Date: Mon, 8 Feb 2016 01:53:06 +0100 From: John Marino <freebsdml@marino.st> To: Warren Block <wblock@wonkity.com> Cc: FreeBSD Mailing List <freebsd-ports@freebsd.org>, Kevin Oberman <rkoberman@gmail.com> Subject: Re: Removing documentation Message-ID: <56B7E6F2.9050906@marino.st> In-Reply-To: <alpine.BSF.2.20.1602071702120.74300@wonkity.com> References: <56B754A8.3030605@marino.st> <alpine.BSF.2.20.1602071702120.74300@wonkity.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On 2/8/2016 1:29 AM, Warren Block wrote: > portmaster's one big feature has always been that it has no > dependencies. That was and is important. One of the motivators for > portmaster was portupgrade's Ruby and ruby-bdb dependencies, which often > broke upgrades. 1) poudriere is exactly the same, portmaster is not unique there. 2) In the context of Synth, this is not a legitimate criticism. One always the option of official packages. Synth could even build itself in a short time if someone has a "no official packages" policy. > I have not tried Synth due to the Ada dependency, and so do not know if > it has other portmaster-like abilities, like installing or upgrading a > port from the command line with just an origin (portmaster devel/git) or > whether it can build or upgrade a port or group of interdependent ports > on the host system rather than in a chroot or jail. hmm? The fact that it builds on a dirty host system is a liability; its a major flaw. The typical Synth build environment is the host base system minus the installed packages, so that's clearly superior. For the rest: Synth was intended to be a portmaster replacement so does all the signficant things that portmaster can (better, faster, and a lot more). Ada is not ruby, nor is it a shell script. To consider a superior characteristic as a liability to the point of not testing is a decision I would question. I would think you'd at least want to know what you are missing. > I have committed a change to the Handbook that rewords the portmaster > entry. It removes "recommended", replacing it with "smallest", and > clarifies and simplifies some other text. If one believes (as I do) that an unmaintained tool should not be presented to new users in such a positive light (or at all), this change doesn't satisfy. Existing users know how to use it and where it is so they don't need it in the handbook. If nothing changes, portmaster should be clearly identified as unmaintained (e.g. deprecation or obvious warning) and then if users still choose to use it, that's their informed decision. John
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?56B7E6F2.9050906>