The old old discussion... (was Re: unsigned host port)

["Ralf S. Engelschall" <rs...@engelschall.com>]

In article <97...@decus.org> you wrote:
> From the fingers of Marc Slemko flowed the following:

> >----------------------------
> >revision 1.93
> >date: 1997/03/29 03:40:34;  author: fielding;  state: Exp;  lines: +2 -2
> >Host port changed to unsigned short.
> >
> >Submitted by: Ken Coar, PR #276
> >Reviewed by: Roy Fielding, Chuck Murcko
> >----------------------------
> >
> >This change didn't seem to make it to the mailing list.  
> >The reason I bring it up is because mod_rewrite hasn't yet been updated:

>     I'm not sure what you mean.  I received a copy of the update notice
>     from apache-cvs.  The patch that went in didn't include mod_rewrite
>     by design, although that may have been a design error.  ;->

Correct, this is not Marcs fault. He included a hint for me that I should
include also a patch to mod_rewrite. And I immediately had done this: I've
incorporated this into mod_rewrite 3.0.3-dev.   The point of discussion is a
different one:

>     When I submitted the patch I *didn't* include the required change to
>     mod_rewrite, but flagged it for Ralf's attention.  He acknowledged
>     it, but I don't know if he made the change in his own copy and
>     didn't submit a patch to the group, or hasn't made it at all, or
>     what.

I've incorporated it on 29-Mar-1997 according to my ChangeLog.  But because we
committed mod_rewrite 3.0.2 the day before I wanted to spam the list again
with a 2-line patch for just 3.0.3. I wanted to wait for other things, so I
didn't immediately released 3.0.3. In the meantime I've changed things in the
documentation, etc.

>     As I recall, the major feeling was that, if mod_rewrite was to be
>     considered a regular part of the Apache package, the copy the
>     repository had to be considered the master, and Ralf's a copy
>     thereof, rather than the converse.  

<grin> Yes, sure, but how should I make the repository to be the master
without CVS access? Think about it. Without I had to do it in my version and
regularly summit [SYNC] patches to the list. Sorry, this is not my fault: I
wanted _TWO_ times access to CVS so I can directly maintain mod_rewrite in the
Apache repository, but at BOTH DISCUSSIONS the people finally thought that CVS
access is not really needed to provide an up-to-date copy.

So I maintain mod_rewrite in my private master source and only summit [SYNC]
patches _from time to time_ when there is enough stuff to commit.

** NO: I really don't want to start the discussion again about
** CVS access. I'm really lazy to discuss this point any more.
** No more discussion from my side! If people really think it is 
** not neccessary then thats the problem of these people. 
** I've already voted for CVS access two times in the past and
** never received it. Perhaps I'm a too boring or bad hacker 
** for the Apache Group. I don't know... ;-)

>     However, that never seemed to
>     happen.  Or perhaps I'm misremembering.  The question was
>     complicated by the fact that the latest mod_rewrite is always
>     backward-compatible.

backward-compatible? To what? I don't see the point of discussion
here, sorry. Please explain which fact you exactly mean.

>     I can't explain why, but when I've had something that needed
>     application to mod_rewrite, I *haven't* been including it myself but
>     just flagging it for Ralf's attention, and he's typically applied it
>     in a subsequent patch submitted for the module.  {Sigh}  Mea culpa.
>     I'll be better..

Yes, this is not the fault of you, it is the consequence of the CVS access and
the fact that I don't wanted to start a "[SYNC] mod_rewrite 3.0.3" just for
this simple 2-line patch.

>     So: what *is* the story with where the master mod_rewrite lives?

Simple:

- The master should be the Apache repository and the
  primary maintainer should be RSE. Anyone can submit patches and when RSE
  really thinks it is a bad one he can vote it with -1. Else they get
  committed by any one of the list. But RSE don't want work this way unless he
  also has access to the CVS. Because I don't want to be the primary
  maintainer of a piece of code while I don't have direct access to this piece
  of code.

- Because RSE is still not able to maintain mod_rewrite directly at the Apache
  repository, he maintains it privately at his own closed site. And to bring
  the repository up-to-date from time to time he has to submitt "[SYNC]
  mod_rewrite 3.0.x" patches with the hope that after voting someone actually
  commits them. But this sync-patches occur only when there is really enough
  stuff because it is silly release new versions for every little patch.

Think about this situation and what is better.
I think the first one. But... :-(

                                       Ralf S. Engelschall
                                       rse@engelschall.com
                                       www.engelschall.com

Re: The old old discussion... (was Re: unsigned host port)

[Rob Hartill <ro...@imdb.com>]

On Tue, 8 Apr 1997, Ralf S. Engelschall wrote:

> Simple:
> 
> - The master should be the Apache repository and the
>   primary maintainer should be RSE. Anyone can submit patches and when RSE
>   really thinks it is a bad one he can vote it with -1. Else they get
>   committed by any one of the list. But RSE don't want work this way unless he
>   also has access to the CVS. Because I don't want to be the primary
>   maintainer of a piece of code while I don't have direct access to this piece
>   of code.

That's fine as long as the code is treated in the same way as the rest...

Anyone can offer patches and any of the CVS prived users can commit them.
Ralf can have a veto (-1) just like everyone else currently has over any
code change. Ralf, you'd also need to play by the same rules with respect
to what type of changes are okay at any given time - e.g. you can't demand
a major update in the middle of a beta phase and you should offer large patches
and idea changes to the group before commiting them. That means a change
because you'd no longer be able to invent new functionality and add it
without first gaining group approval.

> - Because RSE is still not able to maintain mod_rewrite directly at the Apache
>   repository, he maintains it privately at his own closed site. And to bring
>   the repository up-to-date from time to time he has to submitt "[SYNC]
>   mod_rewrite 3.0.x" patches with the hope that after voting someone actually
>   commits them. But this sync-patches occur only when there is really enough
>   stuff because it is silly release new versions for every little patch.
> 
> Think about this situation and what is better.
> I think the first one. But... :-(

Another option is to take mod_rewrite out of the Apache distribution
and point people at it at Ralf's site. If you want full control over
the code, I suspect this is your only option.

All this is IMO.

--
Rob Hartill                              Internet Movie Database (Ltd)
http://us.imdb.com/tour   .. a site for sore eyes.