Date: Fri, 27 Jun 2014 16:09:58 +1000 From: Kubilay Kocak <koobs@FreeBSD.org> To: Matthias Andree <matthias.andree@gmx.de>, freebsd-ports@freebsd.org Subject: Re: RFC: Remove single MASTER_SITE warning in portlint? Message-ID: <53AD0AB6.3080304@FreeBSD.org> In-Reply-To: <53AD075C.9080409@gmx.de> References: <53AD060E.9080601@freebsd.org> <53AD075C.9080409@gmx.de>
next in thread | previous in thread | raw e-mail | index | archive | help
On 27/06/2014 3:55 PM, Matthias Andree wrote: > Am 27.06.2014 07:50, schrieb Joe Marcus Clarke: >> Apparently there was some discussion on IRC about this. A PR (Bugzilla) >> has been opened requesting the portlint warning about only a single >> MASTER_SITE to be done away with. >> >> https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=191079 >> >> This was added before my time with portlint, and I could go either way >> on it. We do have a list of CDN MASTER_SITE aliases to avoid warning >> for silly things. But the question remains: do we need this warning at >> all? > > I would probably prefer an option to make it actually run curl -I (send > HTTP HEAD commands) or thereabouts against make fetch-urlall-list to > check if the file is fetchable, either way you go with the warning. We > could then have warnings about unfetchable URLs. I know we have > separate online services for that, but it would be fitting for portlint. > > _______________________________________________ > freebsd-ports@freebsd.org mailing list > http://lists.freebsd.org/mailman/listinfo/freebsd-ports > To unsubscribe, send any mail to "freebsd-ports-unsubscribe@freebsd.org" > That's an interesting thought Matthias. Technically you could even check for multiple RR's for a host lookup, thereby making the check more robust to false positives. I believe the user experience for distfile fetching could be improved (for transient and permanent failures) in other ways too, and I'm not yet sold either way on the question of this check being removed. For the sake of argument (I haven't though about my position on this yet), we (and portlint) could enforce 2 MASTER_SITES. Mirroring is a solved problem and we have the resources to do so, why might we not go down that route? Koobs
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?53AD0AB6.3080304>