On 05/27/2015 10:56 AM, Georg Baum wrote:
Scott Kostyshak wrote:
On Mon, May 25, 2015 at 7:24 PM, Uwe Stöhr <uwesto...@web.de> wrote:
I did not update all references. Work was in progress. Now everything
should work.
My personal opinion is the following: try your best to not commit work
in progress. If you would like to commit work on progress, do not do
it on master. Create a different branch and work on that. Push it to
the features remote if you want. Feel free to ask if you're not sure
what I'm talking about.
I'd formulate it a bit more radical: If you submit work in progress, which
causes tests to fail (especially without communication), then you are
actively waisting other developers time. If this continues, although several
developers complained and offered help, then I'll get really upset.
It's perhaps worth saying that this is at least some change from our policy
back when we used svn. When I was working on HTML output, for example,
people encouraged me to commit as I went along, rather than to use a
feature branch. This was largely because merging, etc, in svn was so hard.
Now, with git, it is a completely different situation. It is easy to
create local
branches, etc, and to work there, and there is no reason to commit work
that is not complete.
Richard
