Well let's hold off then and do a clean vote. I don't think its a good
precedent for a first release to have a mentor vote -1, especially one
who has been as involved/responsive as you have been. If it helps, I'll
also change my vote to -1 :-)
However, just so I know what needs to change to get you to a +1:
- release artifacts must include incubating in their names
- DISCLAIMER must be present
- review of hashes (per your note below)
- NOTICE: question on this one. I said to remove it, as there was an
admonition in the Apache documentation that said NOTICE file should not
include unnecessary legal notices. As far as Runtime is concerned, this
notice is unnecessary -- although if the ASF finds it necessary, we're
fine with it being there.
- Instructions to compile the source repos (this is covered in the
documentation, but it shouldn't be hard to put this in text format in
the source release.)
- Sign the artifacts with an apache email address
- Add apache to the name
Is that right? All of these seem perfectly reasonable to me. I just
want to make sure we're chasing down the list.
Sterling
On 2/22/16 7:50 PM, Justin Mclean wrote:
Hi,
Sorry but it’s -1 (binding) from me.
To be clear that doesn’t stop other people voting +1, and if you get 3+1 you
can still put it up on the IPMC general list for a vote. You’re also welcome to
try and change my mind, anyone can change their vote after initial voting. All
the -1 means is I wouldn’t release it, but what makes a release good enough
quality to release is going to vary form person to person an that’s all OK.
I would however expect that in the currently form it may not pass an IPMC vote.
It’s very close however and there only a couple of missing things.
I checked:
- release artefacts are missing incubating from their names [1][2]
- signatures OK but not sure re hashes
- missing DISCLAIMER in release artefacts [3]
- LICENSE(s) all good
- NOTICE good but missing original developer (runtime)
- newt doesn’t have a REAME at the top level
- no unexpected binary files in the releases
- all Apache source file have Apache headers / no double headers I could find
- not sure how to compile the source repos - some instruction on this in the
releases would be nice
How were the hashes generated?
I’m seeing this:
$ openssl sha1 larva-0.8.0-b1.tgz
SHA1(larva-0.8.0-b1.tgz)= 99b15843d0a5af3f3d7dbdcb52afb80144ee1255
$ cat larva-0.8.0-b1.tgz.sha
/Users/ccollins/tmp/rel/bin/larva-0.8.0-b1.tgz:
51915329 EE9E17F8 7517C2B6 1C99268B 9AAA478D 2C85AA0B B036276D 4B980A11 9BE18DEB
471E762A A80CB4D5 7478390E 60A0EAE1 0481F723 5FFE83A8 6990D700
You probably want to remove "/Users/ccollins/tmp/rel/bin/larva-0.8.0-b1.tgz:”
from that file.
Some possible improvements:
- Re naming it's a good idea to add apache to the name as well as I believe it
gives some extra legal protection / shows it’s an apache product.
- It a good idea to sign the artefacts with an apache email address.
Thanks,
Justin
1. http://incubator.apache.org/guides/releasemanagement.html#naming
2. http://incubator.apache.org/incubation/Incubation_Policy.html#Releases (note
the word MUST)
3. http://incubator.apache.org/guides/releasemanagement.html#check-list
