version confusion

[jon * <jo...@clearink.com>]

on 12/13/99 4:15 PM, costin@locus.apache.org <co...@locus.apache.org>
wrote:

> -     public static final String Version = "2.2";
> +     public static final String Version = "1.0R1";

Hmmm..could going back a revision cause confusion? Maybe we should make it
3.0 instead. That way, we kill a bunch of birds with one stone...especially
the one where people who have stupid MIS departments that won't use 1.0
versions.

-jon

Re: version confusion

[Jason Hunter <jh...@acm.org>]

Costin.Manolache@eng.sun.com wrote:

> IMHO we are moving a bit too slow, so I would prefer to have
> discussions after a check in if someone disagree ( or wants more
> informations). Also, a -1 on any patch is enough to roll it back.
> 
> I'll try to give more details in my check-ins.

OK, if I can see more details on check-in I'll be happy.  It takes a
well-informed populace to place good votes, after all.  :-)

-jh-

Re: version confusion

[jon * <jo...@clearink.com>]

on 12/14/99 10:24 AM, Costin.Manolache@eng.sun.com
<Co...@eng.sun.com> wrote:

> Also, a -1 on any patch is enough to roll it back.

Correction for an important point...

-1 + a good reason.

Without the good reason, it can still go ahead imho. That is why I gave all
my good reasons for the version number on the PMC list.

-jon

Re: version confusion

[Co...@eng.sun.com]

> I suppose that's where I'm going with this.  I'd like to see either
> discussion before check-in, or a full check-in message that explains
> the *why* of the change.  Be complete, give reasoning.  I don't have
> time to read all the diff's and would like to know what's happening
> based on check-in messages, so when I leave hibernation I'll not be
> lost.

Hi Jason,

Most of the changes are cosmetic, including the version stuff, or bug
fixes.
We had one very good patch ( BufferedIS ), and another one waiting
( Gregorian callendar ), both improving performance.

Ther was no functionality change or anything else that would confuse
someone - all "big" changes are waiting for a release so we'll not affect
the stablitiy of the code.

IMHO we are moving a bit too slow, so I would prefer to have
discussions after a check in if someone disagree ( or wants more
informations). Also, a -1 on any patch is enough to roll it back.

I'll try to give more details in my check-ins.

Costin

Re: version confusion

[Jason Hunter <jh...@acm.org>]

Anil K. Vijendran wrote:
> 
> Hi Jason,
> 
> No change like this is final unless there is enough consensus.
> 
> The way things like this have been resolved in the past is someone
> checks in stuff and whoever has an objection just replies to that
> commit message and there's a dialog right there.

Are we all OK with that?  It differs from what I was made to understand
at the project's beginning.  

I express concern because I personally am using the check-ins to keep
tabs on what's changing (while I'm in "hibernation mode" this month). 
When I see check-ins like the version change, it makes me wonder why 
the version was changed.  Who agreed to it?  What was the rationale 
for choosing this number?  If this is R1 what are we having R stand 
for?  I presume Release, but it's not stated.

I wonder the same with javac.jar being checked-in.  At one point we
discussed it and decided against it for JDK 1.1 compat reasons.  
Or so I thought.  Why is it checked-in now?  Did something change?

I suppose I can raise flags on all these check-ins and ask the
questions.  But watching Tomcat changes is not my primary job (at least
right now), so I tend to just trust people know what they're doing and
don't pipe up.  But that doesn't help ensure quality!  A little
discussion before submit (or a complete check-in message giving
justification) would make sure we all understand what's going on, and
give the maximum chance for early error detection.

I suppose that's where I'm going with this.  I'd like to see either
discussion before check-in, or a full check-in message that explains 
the *why* of the change.  Be complete, give reasoning.  I don't have
time to read all the diff's and would like to know what's happening
based on check-in messages, so when I leave hibernation I'll not be
lost.

-jh-

P.S.  For the record Sam, I have the utmost respect for the members of
tomcat-dev, especially the people who are making the check-ins that I'm
curious about.  That's the conundrum.  I expect they're doing the right
thing but need to know the reasoning behind the changes so I can
double-check for quality, and make sure my later changes follow suit.

Re: version confusion

["Anil K. Vijendran" <An...@eng.sun.com>]

Hi Jason,

No change like this is final unless there is enough consensus.

The way things like this have been resolved in the past is someone checks in
stuff and whoever has an objection just replies to that commit message and
there's a dialog right there.

Jason Hunter wrote:

> > on 12/13/99 4:15 PM, costin@locus.apache.org <co...@locus.apache.org>
> > wrote:
> >
> > > -     public static final String Version = "2.2";
> > > +     public static final String Version = "1.0R1";
> >
> > Hmmm..could going back a revision cause confusion?
>
> >From the way I understood this process, changes such as this one should
> be voted on before making a change.  Only clear "bug fixes" are to be
> checked in straightaway.  Perhaps decisions are being made offline, but
> that's not conducive to good net collaboration.  I've seen a few other
> controversial check-ins (adding javac.jar to the classpath) that should
> be talked about first as well.
>
> In other words, I think we need to see a little more talking and fewer
> one-manned check-ins.
>
> -jh-
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: tomcat-dev-unsubscribe@jakarta.apache.org
> For additional commands, e-mail: tomcat-dev-help@jakarta.apache.org

--
Peace, Anil +<:-)

Re: version confusion

[Jason Hunter <jh...@acm.org>]

> on 12/13/99 4:15 PM, costin@locus.apache.org <co...@locus.apache.org>
> wrote:
> 
> > -     public static final String Version = "2.2";
> > +     public static final String Version = "1.0R1";
> 
> Hmmm..could going back a revision cause confusion? 

>From the way I understood this process, changes such as this one should
be voted on before making a change.  Only clear "bug fixes" are to be
checked in straightaway.  Perhaps decisions are being made offline, but
that's not conducive to good net collaboration.  I've seen a few other
controversial check-ins (adding javac.jar to the classpath) that should
be talked about first as well.

In other words, I think we need to see a little more talking and fewer
one-manned check-ins.

-jh-

Re: version confusion

[Stefano Mazzocchi <st...@apache.org>]

jon * wrote:
> 
> on 12/13/99 4:15 PM, costin@locus.apache.org <co...@locus.apache.org>
> wrote:
> 
> > -     public static final String Version = "2.2";
> > +     public static final String Version = "1.0R1";
> 
> Hmmm..could going back a revision cause confusion? Maybe we should make it
> 3.0 instead. That way, we kill a bunch of birds with one stone...especially
> the one where people who have stupid MIS departments that won't use 1.0
> versions.

I agree with Jon, 3.0 sounds better.

-- 
Stefano Mazzocchi      One must still have chaos in oneself to be
                          able to give birth to a dancing star.
<st...@apache.org>                             Friedrich Nietzsche