Justin Erenkrantz wrote:
> No, as Joe said, you ignored prior comments - this isn't the first
> time that these concerns have been raised.
I did not ignore prior comments, I responded to the comments, and my
responses were ignored.
As a result of this, I had no clear indication of what further action
was necessary. Was my response sufficient to address Joe's concerns? Was
there a compromise that could have been reached? No way to tell until
Joe had responded, and Joe chose not to do so until suddenly piping up
six months later and stamped on the code.
This is exactly the way this issue should not have been handled.
>> You have raised some valid points which need to be addressed in the current
>> code. Unfortunately you delivered those points with the blunt instrument of
>> the -1, and as a result the SSL part of the code has effectively been
>> withdrawn, and any chance of those issues ever being addressed has been
>> vastly reduced if not eliminated entirely.
>
> If and when you wish to address *all* of the issues, I (and I think
> Joe too; but I can't speak for him) will be happy to see it back on
> trunk.
Joe has vetoed the code on trunk, and took it upon himself to remove the
code entirely. I think Joe has made it pretty clear where he stands on this.
> But, releasing a half-completed API in 1.3 is extreme badness.
Absolutely.
The solution is simple: leave it on trunk, remove it from the v1.3.0
branch until the issues are resolved, so that the folks wanting to see
v1.3.0 released are not prevented from making that happen. I don't think
anybody would have objected to the suggestion to do so.
Regards,
Graham
--
|