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: cvs commit: apache-1.3/src/modules/experimental mod_mmap_static.c
Date Thu, 11 May 2000 21:54:34 GMT
>> Why are we committing willy-nilly to the 1.3 tree?  Chuck tried to get 3
>> +1's for mod_proxy, and he was toldby Roy:
>> 	Basically, my opinion is that no further changes of any kind should
>> 	be made to the 1.3 tree.  If you can get 3 volunteers to do a real
>> 	code review of that branch, fine, but I don't have the time to do
>> 	that and keep up (barely) with the pace of 2.0 development.  So the
>> 	answer is: commit it when you get three +1s from people who have
>> 	actually tested the code.  People who just think its a good idea
>> 	don't count -- I think its a good idea, but that doesn't make it worth
>> 	doing a series of 1.3 releases just to make it work.

I should further conditionalize this to say that I don't think further
changes should be made to the *operational code* on the 1.3 branch --
for me this includes anything that might impact the HTTP interface,
but does not include our support files (what Ken tweaked and I retweaked)
or the Win32 build (which had worked for only a select few developers).
That is my opinion because I know exactly what is wrong with the core code
and why it can't be fixed without breaking backwards compatibility, so
introducing changes to it at this point is very high risk for very
little reward, and distracts the same people from contributing to 2.0.

Whether people continue work on it as review-then-commit or
commit-then-review doesn't matter to me, provided that the review part
is no longer optional for the 1.3.x release.  That just means the people
who want to release their code changes in 1.3.x need to coordinate some
sort of "I'll review your code if you review mine" agreement for
each of the changelog entries.  I trust that anything that has been
reviewed by at least three of us will be worthy of being called Apache.
People who don't have the patience for that type of careful review
shouldn't be messing with the stable branch at all.

Does that sound reasonable?  The thing I am trying to avoid is having
the tail wag the dog when it comes to peer review for a "stable" release.


View raw message