Should Express 5.0 require a minimum of Node.js 4?

[dougwilson]

This is a discussion regarding having Express 5.0 drop support for Node.js 0.10. One of the reasons for the delay in Express 5.0 was the announcement of the official merging of Node.js and io.js, and we were looking forward to it's outcome to determine what we should do with Express 5.0.

Here at Express, we value very long-lived major versions, allowing users to confidentially build their new server on Express and not have to worry about needing to upgrade across a major, breaking change for years. Express 4.0 has been out for over a year now and Express 3.0 was supported for over 3 years !!

When Express 5.0 comes out, Express 4.x line will continue to be fully supported, delivering our users bug fixes, new features, and more!

To this end, we have been carefully watching io.js/Node.js in relation to Express 5.0 such that we can continue this great support history without accidentally locking us into the past for a long time.

Please share your thoughts here.

[Fishrock123]

Fwiw 0.10 will be EOL'd October 2016, and 0.12 at the end of year, 2016.

The former is already in maintenance mode, and 0.12 will go into maintenance-only next April. 0.12's EOL was cut short due to OpenSSL ending support for 1.0.1 at end-of-year 2016.

More info: https://github.com/nodejs/LTS (Slightly outdated)

I'd personally vote for yes (biased I suppose), and try to encourage a larger move to 4.0.0 sooner than later for users.

[Twipped]

Dropping 0.10 for 5.0 sounds good to me. Express 4 isn't that different, right?

[alanpurple]

+1
drop support for 0.10

[hacksparrow]

Highly in favor of dropping support for 0.10.

[talon]

Open source should move as fast as possible w/ option for sponsored LTS. Drop it.

[aredridel]

What's the cost of not dropping 0.10?

[andineck]

Moving forward to node 4.0 seems more important than supporting 0.10

[dougwilson]

So, just to be clear: Express 5.x dropping support for Node.js 0.10 does not mean Express will drop support for Node.js 0.10. Just like Express 3.x has lived in parallel with Express 4.x for almost 2 years now, Express 4.x will continue to live alongside Express 5.x.

To answer @aredridel , not dropping 0.10 from Express 5.x means that 5.x would be hard-pressed to consume and support various features in the newer V8 and Node.js platform (things like Intl, generators, promises, and more). Because we want to make longtivity promises, this means that once Express 5.0 comes out, an Express 6.0 is unlikely to appear until 1 year or more later, putting it further behind the curve.

Here is a little mock-up support example:

Year             0      1      2      3      4      5      6

Express 3 (0.8 ) +--------------------+
Express 4 (0.10)               +--------------------+
Express 5 (4?  )                      +--------------------+

Typically we'll have 2 active Express majors at a time (more than that gets harder, and the benefit decreases). The idea is that we can then perhaps up the minimum Node.js version with a major, since the previous major will still be around for a long time to come after the release. If you look back at the history of the 3.x line while 4.x has existed, we continued to backport every single new feature from Express 4.x into 3.x that was possible with 3.x's design. We intend to do the same with 4.x/5.x, which means 4.x will keep up; the release of 5.x does not move 4.x into "maintenance mode", either.

[Twipped]

Perhaps a better question to ask is, what does Express gain by dropping 0.10 but not 0.12? Most of the improvements you mention cant adopted until 0.12 is abandoned.

[Fishrock123]

Perhaps a better question to ask is, what does Express gain by dropping 0.10 but not 0.12? Most of the improvements you mention cant adopted until 0.12 is abandoned.

Good point.

If we drop 0.10 we should probably also just drop 0.12.

Stats are that 0.12 doesn't have nearly the usage of 0.10, and is, quite frankly, rather buggy.

[dougwilson]

@ChiperSoft @Fishrock123 , oh, my bad. I do not like to think of 0.12 as even exiting, as there are so many issues with the V8 in that version, it's laughable that people can even use that version of Node.js without getting various illegal instruction errors from corrupted machine code generation.

Yes, this topic assumes that the minimum for Express 5 would be Node.js 4. I have updated the title to correspond :)

[aredridel]

Yeah. I've no real interest in generators, but promises I do -- that said, they work fine on 0.10, so that's not an issue.

Intl is near and dear to my heart, but I'd be intrigued to see what actual plans are before throwing out 0.10.

As it stands, if 0.10 is thrown out, that leaves basically 1.8.7, 2.3.0 and 3.3.0 as the only stable versions. 4.0.0 is not yet. It will be, and it's sure intended to be the future, but I really do value backward compatibility too.

[dougwilson]

Hi @aredridel , I understand the concerns. This is only a discussion at this point without any actual commitment. We are looking to understand what the value is, of course. As for promises, we won't be importing any third-party library to do promises, so we would have to look into how to do promises in 0.10 without a Promise global available.

When we implement any feature, we don't want it to be conditional on the version of Node.js you run on, but available on all supported Node.js versions of Express (thus allowing the user to know that running Express 5.2 on a supported Node..js version, they get all the promised features). Because of this, it makes dropping older versions more important, and hopefully should be possible since Express 4 will be fully supported, even adding all new possible features from 5 as they come out :)

If we decide to bump the minimum version of Express 5, yes, it would delay Express 5 a little longer to ensure that Node.js 4 is in a more stable condition when this occurs.

[aredridel]

Yeah. I'm just chiming in in the conversation.

What's the reason for not using an ES6-compatible shim for promises? It's usually a small dependency, and can be conditional on native not being available.

[Fishrock123]

I just want to note that native promises are still super slow.

[aredridel]

True that.

[dougwilson]

What's the reason for not using an ES6-compatible shim for promises? It's usually a small dependency, and can be conditional on native not being available.

We have tried that on a small scale in koa-related repositories, but it never worked very well, and always ended up better if we only used the global Promise function, which most have changed to doing now.

I just want to note that native promises are still super slow.

And, just to calm everyone down on promises, Express and it's dependencies will continue to exclusively use callback-based APIs. The adoption of promises would be in the public API (for those who choose it use it) as well as accepting promises, for those who choose to do so. As in, the current roadmap is a dual API externally, and a callback-based API internally. This should provide people with promises without any performance penalty for those who choose not to use them.

[aredridel]

Good good.

[Twipped]

Can you clarify what you mean by "use promises"? I was under the impression that just meant accepting a thenable as the return on middleware and routes. That shouldn't require any Promise lib, native or otherwise.

On Sep 14, 2015, at 8:39 AM, Douglas Christopher Wilson notifications@github.com wrote:

What's the reason for not using an ES6-compatible shim for promises? It's usually a small dependency, and can be conditional on native not being available.

We have tried that on a small scale in koa-related repositories, but it never worked very well, and always ended up better if we only used the global Promise function, which most have changed to doing now.

I just want to note that native promises are still super slow.

And, just to calm everyone down on promises, Express and it's dependencies will continue to exclusively use callback-based APIs. The adoption of promises would be in the public API (for those who choose it use it) as well as accepting promises, for those who choose to do so. As in, the current roadmap is a dual API externally, and a callback-based API internally. This should provide people with promises without any performance penalty for those who choose not to use them.

—
Reply to this email directly or view it on GitHub.

[Twipped]

Requiring Node 4 would be a pretty huge step up, but as long as Express 4 is still maintained I don't think it's a huge problem. Just need to make it very clear that E5 is a future forward version.

[ritch]

@dougwilson What if we adopted Node's LTS Plan? It might not be perfect (and IMO it is very complicated), but at least as a user I only would have to understand a single set of legacy support semantics.

Either way it would be nice to have some documentation (what you should expect) around long-lived major versions.

[OshotOkill]

Since Node 4 & ES6 finally came out I think the next main version of Express should eye for the future to be a new brand of itself, especially the main version have a long lifecycle.

Developers who want to persist in node v0.10 just only to uphold Express 4, and you have already said that would be also maintained after Express 5 realeased. In my opinion node 0.12 should also not to be considered much about compatible in Express 5, but I know that's not realistic.

[ivarprudnikov]

Correct me if I am wrong but compatibility seems to play a bit bigger role here, think third party modules for express. Compatibility would allow quicker uptake of express 5 if it supports node 0.12 for instance as it would not require and would not rely on developers fixing their modules/dependencies to make it compatible with new express. Or would not require searching for new ones.

How would developer know if a module that was used for express 4.0 or 3.0 will work for 5.0? Would I push code to production in this case? I do not know of any automatic tool that could tell me that (idea for greenkeeper.io devs)

[dougwilson]

Hi @ivarprudnikov , nothing in here would make third-party modules not work with Express 5.0. This is simply about a minimum Node.js version for Express. Can you explain how bumping the minimum version of Node.js for Express would break third-party modules for Express? Perhaps I'm missing something critical here?

[ivarprudnikov]

@dougwilson but if module is not compatible with Node > 0.12 ? And the fact that express would require higher (supported) version means there would be additional steps involved in checking if all dependencies work.
In the case of existing fleet of nodes, upgrade would require extensive QA.
In the case when someone creates new apps, one needs to sanity check if all deps (developer is used to) are compatible with higher node version which on the other hand is a positive exercise.

[dougwilson]

Hi @ivarprudnikov are you not able to do this within a timeline of 1 year? Express 4 is not going anywhere for a long time.

[dougwilson]

Hi all, sorry I am traveling today so my access is off and on. I read through all the comments and will definitely write up my thoughts when I'm at a computer soon :) The thoughts in #2755 (comment) from @wesleytodd most closely align to my thoughts and is nicely worded. I will provide further clarity on my thoughts later today when I don't have to type them on a phone, but was hoping this would at least help a little in the meantime.

[kibertoad]

@wesleytodd Btw, there was an idea voiced earlier about dynamic Express.js imports based on feature checks - then those who are using more modern Node.js and do not rely on direct imports would get the proper non-transpiled version. Do you see any problems with this plan?

[wesleytodd]

Do you see any problems with this plan?

Yes I do. It is more code to maintain, and unless we were exposing a different interface to the user, they wouldn't even get any benefit. So there is no upside to doing such.

[boneskull]

It's my hope the Express team will drop support for unmaintained versions of Node.js as soon as is feasible. Mocha has already dropped support for v0.10 and v0.12, and will drop support for v4 ASAP.

FWIW, "dropping support" does not imply "rewriting"; it can be as straightforward as removing the version from the matrix in .travis.yml.

[dougwilson]

We already confirmed we are dropping support for them in 5. What still needs to be clarified?

[boneskull]

@dougwilson Sorry, it sounded like it was somewhere short of "official"--I saw v0.10 and v0.12 were dropped already, but not v4. But if that's the plan, then great! 👍

[dougwilson]

When 5.0 is released it will support down to the LTS version at that time.

[jonchurch]

Superseeded by #5595

Node 18 will be minimum supported for v5