On Sat, 2008-01-05 at 10:54 -0700, Wendy Smoak wrote:
> On Jan 5, 2008 10:06 AM, simon <[EMAIL PROTECTED]> wrote:
> Best wait for him then?  As far as I can tell, it did get enough votes:
> http://www.nabble.com/-VOTE--release-for-MyFaces-1.2.1-td14440257.html
> 
> At the end, Matthias mentions that it didn't pass the TCK.  I don't
> think that invalidates a PMC vote to release, though we are under some
> restrictions about calling it an official JSF implementation.  (I"m
> not clear on the details here.)

But the problem is that the files published are only a *partial* release
(myfaces-project and myfaces-impl, but not myfaces-build or
myfaces-api). All the artifacts need to go out together to form a valid
release AFAIK. I'm not sure what effect this will have on users, but it
doesn't seem like a good idea to leave them there..

The fact that they were released too early is a mistake. The fact that
only a partial release happened is due to bad file permissions on the
myfaces-build directory, making it unwritable except for the owner. I've
emailed the current owner of that dir asking for g+w to be added.

The bad dirs have now been removed from people.apache.org. However the
maven release process actually *overwrites* the metadata files in the
parent dir, and we have no copies of the old files AFAIK. So if we do
get the bad dirs removed from repo1.maven.org and mirrors, does this fix
the problem or will the bad metadata files still stuff users up?

I guess the other solution is to just do nothing until the *real* 1.2.1
release goes out, overwriting the bad files. I believe that is expected
to happen in the next week or so. But I feel a little uncomfortable
about that, and at the least we would need to notify the user list.

Regards, 

Simon


Reply via email to