Date: Thu, 27 Apr 2006 10:45:12 +0400 From: "Andrew Pantyukhin" <infofarmer@gmail.com> To: "Doug Barton" <dougb@freebsd.org>, "Matthias Andree" <matthias.andree@gmx.de>, "FreeBSD Ports" <ports@freebsd.org> Subject: Re: ports/96379: PLEASE DO NOT SEND OPENVPN 2.0.7 REMINDERS! Message-ID: <cb5206420604262345i6282fb80k1d90d4b6773a2b00@mail.gmail.com> In-Reply-To: <20060426215845.GA6337@merlin.emma.line.org> References: <20060426192442.6E3B654E@k7.mavetju> <20060426201101.GA1433@merlin.emma.line.org> <444FE894.70008@FreeBSD.org> <20060426215845.GA6337@merlin.emma.line.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On 4/27/06, Matthias Andree <matthias.andree@gmx.de> wrote: > On Wed, 26 Apr 2006, Doug Barton wrote: > > > Matthias Andree wrote: > > > Please close ports/96379. > > > > > > I am not accepting patches to update the FreeBSD/DragonflyBSD port > > > security/openvpn to 2.0.7. > > > > You may wish to consider whether the effort to continue telling people = you > > will not update the port will be lesser, or greater, than the effort > > required to update the port. > > My decision (call it ruling if you like) is that I will not ask > thousands of users to update just because a handful of fools can't get > past the version number to the changelog. Most of the requests had > actually been forwarded by committers to ask my feedback... Chill out, man :-) This handful of men are the ones we should praise, not scorn. In fact, they invested their time to learn the basics of making/updating a port, they took the trouble to notice 2.0.7 is out, to try it on their boxes, make a patch and go through a not-so-transparent process of submitting a PR. And what do they expect in return? Well, if they do a great job and the patch gets into the tree they'll get a heart-warming "Committed, thanks!" If not, if they make mistakes, at least we should try not to discourage them. They get no money, fame or glory. Should we really bash them for not being attentive enough? Back to 2.0.6-2.0.7. Below is a fragment of the difference between the versions. As far as I can tell without studying the rest of the code, this does not only fix windows versions. Granted, *nix systems might have get away with a wrong typedef, but it doesn't mean the old code is safe. Only the most important things go to changelogs. What we see here tells us that further investigation might be needed to know the effect of the mistake, even if USE_64_BIT_COUNTERS is defined by default. It doesn't mean that you're wrong about holding the horses and staying with 2.0.6. Saving users the time of download and recompilation is a nice decision. I only ask you to keep an eye out beyond changelogs. We should all go to the source when in doubt. BTW, thanks for a great port (this and many other ones) and for doing such a great job maintaining them all through the years. We use openvpn at several locations and we really appreciate the speed at which your provide updates. Thanks! /* - * Statistics counters. + * Statistics counters and associated printf formats. */ -typedef unsigned long long int counter_type; +#ifdef USE_64_BIT_COUNTERS + typedef unsigned long long int counter_type; +# ifdef WIN32 +# define counter_format "%I64u" +# else +# define counter_format "%llu" +# endif +#else + typedef unsigned int counter_type; +# define counter_format "%u" +#endif
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?cb5206420604262345i6282fb80k1d90d4b6773a2b00>