On 2/20/2015 2:11 PM, Quanah Gibson-Mount wrote:
The two links on
<https://wiki.apache.org/spamassassin/DownloadFromSvn> starting at:
Viewing the SVN tree through the web
are both broken (404)'s.
I can't recreate the issue, sorry.
The info on <https://wiki.apache.org/spamassassin/DevelopmentStuff>
seems to have been better kept up to date.
Please send me your wiki username and I'll get you in the group to be
able to contribute on the wiki!
Also, are there any plans to move to git?
I can't speak for the project unilaterally but I have a lot of
priorities for the project that would rank above moving to git. However,
I did ask Joe back in 2012 to write up a nice paragraph about why git
would be good for SA.
The two things he identified that make git nicer than svn for large
projects: easy forking, and easy merging. The rest is a good
explanation of the benefits so I'll include it:
It is difficult to control permissions for a large number of users
at once, adding inertia to the process of bringing new people into
the project. With a git repo, any user can fork the project and get
their own local copy of the repository. They can commit, rollback,
and all the usual operations, but the changes stay on their local copy.
To merge other users' changes back into the main codebase, there are
two options. Trusted users can `git push`, which isn't substantially
different from `svn commit`. Untrusted users can create a patch and
send it in. A git patch merges the revision history of the user's
copy of the repo. These two features encourage users to commit after
every minor change, in contrast to svn where commits are always
capable of breaking things for everyone else.
If you want to go crazy with git, you can even use it for staging
(http://stackoverflow.com/questions/10662028/git-staging-server).
Commits and pushes at staging are not visible to production. At
stable versions, staging does a tag and push upstream. Production
can upgrade independently from staging's pushes by simply checking
out the latest revision.
So git certainly has merit but I want to focus more on the research
work, modernizing the ruleqa zones, etc. Git right now provides little
merit to the project but if there was a committer who was desperate for
it and was going to bang out a lot of work, etc. I'd certainly support them!
Regards,
KAM
