update ref to 1.0.0.final #267

smaldini · 2015-04-24T21:53:37Z

WIP for #223

viktorklang · 2015-04-25T18:21:42Z

Looks good so far :)

smaldini · 2015-04-25T20:09:03Z

Any feedback on the grammar, I think so far so good on this side but I feel a bit dodgy on this aspect to check the related box :)

viktorklang · 2015-04-25T20:13:59Z

@smaldini I've done a pass, but I'd like to have more eyes on it.

smaldini · 2015-04-25T20:18:40Z

maybe we can start the release of the binary and checks the impls too.

smaldini · 2015-04-28T15:01:59Z

Right let's start the testing process, should I publish final on codehaus before merging @viktorklang ?

viktorklang · 2015-04-28T15:15:08Z

@smaldini I'm not sure, assuming we already validated the integrity of the bits (perhaps do a RC5-final binary comparison before pushing?) it seems like we should be able to push it to maven central.
Wdyt?

smaldini · 2015-04-28T15:21:09Z

@viktorklang I'm usually risk averse and I would just merge and push to maven.

viktorklang · 2015-04-28T15:26:01Z

@smaldini Check the SHA of the RC5 jars and compare to the SHA of the final jars before pushing, and I'll sleep better :)

smaldini · 2015-04-28T15:26:29Z

I've rarely seen 1.0.0 that works for everyone anyway, thats why there is a minor digit hey 💃
But I will test SHA

smaldini · 2015-04-28T15:27:50Z

there's a bunch of TCK javadoc missing but well...

viktorklang · 2015-04-28T15:30:34Z

@smaldini documentation improvement is never done :)

smaldini · 2015-04-28T15:36:40Z

Strange , .asc and md5 are different from version to another.

smaldini · 2015-04-28T15:38:00Z

which is strange since I'm only one commit away. looking at the content

smaldini · 2015-04-28T15:38:46Z

right Manifest is regenerated

Export-Package: org.reactivestreams;version="1.0.0.RC5"
Bundle-SymbolicName: org.reactivestreams.reactive-streams
Bundle-Version: 1.0.0.RC5

viktorklang · 2015-04-28T15:41:31Z

@smaldini Alright, if that's the only diff then let's go :)

smaldini · 2015-04-28T16:00:19Z

It is on its way to maven central, joyful tear.

viktorklang · 2015-04-28T16:03:39Z

Awesome, @reactive-streams/contributors should update to it so we can
announce shortly.

Cheers,
√
On 28 Apr 2015 18:00, "Stephane Maldini" notifications@github.com wrote:

It is on its way to maven central, joyful tear.

—
Reply to this email directly or view it on GitHub
#267 (comment)
.

smaldini · 2015-04-28T16:14:32Z

http://repo1.maven.org/maven2/org/reactivestreams/reactive-streams/1.0.0.final/

viktorklang · 2015-04-28T23:29:56Z

@jsuereth and @rkuhn suspect that the ".final" part of the version number
may trip up version conflict resolution.
Since the ".final" serves no purpose, I propose we pull the "1.0.0.final"
and republish it as "1.0.0" while we still can. Is that OK, @smaldini?

Cheers,
√
On 28 Apr 2015 18:14, "Stephane Maldini" notifications@github.com wrote:

http://repo1.maven.org/maven2/org/reactivestreams/reactive-streams/1.0.0.final/

—
Reply to this email directly or view it on GitHub
#267 (comment)
.

ldaley · 2015-04-28T23:30:30Z

+1

rkuhn · 2015-04-29T07:08:13Z

@alkemist The problem would be that OSGi considers "1.0.0" < "1.0.0.RC5", how do we set a different version in the OSGi manifest, something like "1.0.0.final" is actually the right thing for OSGi.

ldaley · 2015-04-29T07:08:58Z

Still +1.

rkuhn · 2015-04-29T07:09:44Z

background on why using "1.0.0.final" for Maven is problematic: http://mojo.codehaus.org/versions-maven-plugin/version-rules.html (in short: "1.0.0.final" is interpreted as "0.0.0-1.0.0.final" and effectively only string comparison is used)

rkuhn · 2015-04-29T07:10:40Z

@alkemist I take it that your +1 is for the "1.0.0" maven version. In sbt I would know how to fix the OSGi version, how do we do this in gradle?

ldaley · 2015-04-29T07:11:28Z

FWIW, Ivy (therefore SBT) and Gradle understand that final > rc.

benjchristensen · 2015-04-29T07:11:35Z

I also think we should release 1.0.0 and then if OSGi is a mess release 1.0.1 as a patch release.

ldaley · 2015-04-29T07:12:13Z

@rkuhn what exactly do you want to do? Just use 1.0.0.final in the OSGi manifest?

rkuhn · 2015-04-29T07:31:54Z

Yes, that would solve all the issues AFAICS.

viktorklang · 2015-04-29T07:34:09Z

Solving all issues sounds good to me

Cheers,
√
On 29 Apr 2015 09:31, "Roland Kuhn" notifications@github.com wrote:

Yes, that would solve all the issues AFAICS.

—
Reply to this email directly or view it on GitHub
#267 (comment)
.

…e version

smaldini · 2015-04-29T09:32:29Z

Gradle OSGI plugin was already taking care of the bundle version but yeah making explicit is not harmful.

rkuhn · 2015-04-29T09:35:08Z

I tried it out with Viktor’s instructions locally and this was the resulting manifest:

rk:libs rkuhn$ cat META-INF/MANIFEST.MF 
Manifest-Version: 1.0
Bundle-Description: Reactive Streams API
Export-Package: org.reactivestreams;version="1.0.0",1.0.0.release
Bundle-SymbolicName: org.reactivestreams.reactive-streams
Bundle-Version: 1.0.0,1.0.0.release
Bundle-Name: reactive-streams
Bundle-ManifestVersion: 2
Bnd-LastModified: 1424093440000
Bundle-Vendor: Reactive Streams SIG
Bundle-DocURL: http://reactive-streams.org
Created-By: 1.8.0_31 (Oracle Corporation)
Tool: Bnd-2.1.0.20130426-122213

I tried to find information about what that version string means but that syntax is not mentioned anywhere in the freely available docs (I do not have an OSGi specification handy).

smaldini · 2015-04-29T09:36:01Z

Yeah Grade OSGI plugin uses the version you specify :)

ldaley · 2015-04-29T09:36:16Z

Ah, use instructionReplace instead of instruction.

smaldini · 2015-04-29T09:36:18Z

we don't need that extra line

smaldini · 2015-04-29T11:38:44Z

Is that correct enough for another try :) ?

Archive:  reactive-streams-1.0.0.jar
   creating: META-INF/
  inflating: META-INF/MANIFEST.MF    
   creating: org/
   creating: org/reactivestreams/
  inflating: org/reactivestreams/Processor.class  
  inflating: org/reactivestreams/Publisher.class  
  inflating: org/reactivestreams/Subscriber.class  
  inflating: org/reactivestreams/Subscription.class  

smaldini$ cat META-INF/MANIFEST.MF 

Manifest-Version: 1.0
Bundle-Description: Reactive Streams API
Export-Package: org.reactivestreams;version="1.0.0"
Bundle-SymbolicName: org.reactivestreams.reactive-streams
Bundle-Version: 1.0.0
Bundle-Name: reactive-streams
Bundle-ManifestVersion: 2
Bnd-LastModified: 1430236558000
Bundle-Vendor: Reactive Streams SIG
Bundle-DocURL: http://reactive-streams.org
Created-By: 1.8.0_20 (Oracle Corporation)
Tool: Bnd-2.1.0.20130426-122213

viktorklang · 2015-04-29T11:43:15Z

@smaldini no, "Bundle-Version: 1.0.0" is wrong according to for OSGi we need something like "Bundle-Version: 1.0.0.release" (@rkuhn can confirm)

smaldini · 2015-04-29T11:44:22Z

Who uses OSGI again ? sorry fixing that.

smaldini · 2015-04-29T11:46:20Z

  inflating: META-INF/MANIFEST.MF    
  inflating: org/reactivestreams/Processor.class  
  inflating: org/reactivestreams/Publisher.class  
  inflating: org/reactivestreams/Subscriber.class  
  inflating: org/reactivestreams/Subscription.class  

wifi-5-214:libs smaldini$ cat META-INF/MANIFEST.MF 

Manifest-Version: 1.0
Bundle-Description: Reactive Streams API
Export-Package: org.reactivestreams;version="1.0.0.release"
Bundle-SymbolicName: org.reactivestreams.reactive-streams
Bundle-Version: 1.0.0.release
Bundle-Name: reactive-streams
Bundle-ManifestVersion: 2
Bnd-LastModified: 1430236558000
Bundle-Vendor: Reactive Streams SIG
Bundle-DocURL: http://reactive-streams.org
Created-By: 1.8.0_20 (Oracle Corporation)
Tool: Bnd-2.1.0.20130426-122213

smaldini · 2015-04-29T11:46:32Z

now ?

viktorklang · 2015-04-29T11:56:11Z

I'm not sure about: "Export-Package: org.reactivestreams;version="1.0.0.release""
@rkuhn?

For the record, let it be known, OSGi is an instrument for inflicting pain.

smaldini · 2015-04-29T12:11:07Z

I'm going to make a slide about that springio.net tomorrow I think...

viktorklang · 2015-04-29T12:13:23Z

@smaldini a slide about what? OSGi?

smaldini · 2015-04-29T12:15:32Z

Yeah 👍 who cares about specs that are actually changing something in async processing without involving operating systems type containers.

rkuhn · 2015-04-29T12:34:25Z

@smaldini Your latest manifest is looking good!

viktorklang · 2015-04-29T13:18:47Z

@rkhun Just to make sure, is "1.0.0.release" considered to be newer than "1.0.0.RC5" by OSGi? (note difference in upper/lower case)

rkuhn · 2015-04-29T13:25:34Z

Yes, it is, see https://osgi.org/javadoc/r4v43/core/org/osgi/framework/Version.html (it uses String.compareTo).

viktorklang · 2015-04-29T13:29:38Z

@rkuhn Thanks!

Alright, so it looks like we have a solution!

let's switch to version 1.0.0 with OSGi version being 1.0.0.release in the repo
generate the jars
inspect the jars to make sure they are OK
publish 1.0.0
delete the 1.0.0.final tag from the reactive-streams-jvm repository

Sounds good?

smaldini · 2015-04-29T13:34:14Z

The tag is not here yet since the PR was not merged. Right now it can then :)

Sent from my iPhone

On 29 Apr 2015, at 2:29 pm, Viktor Klang (√) notifications@github.com wrote:

@rkuhn Thanks!

Alright, so it looks like we have a solution!

let's switch to version 1.0.0 with OSGi version being 1.0.0.release in the repo
generate the jars
inspect the jars to make sure they are OK
publish 1.0.0
delete the 1.0.0.final tag from the reactive-streams-jvm repository

Sounds good?

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

viktorklang · 2015-04-29T15:22:47Z

@rkuhn @smaldini @benjchristensen @alkemist and @reactive-streams/contributors in general.

I think we ought to merge this PR and publish the fixed 1.0.0 ASAP so we can get unstuck in the release process. @rkuhn, since you discovered the problem and found an appropriate fix, would you like to do the honors of cutting the release, or should our "release general" @smaldini do it?

rkuhn · 2015-04-29T16:08:20Z

Yes, I’ll start with the release now.

viktorklang · 2015-04-29T16:09:13Z

Godspeed

Cheers,
√
On 29 Apr 2015 18:08, "Roland Kuhn" notifications@github.com wrote:

Yes, I’ll start with the release now.

—
Reply to this email directly or view it on GitHub
#267 (comment)
.

update ref to 1.0.0.final

smaldini · 2015-04-29T16:19:48Z

Thanks @rkuhn! I was literally in flight from London Publisher to Barcelona Subscriber, fortunately flights are using backpressure too.

Sent from my iPhone

On 29 Apr 2015, at 5:22 pm, Viktor Klang (√) notifications@github.com wrote:

@rkuhn @smaldini @benjchristensen @alkemist and @reactive-streams/contributors in general.

I think we ought to merge this PR and publish the fixed 1.0.0 ASAP so we can get unstuck in the release process. @rkuhn, since you discovered the problem and found an appropriate fix, would you like to do the honors of cutting the release, or should our "release general" @smaldini do it?

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

viktorklang · 2015-04-29T18:47:59Z

@smaldini There is back pressure for sure, the longer the flight the worse back pain...

update ref to 1.0.0.final

f4fe7d7

change 1.0.0.final to 1.0.0 and make sure OSGI manifest has the bundl…

96f1254

…e version

OSGI fix

712411f

OSGI fix...

810adfe

rkuhn added a commit that referenced this pull request Apr 29, 2015

Merge pull request #267 from reactive-streams/1.0.0.final-wip

9f6d8a6

update ref to 1.0.0.final

rkuhn merged commit 9f6d8a6 into master Apr 29, 2015

rkuhn deleted the 1.0.0.final-wip branch April 29, 2015 16:09

update ref to 1.0.0.final #267

update ref to 1.0.0.final #267

Conversation

smaldini commented Apr 24, 2015

viktorklang commented Apr 25, 2015

smaldini commented Apr 25, 2015

viktorklang commented Apr 25, 2015

smaldini commented Apr 25, 2015

smaldini commented Apr 28, 2015

viktorklang commented Apr 28, 2015

smaldini commented Apr 28, 2015

viktorklang commented Apr 28, 2015

smaldini commented Apr 28, 2015

smaldini commented Apr 28, 2015

viktorklang commented Apr 28, 2015

smaldini commented Apr 28, 2015

smaldini commented Apr 28, 2015

smaldini commented Apr 28, 2015

viktorklang commented Apr 28, 2015

smaldini commented Apr 28, 2015

viktorklang commented Apr 28, 2015

smaldini commented Apr 28, 2015

viktorklang commented Apr 28, 2015

ldaley commented Apr 28, 2015 via email

rkuhn commented Apr 29, 2015

ldaley commented Apr 29, 2015

rkuhn commented Apr 29, 2015

rkuhn commented Apr 29, 2015

ldaley commented Apr 29, 2015

benjchristensen commented Apr 29, 2015

ldaley commented Apr 29, 2015

rkuhn commented Apr 29, 2015 via email

viktorklang commented Apr 29, 2015

smaldini commented Apr 29, 2015

rkuhn commented Apr 29, 2015

smaldini commented Apr 29, 2015

ldaley commented Apr 29, 2015

smaldini commented Apr 29, 2015

smaldini commented Apr 29, 2015

viktorklang commented Apr 29, 2015

smaldini commented Apr 29, 2015

smaldini commented Apr 29, 2015

smaldini commented Apr 29, 2015

viktorklang commented Apr 29, 2015

smaldini commented Apr 29, 2015

viktorklang commented Apr 29, 2015

smaldini commented Apr 29, 2015

rkuhn commented Apr 29, 2015

viktorklang commented Apr 29, 2015

rkuhn commented Apr 29, 2015

viktorklang commented Apr 29, 2015

smaldini commented Apr 29, 2015

viktorklang commented Apr 29, 2015

rkuhn commented Apr 29, 2015

viktorklang commented Apr 29, 2015

smaldini commented Apr 29, 2015

viktorklang commented Apr 29, 2015