[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: svn commit: r1169837 - /subversion/trunk/subversion/libsvn_delta/compat.c

From: Daniel Shahaf <d.s_at_daniel.shahaf.name>
Date: Tue, 13 Sep 2011 00:38:57 +0300

Hyrum K Wright wrote on Mon, Sep 12, 2011 at 16:11:04 -0500:
> On Mon, Sep 12, 2011 at 2:06 PM, Daniel Shahaf <d.s_at_daniel.shahaf.name> wrote:
> > Look.  You're just wrong.  SVN_ERR() always implies svn_error_trace()
> > regardless of SVN_ERR__TRACING, and there is no additional branch here
> > in an optimized build (and last time you were introducing those changes
> > someone demonstrated this via disassembly excerpts).
>
> As you point out, the removal of the additional branch depends upon
> the optimizer. I hardly think that means I'm "just wrong." Partially
> wrong, maybe, but it certainly isn't as final as you make it sound. :)
>

I fail to see your point here. Did you mean to say that my
characterization of your claim as wrong was itself wrong because your
claim is true when the code is compiled without optimizations?

> In any case, I was making a similar change elsewhere, and decided to
> remove some extra cruft. You may not agree with that change, and
> that's fine. But right now, this is all very new code, and
> (unfortunately) I'm the only person actively hacking it, so I'm kinda
> inclined to do what makes my life easier, so long as it doesn't hurt
> anything else.
>

It doesn't hurt anything else, but your log message described the
change as an "improvement", whereas I think it is a no op, so I pointed
that out.
Received on 2011-09-12 23:40:10 CEST

This is an archived mail posted to the Subversion Dev mailing list.