Replying to Yishay and Alex:

Last time I took the release source bundle, changed the versions and ran a 
release build. Even if this worked, it was just a hack and technically we 
published release artifacts that are different than the version we voted on. 
So, someone who downloads the release archive will not be able to reproduce the 
build. I am not going to do this this time as I think this is not the Apache 
way.

The flex-typedefs is not a maven module from the point of view of the compiler 
build. So rat checks the content as part of the root modules check. It will 
check every file in that directory for license headers (even stuff in an 
eventually existing target directory). The exclusion instruction is in the 
pom.xml inside the flex-typedefs directory and isn’t acive when building the 
compiler (they are separate builds). One was to avoid this would be to exclude 
the flexjs-typedefs directory in the compilers rat-check (even if the orginal 
repo doesn’t contain that directory). The better solution would be to do 3 
releases and not integrate the typedefs in the compiler artifact (Which I think 
is the correct way to do it)

I think it is an issue … as a Maven artifact is the binaries plus the poms. 
Changes in the poms can greatly affect the functionality of the binaries. 

Chris


Am 12.06.17, 21:41 schrieb "Alex Harui" <aha...@adobe.com.INVALID>:

    I'm confused.  Can I get a summary?
    
    Are there some files that are being caught by RAT?  If so, what are they?
    
    Are we sure the process should be that the RM should switch away from
    SNAPSHOT before the vote?  If a major problem is found in that RC,
    wouldn't we have deployed bad artifacts under the final version number and
    have to pull them back?  Or abandon that release version and use the next
    version number?
    
    IMO, the main thing folks want from Maven are the JARs which aren't an
    official ASF release anyway.  Seems like we should vote on a source
    package, then set any version numbers and have Maven build the final jars
    from there.  The differences in the source should only be in POMs and
    other configs right?
    
    What am I missing?
    -Alex
    
    On 6/12/17, 3:53 AM, "Christofer Dutz" <christofer.d...@c-ware.de> wrote:
    
    >It should be between the Last call and opening the vote. It is equal to
    >“cutting the release candidate”.
    >
    >So, the LAST CALL thread is finished and the RM writes that he’s going to
    >cut a release … AFTER THAT he does these steps and THEN he opens the vote
    >thread. I never said anything else than that.
    >
    >Chris
    >
    >
    >Am 12.06.17, 12:30 schrieb "piotrz" <piotrzarzyck...@gmail.com>:
    >
    >    Chris,
    >    
    >    I'm a bit confused. You have said that I shouldn't do this as part of
    >    VOTING, LAST CALL:
    >    
    >    "No, 
    >    
    >    The removing the SNAPSHOT, tagging and setting the new version should
    >be,
    >    more or less, one step.
    >    "
    >    
    >    Now you are saying just opposite. So again when I should do this
    >(Last Call,
    >    Voting) step ?
    >    
    >    "1) In order to have a proper Maven release, the versions of the
    >maven build
    >    should be changed to “0.8.0” (omit the SNAPSHOT). "
    >    
    >    Piotr
    >    
    >    
    >    
    >    -----
    >    Apache Flex PMC
    >    piotrzarzyck...@gmail.com
    >    --
    >    View this message in context:
    >https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fapache-fle
    >x-development.2333347.n4.nabble.com%2FDISCUSS-Discuss-Release-Apache-FlexJ
    >S-0-8-0-RC1-tp62274p62341.html&data=02%7C01%7C%7C764b156340ed4161762808d4b
    >1813b7a%7Cfa7b1b5a7b34438794aed2c178decee1%7C0%7C0%7C636328616008847001&sd
    >ata=pnXSK31V8HvCRI9NlEVlGD0SgCczOCQYlw0PyoVZnfQ%3D&reserved=0
    >    Sent from the Apache Flex Development mailing list archive at
    >Nabble.com.
    >    
    >
    
    

Reply via email to