[Twisted-Python] svn-reorg blocking

[Jonathan Lange]

On 27/10/05, James Y Knight <foom at fuhm.net> wrote:
> On Oct 26, 2005, at 12:10 AM, Jonathan Lange wrote:
> > I strongly and vigorously oppose merging this branch until there are
> > unit tests for James' modifications to Trial.
>
> Nobody has suggested that such a thing should be done; there is no
> need for strong and vigorous opposition to a position nobody has
> advocated.
>
> > Instead I recommend
> > moving serialport up a level (as you suggested earlier).  That appears
> > to be the simplest thing that could possibly work.
>
> This statement makes no sense in this context. Fiddling with
> serialport does not remove the need to make changes to trial that
> require new unit tests. As I said before, I don't even think it
> removes the need for the attribute to mark ignorable subpackages
> either, but, I'm not going to argue that point further.
>

My mistake. I remembered that there weren't tests for __path__ support
in the middle of the email -- and by then I had already used the word
"vigorously", which I've been aching to use in a sentence for some
time.

However, I still hold that the attribute to mark ignorable subpackages
is currently unnecessary.  If it turns out that the attribute *is*
necessary, I'm more than happy to see it added.

Still, delight in vocabulary and strong disagreement aren't excuses
for sloppy logic and a disregard for the facts -- so I apologise.  If
it helps, I'll own that my paranoia springs from my experiences
maintaining Trial in 2002-2003, when all people added "useful"
features with little consideration (leading to bloat, leading to The
Rewrite).

While the circumstances have changed (not least because your code is
considered), I'm still going to hang on to that paranoia, if only to
slow things down.

cheers,
jml