> On May 2, 2021, at 10:22 AM, Alex The Rocker <[email protected]> wrote:
> 
> Hi David,
> 
> Thanks for your answer, this is perfectly clear.
> 
> Will this upcoming TomEE 8.0.7 release include latest Tomcat / CXF
> dependencies that fixes CVEs that showed up since 8.0.6 release ?

It'll definitely contain the latest CXF as many of the TCK failures were there. 
 We're actually building against the unreleased 3.5.0-SNAPSHOT, but hopefully 
we can back that down to the last release, 3.4.3. I don't know the status of 
the Tomcat version as I've not been doing those TCK fixes, but I suspect it's 
the latest latest.

If anything is missed, we can roll a 8.0.8 immediately even as soon as Tuesday. 
 Basically, if we don't get a release vote up in the 8 to 16 hours, we'll miss 
the Jakarta EE 9.1 release window.

-David

> 
> Le dim. 2 mai 2021 à 19:09, David Blevins <[email protected]> a écrit :
>> 
>>> On May 2, 2021, at 9:39 AM, Alex The Rocker <[email protected]> wrote:
>>> 
>>> I am a bit confused : I thought that renaming of javax.* into
>>> jakarta.* packages for the EE API is only targeted for TomEE 9.x.
>> 
>> All the source code is in TomEE 8.0 and TomEE 9.0.x is just created through 
>> bytecode transformation.  The long and short of that is they pass/fail 
>> almost the exact same tests as it's the same code. We can do library 
>> upgrades in the `tomee-jakarta` repo and add a patch file for third-party 
>> dependencies here and there, but any fixes on the TomEE side go into 8 and 
>> are automatically picked up in via bytecode transformation of 9 as well.
>> 
>> In fact, I've been doing most my local work exclusively running the Jakarta 
>> EE 8 TCK and just letting the automated systems do the running of the EE 9 
>> TCK.
>> 
>>> In such case, what would be the content of a TomEE 8.0.7 release?
>> 
>> Everything in this list is fixed, except about 10 tests which are not part 
>> of the Web Profile and therefore not required for certification and will be 
>> fixed later.  Though it says "EE 9" these failed for both 8 and 9 TCKs:
>> 
>> - https://issues.apache.org/jira/browse/TOMEE-3140
>> 
>> We're still one TCK test shy of passing and have some work to do on the API 
>> jars to pass the signature tests, but it's looking good.
>> 
>> Brief pause for me to take a nap, however, as I've been up for 24 hours and 
>> keep nodding off at the keyboard :)
>> 
>> 
>> -David
>> 
>> 
>>> 
>>> In any cases, I'd be happy to see a TomEE 8.0.7 release soon, in order
>>> to get latest CVE fixes since 8.0.6.
>>> 
>>> Kind regards,
>>> Alexandre
>>> 
>>> Le sam. 1 mai 2021 à 03:32, David Blevins <[email protected]> a écrit 
>>> :
>>>> 
>>>> Heads up that we are narrowing in in the last few TCK issues and there is 
>>>> still some chance we can be Jakarta EE 9.1 Web Profile certified in time 
>>>> for the Jakarta EE 9.1 release vote Monday.
>>>> 
>>>> It would be super super and I mean *super* tight....
>>>> 
>>>> However, if we can get it done we'll need to do a release vote by no later 
>>>> than Sunday afternoon and file our certification request.  We don't need 
>>>> to have concluded our vote to make the Jakarta EE 9.1 release ballot, we 
>>>> just need final binaries of our own to be at least in staging and in the 
>>>> process of our own vote.
>>>> 
>>>> We do need that vote pass, however, so that would require some pragmatism 
>>>> on all our parts.
>>>> 
>>>> For that reason I recommend we do not try to push out a 9.0.0 final, but 
>>>> go ahead with 9.0.0-M7. If there are some issues with the binaries we put 
>>>> up for vote, unless they are legal issues, we can still release them and 
>>>> immediately fix the issues next week in a subsequent 8.0.8 and 9.0.0-M8.  
>>>> There's no reason to "wait", we can simply release twice.  Version numbers 
>>>> are free.
>>>> 
>>>> The Jakarta EE 9.1 release vote lasts for two weeks and an announcement 
>>>> would happen some days after that.  Ff we did want to push out a 9.0.0 for 
>>>> the announcement, we'd have at least till May 17th to do that, perhaps 
>>>> even the 20th.
>>>> 
>>>> The reason we want to get certified in time for the ballot is recently 
>>>> there was a change that implementations listed on the ballot get a special 
>>>> place at the top of the specification page.  Any implementations that come 
>>>> even one day later cannot be included and will not be accepted or given 
>>>> special designation.  This lasts forever and is a permanent advantage to 
>>>> those in the list.  It's also a permanent *disadvantage* to those not on 
>>>> the list.  It's eat or be eaten.
>>>> 
>>>> So that's what we're going for:  A staged binary up for a vote here, 
>>>> passing the TCK, in time to be listed on the Jakarta EE 9.1 release ballot 
>>>> Monday.
>>>> 
>>>> 
>>>> -David
>>>> 
>> 

Reply via email to