[Zope-dev] Zope Head (2.8) breaks "refresh"

[Tres Seaver]

On Sat, 2003-12-13 at 22:10, Stuart Bishop wrote:

> Until following the steps in doc/INSTALL.txt fires up a single ZEO
> server bound to the loopback address and a single ZEO client (with
> authentication on), the vast majority of developer and production
> installations will be ZEO-less. This also makes ZEO-less installations
> the best tested and most stable environment, or at least people will
> assume so.

Can't argue about the assumptions, nor about how people are likely to
start off using Zope;  nevertheless, *every* ZC developer routinely uses
ZEO, even for sandbox development.  None of us would even contemplate
*not* using ZEO in production:  its "testedness" is pretty well
established.

Here are a couple of compelling reasons to run with ZEO all the time:

  - You can't scale Zope across hardware without ZEO, which means you
    need to assume that you will (or at least may) need it in
    production;  hence, you might as well develop and test your app /
    site with ZEO.

  - ZEO lets you get a debug session open on your database without
    stopping the live site (this can be a lifesaver, particularly in
    development and site configuration;  ever lock yourself out of the
    site by messing up the configuration of your root user folder?)

  - ZEO speeds restarts, often dramatically (verifying persistent caches
    used to be problematic, but ZODB3 3.2 has a fix for the common
    case). 

  - ZEO makes ZODB-dependent unit tests run faster (another facet of
    the restart problem).  This has been particulary true for testing
    products installed in INSTANCE_HOME, because the machinery for
    stitching together the __path__ of Products was tightly coupled to
    appserver startup.  We might be able to fix that in future Zopes.

Tres.
-- 
===============================================================
Tres Seaver                                tseaver at zope.com
Zope Corporation      "Zope Dealers"       http://www.zope.com