httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Roy T. Fielding" <field...@kiwi.ics.uci.edu>
Subject Re: incorporating? (was: Copyright & donating code)
Date Tue, 26 Jan 1999 03:59:29 GMT
>Let me just jump in here to say I agree with Roy, you are reading too much
>into this.  The big debate afterwards was not about a bug in the 1.3.4
>patch itself, but about the approach to working abound a bug in Lynx.  I
>can't say I'm too happy about how this debate ended, as implied threats to
>nuke previously accepted code are not a nice way to treat a new
>contributor, but that is another matter.

To clarify, it wasn't a threat -- it was a fact.  What is important to
us as a group is that Apache do the right thing in the right way to best
serve our actual customers.  Most of the time this corresponds to what
is written in the IETF RFCs, but not all of the time, and that includes
the RFCs that I wrote myself.  When the right thing differs with what
the RFC says, then we either do the right thing and fix the RFC later
or disable the feature until we can get around to fixing the RFC.
I can't tell you how many times I had this argument (from the other side)
with Robert Thau.

The question, then, is what is the right thing?  That's what I meant by
saying that Koen needs to demonstrate why the current implementation
is buggy through some actual scenarios, rather than just pointing at
the RFC and reminding me how much time he spent working on it (of which
I am very much aware already).  Keeping in mind, of course, that the fix
must be less harmful overall than the thing being fixed.  The alternative
for something like TCN is to remove those bits that might confuse the
client into thinking we are following the RFC completely.  If I thought
that was a good alternative, I wouldn't have invested my vacation time
in applying the original patches.

....Roy

Mime
View raw message