it was discussed between Mike and Wolf to follow some code formatting rules for Spark. Wolf did apply a code formatter to the code. Any objections against applying this patch in the SVN? The builds for us are working nicely.
As this is a larger Change (although only in formatting), I would like to ask you about your opinion about this and your agreement to apply the code formatting patch.
I’m not a developer, so i cannot judge, but in general i don’t mind. I have pm’ed other devs about this thread.
Not that I am a Spark dev, but I am a dev and will chime in with my 2 cents worth. I would recommend that code formatting should only be applied against code that is being edited. I find the cost of formatting all the files to be too high. That cost being the effective destruction of the files history.
Yes, I know the history is still there, but the problem is that the ability to compare changes before and after the code was formated can become severely handicapped if not impossible. This makes tracking down bugs after code was changed very difficult since a code change of a couple of lines is hidden amongst the many other changes that occur simply because of the formatting changes.
In the past, teams I have worked on would compromise by setting new formatting rules to only apply against edited code, thereby only affecting code that was being changed explicitly.
I have the same opinion like rcollier. I think we only work on coding conventions when we edit code.
To reduce the wrong coding conventions I will add checkstyle. Checkstyle check’s for the “Sun coding conventions”.
The overall voting was against an overall code formatting patch. We won’t do it.