> Let us get more tests first and then do refactor.
I disagree. There has been a lot of work contributed by a various folks.
The proposed changes were proposed BECAUSE of VERIFIED AND TESTED problems.
The problems surfaced months ago, and surfaced again last week. The impact
is felt by users, and verified by Danny, who can crash James by overloading
it within minutes. Andrei found this problem months ago, and proposed the
changes. It took time to whip those changes into shape, accounting for
changes in Avalon code. Now Peter has them ready for review, testing, and
deployment.
> If these bugs can be resolved without restructuring code, you should check
> them in.
> If you need to re-architect or re-factor for these fixes, you should post
your
> proposal and your fixes.
I really don't consider these changes to be architecturally significant.
They are a pretty straightforward class re-factoring. The ARCHITECTURE is
still the same.
I'm glad that you're taking time to be active again, Harmeet, but it seems
that you do have some items to catch up on. Peter has been a most active
contributor here since mid-summer. Probably 90% of the commits since he
received Committer status were done by him. He's read through most of the
James code base. He's participating on the Avalon developers list. He's
tested, integrated, etc., (which is good since other Committers have been
busy with their other commitments). I think that he's earned trust and
respect.
So because I already understand these changes, and I don't see a problem
with them, I really don't want to see them held up without a clearly
explained reason. You asked for a few days to review the changes, but I
still haven't heard what you think is wrong with them. What I seem to hear
is that you don't understand what was wrong with the old code. Maybe I am
just missing it. Do you have any concrete problems with the changes, or you
just don't understand why it is necessary to change?
> Testing is really important and should be encouraged as much as possible
> rather than building new hoops for it.
Testing is VERY important. But as a number of people have noted, real world
traffic can crash James. These fixes correct that problem. So I don't see
the need to wait for yet another test before we can move forward, unless
there is a REASON.
I'm interested in seeing that SMTP and POP3 work properly, and this set of
long awaited server and handler changes addresses documented issues, doesn't
change the architecture, and is a straightforward re-factoring. It is vital
to test it to make sure that it works, but that is true of any change.
I hope that your reaction isn't preconditioned because you don't like other
changes that Peter made to the NNTP code. Perhaps Peter would have been
better off if he had first done the service changes, and after those were in
place, worked on NNTP. Personally, I had suggested that NNTP be deprecated
it into proposals/ until it was fixed, but Peter felt that it was important
for NNTP to work, so he took the time to do what no one else did this year,
and has been working on it.
Danny and I have had disagreements on API, but this is the first discussion
I've seen since joining the list where it sounded to me that it bordered on
getting personal. I hope that we can dismiss that issue, and focus on the
code. James needs all of the developers it can get, and we certainly don't
need developers leaving out of frustration. As with many Open Source
projects, it is hard enough getting people with the time, interest and
energy to contribute. That is one of the reasons why Peter wanted to get
NNTP and IMAP working. He feels that the better the base, the easier it
will be to attract more helpers.
--- Noel
--
To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]>
For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
