Thanks to visit codestin.com
Credit goes to github.com

Skip to content

Conversation

@huntc
Copy link
Contributor

@huntc huntc commented Apr 4, 2013

Replaces #898. Uses a new library at the typesafe hub. This library [1] provides pipelining support to Netty in the form of a handler.

The changes here constitute the changes required to Play in support of the new library. Some re-factoring of Play's upstream handler has been done in the process.

Given that the new library exists as a snapshot (at least for now), I've also enhanced the build to consider snapshots during the building of Play in snapshot form. Snapshots will not be resolved though when distributing Play - this is what we want of course.

[1] https://github.com/typesafehub/netty-http-pipelining
@Gissues:{"order":50,"status":"done"}

huntc added 4 commits April 2, 2013 17:40
…ecking whether the connection is still open before deciding to continue with the iteratee.
…sage.

Also added the typesafe snapshot repo into dependency resolution when we're building a snapshot release.
@jroper
Copy link
Member

jroper commented Apr 4, 2013

Missing a space? :)

@cloudbees-pull-request-builder

play2-master #997 FAILURE
Looks like there's a problem with this pull request

@ghost ghost assigned huntc Apr 4, 2013
@huntc
Copy link
Contributor Author

huntc commented Apr 4, 2013

The build is complaining about not being able to resolve Akka actors:

sbt.ResolveException: unresolved dependency: com.typesafe.akka#akka-actor;2.0.1: not found

I'm blowing away my local ivy cache and re-testing. All tests previously passed ok.

I shall report back.

@huntc
Copy link
Contributor Author

huntc commented Apr 4, 2013

I can reproduce the issue locally having blown away the ivy cache. I will fix it.

…understanding - and also to fix an issue where dependencies were not being resolved from a ivy2 layout (at least I think that was the issue).
@huntc
Copy link
Contributor Author

huntc commented Apr 4, 2013

I have now fixed up the dependency resolution issue. I have verified the fix by blowing away my local ivy cache and running the entire test suite.

@cloudbees-pull-request-builder

play2-master #999 SUCCESS
This pull request looks good

@jroper
Copy link
Member

jroper commented Apr 5, 2013

Looks fine to me.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I notice that there's an import for Channels._ for this file. I prefer Channels.close() over just close(), but if there are other usages in the same file you might want to check that their style is consistent with yours. Or, if there aren't other usages, the import can be removed.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Sorry, I see the import was removed. :)

@richdougherty
Copy link
Member

Looks good to me!

@huntc
Copy link
Contributor Author

huntc commented Apr 9, 2013

Thanks @jroper and @richdougherty!

@huntc
Copy link
Contributor Author

huntc commented Apr 11, 2013

This issue depends on issue #995 so that we can get the warm and fuzzies about our changes before merging the PR.

@huntc
Copy link
Contributor Author

huntc commented Apr 12, 2013

Now no longer waiting for #995 as we've hit a blocker. @jroper and @richdougherty have inspected the code and the customer that originally raised the issue has tested it and confirmed that all now appears well.

jroper added a commit that referenced this pull request Apr 12, 2013
@jroper jroper merged commit a09dfd3 into playframework:master Apr 12, 2013
@huntc huntc deleted the http-pipelining branch April 12, 2013 07:44
@cloudbees-pull-request-builder

play2-master-PRs #36 SUCCESS
This pull request looks good

@pk11
Copy link

pk11 commented Apr 12, 2013

this looks awesome! Wonder if we could also enable SPDY support[1] along similar lines? (unfortunately google does not seem to be too keen on turning on pipelining for chrome by default[2] which is a bummer)

[1] http://www.javacodegeeks.com/2012/05/netty-using-spdy-and-http-transparently.html
[2] https://insouciant.org/tech/status-of-http-pipelining-in-chromium/

@huntc
Copy link
Contributor Author

huntc commented Apr 12, 2013

Thanks @pk11! I think supporting SPDY makes sense and we should raise a ticket for it.

Btw http pipelining support was motivated by a customer requiring it, so there's definitely usage. I believe the Android world embraces http pipelining as well.

@pk11
Copy link

pk11 commented Apr 13, 2013

@huntc that's true. In fact the iOS app we are building will use this new HTTP pipelining feature as well (iOS is supporting pipelining as of iOS 5.0[1]) SPDY would be useful mostly with desktop Chrome/Firefox. Thanks again!

[1]http://www.guypo.com/mobile/ios5-top10-performance-changes/

@pk11
Copy link

pk11 commented Apr 13, 2013

(I created a ticket for SPDY support #1005)

@huntc
Copy link
Contributor Author

huntc commented Apr 13, 2013

Thanks Peter!

Sent from my iPhone

On 13/04/2013, at 10:12 AM, Peter Hausel [email protected] wrote:

(I created a ticket for SPDY support #1005)


Reply to this email directly or view it on GitHub.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

5 participants