On Thu, Nov 25, 2010 at 19:10, Robert Niederreiter <r...@squarewave.at> wrote:
> Hi Folks,
> Why does recent zc.recipe.cmmi sticks hardcoded to version < 1.5.0b1 of
> zc.buildout in setup.py?

Digging into the logs, I've found that the change is introduced in
this changeset:
http://svn.zope.org/zc.recipe.cmmi/trunk/?view=rev&rev=118326 (the
exact change to that file is here:

Buildout 1.5 has major differences from previous versions (one being
the use of a custom site.py and running python with "-S" option), and
that makes it different to test against multiple versions. Currently
all the tests for zc.recipe.cmmi pass with Python 2.[4-6] (and
zc.buildout 1.4.4), but upgrading to buildout 1.5 introduces three
test failures in all three Python versions, due to the changed
behavior of buildout.

With the current Python trunk (2.7.1rc1+), things are a bit different.
Using zc.buildout 1.4.4, the test for zc.recipe.cmmi never finishes,
instead it fires up dozens of Python processes in a seemingly infinite
loop, and consumes all system resources. Upgrading to buildout 1.5.2,
things are a bit better: now the test finishes, but with 4 failures.

It seems it is quite hard to write tests against both buildout 1.4.*
and 1.5.*, but other than the test failures, the package seems to work
just fine with the newest zc.buildout version.

Zope-Dev maillist  -  Zope-Dev@zope.org
**  No cross posts or HTML encoding!  **
(Related lists - 
 https://mail.zope.org/mailman/listinfo/zope )

Reply via email to