Why Clang

Chad Perrin perrin at apotheon.com
Mon Jun 18 21:00:51 UTC 2012

On Mon, Jun 18, 2012 at 10:30:23PM +0200, Wojciech Puchar wrote:
> >>scratch and this resulted with thing 5 times larger,
> >
> >*YOUR* measurement of sizes was faulty.  <grin>
> be more exact.

I believe Robert Bonomi (you didn't include attribution for the previous
email, I notice) *was* more exact, in that the rest of his email
explained what he thought of your glossing over the various factors that
might contribute to binary size.

I notice you ignored most of it in your response, too.

> >
> >I'm sure that you _also_ are aware that a larger program size does *NOT*
> >necessarily mean 'bloat'.
> of course. really i can write programs.
> and really - i don't understand all this fuss about "better error
> reporting".
> Really i don't have problems to read gcc error messages when i
> compile my programs.

I can generally puzzle out what caused various GCC warning and error
messages when trying to compile my own code, given comparison of what's
going on in the messages with what's going on in my code and reasoning
through the connections between different parts of the code.  That sort
of thing is required probably 70% of the time, in my experience.

With Clang, by contrast, I find that's required only about 20% to 30% of
the time.  Otherwise, the warning and error messages tend to get me a lot
closer to the actual point of failure than GCC.

*That* is what "all this fuss about 'better error reporting'" is about.

Chad Perrin [ original content licensed OWL: http://owl.apotheon.org ]

More information about the freebsd-questions mailing list