Florian MOGA wrote:
IIRC the purpose of having this code in trunk and in the default
build was to ensure that it doesn't get broken by changes to trunk.
My understanding is that if code is outside trunk, it's up to the
author(s) to keep it in sync with any trunk changes. For code inside
trunk, the community takes on this responsibility and if someone makes
changes in trunk they should update other trunk code as needed.
In exceptional cases where this isn't possible, they should create a
JIRA and temporarily remove the failing module(s) from the build.
This implies that code can be moved to unreleased/ only from contrib/.
The scenario which Ant was mentioning when you'd want to make some
experimental modifications on trunk code and wouldn't like to break the
build, in this case you can't move the code from trunk/ to unrealeased/.
Is that correct and what's the answer for that? (personal sandbox folder
maybe)
When I'm doing this I usually use a local copy until I'm convinced
that the experiment was successful :-) I did one of these recently
that wasn't successful, so I just deleted the changes.
For experiments that need to live in svn so that others can review and
comment on them, I think a sandbox is the right place.
Simon
