Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 11 Jul 2016 12:50:18 +0000
From:      Alexey Dokuchaev <danfe@FreeBSD.org>
To:        marino@freebsd.org
Cc:        ports-committers@freebsd.org, svn-ports-all@freebsd.org, svn-ports-head@freebsd.org
Subject:   Re: svn commit: r418247 - in head/devel/gps: . files
Message-ID:  <20160711125018.GA72283@FreeBSD.org>
In-Reply-To: <e76649e2-65d7-9400-493d-f31fb1d17697@marino.st>
References:  <201607082228.u68MS4Kf096971@repo.freebsd.org> <20160711122341.GA57651@FreeBSD.org> <e76649e2-65d7-9400-493d-f31fb1d17697@marino.st>

next in thread | previous in thread | raw e-mail | index | archive | help
On Mon, Jul 11, 2016 at 07:45:01AM -0500, John Marino wrote:
> On 7/11/2016 07:23, Alexey Dokuchaev wrote:
> > On Fri, Jul 08, 2016 at 10:28:04PM +0000, John Marino wrote:
> >> New Revision: 418247
> >> URL: https://svnweb.freebsd.org/changeset/ports/418247
> >>
> >> Log:
> >>   devel/gps: Fix compiler error highlight on FreeBSD
> >>
> >>   On DragonFly, GPS has been properly highlighting code errors during
> >>   compilation, but on FreeBSD nothing happens.  Luckily Rod Kay has been
> >>   persistent in diagnosing the cause and he finally isolated it to a
> >>   too-short timeout interval for Expect calls.  The timeout period was
> >>   set to 1 millisecond, the absolute shortest interval possible.  While
> >>   Linux and DF could complete the function call during this time, FreeBSD
> >>   requires at least 10 milliseconds to do the same.
> >
> > Fascinating.  Did you guys try to understand what makes Linux and DF over
> > a magnitude faster than FreeBSD?  Is there something that needs to be
> > fixed?
> 
> I didn't spend much time on it because it's nothing that can be fixed
> with regard to GPS.  I believe the timeout of "1" is itself a logic
> mistake that FreeBSD exposed.

That's understood, right.

> With a cursory glance, I'd guess the problem is centered around polling
> file descriptors -- I'd particularly suspect the __gnat_expect_poll
> function in gcc/ada/expect.c to hold the answer.
> 
> short answer: yes, the implementation of the culprit should be improved
> at the OS-level but you are talking about FreeBSD 12.  I wrote this
> descriptive commit in the hopes it would catch someone's eye.

Noted.  (Thanks for doing a great job with your commit logs, they are of
top notch.  Keep it up.)

./danfe



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20160711125018.GA72283>