Re: Patches with independent changes

[Eli Zaretskii]

> Date: Sat, 25 Jan 2014 09:01:21 -0800
> From: Paul Eggert <address@hidden>
> CC: address@hidden
> 
> Eli Zaretskii wrote:
> > There was no second change.
> 
> Sure there was.  It would have been easy to apply a simpler patch that 
> contained just the single change needed to fix the porting bug, namely 
> to call chmod rather than fchmod (attached).  Instead, you applied a 
> more-complicated patch that contained multiple independent changes.

No, I applied a _single_ change, which in effect simply restored the
previous code, but only for WINDOWSNT.  The new fchmod code had no
influence on what I added, and wasn't touched.

> Moving on to trunk bzr 116064:
> 
> >> These independent changes weren't needed to fix the bug.
> >
> > Yes, they were needed
> 
> Obviously they were not needed to fix the bug, as the bug would have 
> been fixed without them.  They were "needed" only in the sense that it's 
> nicer for maintainers and users if Emacs is simpler and smaller.

No, it's not "nicer", it's necessary, because otherwise the build
would have failed when compiler warnings are turned on and used as
errors.

> > Your changes were different in kind
> 
> I don't see why.

I doubt that.

> We all install patches containing multiple independent changes, only
> some of which are needed to fix a bug.  And that's OK.  The
> important thing is that changes in a patch should all be related, so
> that it makes sense to install them together.

Then any number of patches can be installed in a single commit,
because they are all "related" -- after all, they are all about Emacs.