HTTP/2 interop pains | Hacker News Reader

HTTP/2 interop pains

daniel.haxx.se

120 pointsrobin_reala27 comments

Is this a mistake or Google trying to force the standards process in their favour? Users seem to be blaming Firefox, not Google.

This doesn't drive the standards process in any particular direction, this is just a broken implementation that doesn't adhere to the current standard draft!

It could drive the standards process in a particular direction if it would compel Firefox to implement Google's HTTP2 vision, so that beta users can log into Google services again. Then, suddenly, two major browser vendors support a particular blend of HTTP2, significantly increasing the chance of that road being taken.

It sounds overly mischievous though (and I doubt Mozilla will bend), so my bet is that it's a mistake.

I'm reminded of this passage from Mark Pilgrim's Dive into HTML5

> But none of this answers the original question: why do we have an <img> element? Why not an <icon> element? Or an <include> element? Why not a hyperlink with an include attribute, or some combination of rel values? Why an <img> element? Quite simply, because Marc Andreessen shipped one, and shipping code wins.

> That’s not to say that all shipping code wins; after all, Andrew and Intermedia and HyTime shipped code too. Code is necessary but not sufficient for success. And I certainly don’t mean to say that shipping code before a standard will produce the best solution. Marc’s <img> element didn’t mandate a common graphics format; it didn’t define how text flowed around it; it didn’t support text alternatives or fallback content for older browsers. And 17 years later, we’re still struggling with content sniffing, and it’s still a source of crazy security vulnerabilities. And you can trace that all the way back, 17 years, through the Great Browser Wars, all the way back to February 25, 1993, when Marc Andreessen offhandedly remarked, “MIME, someday, maybe,” and then shipped his code anyway.

> The ones that win are the ones that ship.

I'm not convinced it's a mistake, and it wouldn't be the first time a browser maker tried to influence the direction of web.

It sounds like it doesn't work in Chrome either, so it's just broken.

Was it typical behavior of SPDY before to block the connection unless it could pass cookies through? If not, then it definitely feels like Google is trying to force cookie-usage through the protocol, and by shipping it like this, it forces browser vendors to adopt it as-is.

How do you imagine logging in without a cookie? And the browser in question supports cookies, it's just a bug.

This whole thread is just full of some weird anti-Google paranoia way beyond any healthy level of scepticism.

It is a mistake that Google rolls out a new protocol implementation without testing it in the two known major products that support it.

> Is this a mistake or Google trying to force the standards process in their favour? Users seem to be blaming Firefox, not Google.

Given Google's history around this whole HTTP/2.0 shenanigans, I see little reason to see anything in their favour.

They are probably gambling that Firefox will fold (like they forced them to with DRM) so that once again, they can have it their way internet-wide without having to wait for annoying things like feedback from others and standards-committees and such.

They will probably win.

Google directly funds the Mozilla foundation with about 300 million USD per year. If Google wanted to see firefox fall they have a very direct way of achieving it.

Not true - they don't pay this money out of charity. I am guessing other companies with interest in having a leading search engine (MS and Yahoo primarily, but also Apple, Amazon,...) would probably be happy to take Google's place if they backed out (maybe not for the same amount though).

This could be seen as a technique to undermine Firefox and get more users to Chrome, which would devalue Firefox's contribution to Google. As Google appears to embrace standards but continues to implement little flaws and Chrome-only behaviors (that is, extend), they will eventually deplete Firefox's market share to the point where such a monetary transfer is no longer beneficial (extinguish).

People always bring up the fact that there is some value currently provided by Google's payment to Firefox like there's no problem with the vast majority of the Mozilla Foundation's funding coming from a competing entity. It's still a dangerous situation, even if Google is in a position where they're willing to continue paying the "default search engine" toll for the time being. Even their willingness to do that is compassionate, because we all know that Firefox would only harm itself by switching to another default search provider.

Agreed, most of your points are valid and it is a precarious position for Mozilla. I just disagreed with the notion that their funding could be cut off this easily - even if Google retracted I'm sure they could reach some agreement with Facebook for example.

I still hope Mozilla is investing the money wisely, because this arrangement is temporary in any case. And I'm hopeful that FirefoxOS takes off!

Oh, come on, this is a pure conspiracy theory with zero logic behind it. The difference in Google's implementation is something almost certainly too small for Google to care about one way or the other, and easily accounted for by someone forgetting to update a small part of the code from draft 13 to draft 14. There is no reason to believe Firefox would "fold", rather than waiting for Google to fix it, and disabling HTTP2 in the meantime [possibly only for Google] if it hasn't been fixed by the time the stable release is supposed to occur. And Google has, in fact, confirmed that they will fix it:

https://bugzilla.mozilla.org/show_bug.cgi?id=1059074#c22

Google is currently working on a fix: https://bugzilla.mozilla.org/show_bug.cgi?id=1059074#c22

"The Firefox Nightly and beta versions have HTTP2" Seems an extremely long bow to draw when it is only affecting a tiny group of beta/nightly testers...

Beta is the next release, less than six weeks and ALL FF users will be affected unless Mozilla disables it.

This is a ridiculous claim. The spec had continued to change and it is entirely understandable that there will be bugs. This is what interop testing is for. I've worked with the Google folks on the httpbis group and they are committed to the standards process.

hurley got confirmation from Google that they're working on a fix: https://bugzilla.mozilla.org/show_bug.cgi?id=1059074#c22

SPDYCheck.org has shown that Google has been advertising support for the suspect h2-14 draft spec for several days now.

http://spdycheck.org/#google.com

From the list of protocols:

Success! SPDY is Enabled! Hurray, this website is using SPDY! The following protocols are supported:

spdy/5a1 h2-14 spdy/3.1 spdy/3 http/1.1

Does it work on Chrome (nightly or with that support enabled)? Does anybody disagree that this is a no brainer for Mozilla: they should work that around ASAP (with a HTTP/2 blacklist for example). The impact is so high and it affects too many users. Also note that Google in its deepest feelings wants FF down so this may by even intentional since people will always blame the client.

Why should Mozilla act? This only affects nightlies and beta, as the article stated. If you use one of those, you should expect things to occasionally not work.

When I asked my 11-year-old brother what browser he used (since I saw chrome and firefox on the desktop), he informed me that he used "Firefox Nightly" (which I hadn't noticed, but was also present), because "it's a little better". He had no idea what it was. I don't know how he found it in the first place.

> I don't know how he found it in the first place.

4chan, directly or indirectly. They have some crazy obsession with it.

Beta is the next release. It's about to ship in a week or so.

In no way does Google want Firefox down. Besides basically funding Firefox by sponsoring Mozilla, Chrome's official position is simply to drive forward web technology. The Chrome team may want to make their browser better than the competition, but they also want the competition to challenge, and help improve Chrome and the web.

This will simply boil down to someone not fully testing the implementation, especially since it's an extra, not a base feature. If this was an issue with Firefox stable, it would be more pronounced and worrying.

In no way does Google want Firefox down.

Of course they do. One less browser to support for all their properties, one annoying group in the standardization process less that keeps clamoring for annoying things like user privacy, proper spec documentation, independent implementations or god forbid, tried to oppose DRM.

If Google wouldn't want Firefox down, they wouldn't aggressively market Chrome on the search pages that Firefox users are served.

The "sponsoring" argument is totally lame. Google pays hard cash to shove their search machine down Firefox users' throat and collect the advertising dollars from it. That isn't charity.

Reply on news.ycombinator.com