Re: Future squid development

On 4 Oct 2000, at 8:58, Henrik Nordstrom <hno@hem.passagen.se> wrote:
>
> My goal HEAD should stay reasonably stable, but will quite likely be
> somewhat less stable than the latest STABLE branch. HEAD is not bleeding
> edge developments.

 What is reasonably stable? how many "crashes per million" requests? ;)
 For me reasonably stable is usable on a production box, because this
 the only place I'm using Squid.
 HEAD can be and is reasonably stable since branching from STABLE tree
 up until first DEVEL branch is merged into HEAD. Since that point
 claim "reasonably stable" isn't warranted until tested in numerous
 production environments. And as noone really uses HEAD in production,
 this stability claim is left unwarranted until next major release.
 I personally don't believe that HEAD can be made reasonably stable.
 Reasonably stable and creatively progressing development tend to be
 mutually exclusive. ;)

> To keep things moving I think the evaluation process needs to be
> incremental to keep the HEAD revision reasonably clean at all times, not
> only when a new STABLE branch is about to start.

 I agree. I'm afraid I've been confusing you people with my loose using
 of term HEAD (and lack of understanding of how CVS works).
 I hope I've managed to better express myself in reply to Adrian.

 In two words, I propose branching from HEAD a parallel tree to STABLE,
 that gets same bugfixes that STABLE receives, but also receives some
 of new features that HEAD receives, but not all, only those that do
 not need extensive stability-proof record. That way, this FEATURE tree
 stays based on latest STABLE, includes some new features committed to
 HEAD, and shouldn't become a headache to merge with HEAD - its based
 on HEAD, only a subset of it. Does that make sense?

 What this gives to people, is a branch that is untouched by DEVEL
 branch merges, yet includes small changes to STABLE that people find
 useful. And it allows to keep STABLE totally untouched by new feaures.

 The only assumption is that HEAD accepts patches against STABLE or
 this new FEAT tree. Is this possible?

> So my opinion is that it is better to simply try to have HEAD reasonably
> stable at all times.
..
> Partly agreed, and I think you will see that the proposed approach
> actually gives this by restricting HEAD to only have changes which has
> passed a review and found reasonably stable.

 Well, I personally don't believe there is such thing as always-stable
 Beta software. It is imho unrealistic to restrict major changes to
 only those that has production quality - this slows down development
 dramatically, causes some branches never finish. Eventually this means
 that HEAD is not usable on production boxes, meaning that is will not
 be used on production caches, and this leaves several useful new
 features from being used and tested by endusers.

------------------------------------
 Andres Kroonmaa <andre@online.ee>
 Delfi Online
 Tel: 6501 731, Fax: 6501 708
 Pärnu mnt. 158, Tallinn,
 11317 Estonia
Received on Wed Oct 04 2000 - 04:27:41 MDT