see a big problem. Existing applications would require a lot of work to update everything to Vue 3, so I'm sure most will stay on Vue 2 till their own EOL. They will keep their existing IE11 support, developer experience is still acceptable. Regarding the RFC's argument for IE11 for screen reader users: I see the problem, but those users will face a lot of troubles in the coming months, since IE support will fade out for a lot of popular products. If they need IE11 for internal applications, those will not jump on the Vue 3 train either. So for me this should have no effect on a decision, since internal technical debts should stay interal 😃 |
Beta Was this translation helpful? Give feedback.
-
|
IE is usually not a problem for the general public, it's only a problem for legacy entreprise systems, some even still using windows XP that have no other choice than to stay on IE |
Beta Was this translation helpful? Give feedback.
-
|
Drupal core is dropping support for Internet Explorer 11 in Drupal 10 Excerpts from the link above:
You can also read the discussion that led to that decision |
Beta Was this translation helpful? Give feedback.
-
|
No problem at all for dropping support for IE11, what I'm concerned about tho is the safari explorer, because that browser is gonna be the next internet explorer |
Beta Was this translation helpful? Give feedback.
-
|
I felt this way until recently. Apple scooped up several Firefox devs when Mozilla nearly died last year, and their new focus on privacy makes me pretty optimistic about the browser's future. Though as much as current Safari is looking good, I think the primary issue with web support is now deciding whether to exclude users who are locked to certain older versions of Safari because they are on older mobile hardware many of whom legitimately cannot afford to upgrade (like iPhone 6 which tops out at Safari 12.5.2) |
Beta Was this translation helpful? Give feedback.
-
|
One simple question I have that I didn’t see an answer to in the proposal: does this mean Vue 2 will have an effective life until 2023? (Edit: I had this date wrong) That’s Microsoft’s own estimated timeline for IE11 to finally be rendered into absolute insignificance that we truly never have to think about it again. This is due to be the year that the last of all support contracts that include OS versions that only shipped with IE11 ends for them (that’s a pretty significant signal as to the life of their EOL platforms) Of course by that time we will all be complaining about supporting Edge 15 😂 |
Beta Was this translation helpful? Give feedback.
-
Beta Was this translation helpful? Give feedback.
-
The thing is, even if you had Vue 3 with an ES property-based reactivity module in place of the proxy one, those shiny Vue 3 plugins would most likely still not work for you. Those plugin devs that dropped Vue 2 won't restrain themselves to use IE-11 compatible reactivity API only. The fact that proxies can't be emulated in older browser is the deal breaker. IE11-compatible-Vue3, if it existed, would not be the same framework as Vue 3 and most libraries/plugins written for Vue 3 wouldn't work in IE 11.
Which is still 0.5%, even if the absolute number looks bigger. |
Beta Was this translation helpful? Give feedback.
-
That's understandable. For most websites, supporting IE is not a necessity and due to the technical limitations I can definitely see why dropping IE will probably be the best course of action in this scenario. Unfortunately, this means I'll have to look elsewhere for another framework/library that will fit our needs. Thanks for taking your time to reply. |
Beta Was this translation helpful? Give feedback.
-
|
Confused why you need to move to a different framework? They said they're going to maintain Vue 2.7 which has IE 11 support. If ecosystem size/support is your primary issue, your options for another framework is quite limited. I can't imagine the other major frameworks supporting IE11 in the next few years either. The Vue ecosystem primarily supports Vue 2 right now, but by the time it's predominantly Vue 3 in a year or two, you might not have the IE requirement. Even if you do, I think it will take less work to fork the necessary plugins and backport patches/features yourself. |
Beta Was this translation helpful? Give feedback.
-
Your numbers are off by a magnitude. 1% of 1.5 billion = 1500 million / 100 = 15 million. Quite the difference from 150 million.... |
Beta Was this translation helpful? Give feedback.
-
You’re absolutely right! My bad. Accidentally put 0.1 instead of 0.01 into the calculator in a rush. |
Beta Was this translation helpful? Give feedback.
-
|
I totally agree with this proposal. I think IE is dragging down in innovation and just makes the developer's life just plain frustrating. Can't remember the last project where I had to support IE 11. |
Beta Was this translation helpful? Give feedback.
-
|
I believe this is reasonable course of action. What I'm more afraid of is the how well Vue 2 will be maintained, and possibly for how long. I understand it's a big burden to maintain (and further develop) all the libraries you do, and you do it (sort of) for free, but the activity in the Vue 2 repo has been really slow since Vue 3 was released, as can easily be seen by the number of commits vs issue and PR trends. Issues such as vuejs/vue#11718 should have been fixed a long time ago. (I know vite and all that, but the fact is that webpack is the main bundler atm and vite is the new unknown kid on the block) Perhaps a promise to actively maintain Vue 2 until IE11 has less than 0.X% of global usage would help? And show some proper effort into Vue 2 until that day happens. |
Beta Was this translation helpful? Give feedback.
-
|
我也同意放弃IE，IE简直拖后腿。但还是希望以后有人可以出个vue3降级vue2的指南，我曾遇到一个项目，开发初期和甲方再三确认过不需要兼容IE，但在项目二期的时候他们又反悔了，鉴于种种原因我们没有拒绝的权力。 |
Beta Was this translation helpful? Give feedback.
-
|
赞同,Vue 3 不支持IE,无可厚非.但希望Vue 2的维护时间能长一些. |
Beta Was this translation helpful? Give feedback.
-
vue2 和 vue3 应该作为长期维护版本并行 |
Beta Was this translation helpful? Give feedback.
-
|
科技强国的底气[doge] |
Beta Was this translation helpful? Give feedback.
-
|
后面怎么解决的 |
Beta Was this translation helpful? Give feedback.
-
|
ios9 will also not be supported? |
Beta Was this translation helpful? Give feedback.
-
|
I’d assume so considering Proxy objects are not supported on that version. |
Beta Was this translation helpful? Give feedback.
-
|
It's been 10 days since the RFC reached final comments status. Any major holdup or particular reason it hasn't been merged? |
Beta Was this translation helpful? Give feedback.
-
|
Unfortunately, my organization will continue to use IE11 for now and in the next few years. What I am more concerned about is, will Vue 2x continue to add new features and bug fixes? Or will it stop adding new features after a period of time and only keep bugs Fix it? Thanks! 很遗憾,我所在的机构目前以及今后几年都将继续使用IE11.我比较关心的是,Vue 2x会一直持续添加新特性和Bug修复吗?还是会在一段时间后停止添加新特性仅保留Bug修复?谢谢! |
Beta Was this translation helpful? Give feedback.
-
|
FYI "RFC: Internet Explorer 11 support deprecation and removal" angular/angular#41840 |
Beta Was this translation helpful? Give feedback.
-
|
The vue-next readme should be updated accordingly ("Note: IE11 support is still pending.") |
Beta Was this translation helpful? Give feedback.
-
|
Considering there has been no major objection, merging. Regarding 2.x maintenance: we will also pick off some long standing 2.x bugs while working on 2.7, and 2.7 will be maintained for at least 18 months. |
Beta Was this translation helpful? Give feedback.
-
|
This RFC is now finalized. https://github.com/vuejs/rfcs/blob/master/active-rfcs/0038-vue3-ie11-support.md |
Beta Was this translation helpful? Give feedback.
-
Due to this big change, I assume a lot of organisations will be stuck on Vue 2 for a while to come. This means that we need to make libraries that support both vue 2 and vue 3 or maintain two branches for libraries. Both of those options aren't super clear to me at the moment. Are there any examples of libraries supporting both versions? I totally understand why IE11 isn't supported, but I fear a python 2/3 rift is easy to happen. The previous major version was in 2016 when Vue was much smaller, and there weren't as many libraries yet. |
Beta Was this translation helpful? Give feedback.
-
|
One option to support both is vue-demi. https://github.com/vueuse/vue-demi |
Beta Was this translation helpful? Give feedback.
-
|
Vue demi definitely helps! What I'm struggling with is deciding to use both babel transforms and have separate entry points for vue 2 & 3 (and what to do with files that use createElement instead of templates). I know this isn't really the right place to ask, but I'm wondering if there's any libraries that support both I can look at @ChrisShank? Thanks! |
Beta Was this translation helpful? Give feedback.
-
|
You should check out how |
Beta Was this translation helpful? Give feedback.
-
|
proxy attribute support This has nothing to do with IE，We need to support iOS 9 on our mobile devices. Too bad. Hope iOS9 and lower version Android devices will be phased out in a few years. |
Beta Was this translation helpful? Give feedback.
-
Why? |
Beta Was this translation helpful? Give feedback.
-
|
I want to say that the reason IE is not supported is because VUE3 uses ES6 features，Our team does more business on mobile terminal，Therefore, for our team, we will pay more attention to the lower version of mobile phones， |
Beta Was this translation helpful? Give feedback.
-
|
I'd say that the reasoning is the same between having to support IE11 and having to support other old browsers. The same cost applies. |
Beta Was this translation helpful? Give feedback.
-
|
Yes, I agree！It's okay to be radical |
Beta Was this translation helpful? Give feedback.
-
|
Microsoft announced today the official EOL of IE11 - June 22, 2022. |
Beta Was this translation helpful? Give feedback.
-
|
Fwiw, the link to the RFC has moved. (Well, it's no longer active, but this is the repo's archived copy.) Good read about IE11 considerations in general; thanks. Here are a few highlights from the RFC for TL;DC[lick]:
|
Beta Was this translation helpful? Give feedback.
and others
-
Status
This RFC has been finalized and merged.
Summary
Links
Beta Was this translation helpful? Give feedback.
All reactions