Date: Tue, 1 Aug 2017 11:59:31 +0000 From: Andrew Hamilton-Wright <andrew.hamilton-wright@uoguelph.ca> To: Benjamin Kaduk <kaduk@mit.edu> Cc: "freebsd-doc@FreeBSD.org" <freebsd-doc@FreeBSD.org> Subject: Re: why recommend "portmaster -af" for full port rebuild? Message-ID: <YTXPR01MB02701E165DC7E066CCF87CEDC4B30@YTXPR01MB0270.CANPRD01.PROD.OUTLOOK.COM> In-Reply-To: <20170801020406.GW58771@kduck.kaduk.org> References: <62F28CA2-54E4-447A-B290-E57A62229D47@uoguelph.ca>, <20170801020406.GW58771@kduck.kaduk.org>
next in thread | previous in thread | raw e-mail | index | archive | help
I did experiment with '-d', but not with '-D', before switching back to usi= ng "portupgrade -af" which in my experience has been much more satisfactory= . Thanks for the tip -- the next time I try "portmaster" I will look into it. The main argument against a "poudriere" based solution is the amount of con= figuration that that tool requires, as I understand it. Both "portmaster" = and "portupgrade" work out-of-the-box. I am guessing that the recommendation of "portmaster" is likely based on th= e fact that that port doesn't require the dependencies that "portupgrade" d= oes, and if this is the case, that is certainly a valid point -- however I = do think that it is fair to mention on the install page that one should exp= ect a good bit of babysitting the rebuild if one were to follow the advice = written there. This time around, the babysitting included not only the usual "shall I dele= te" questions regarding build products, but also several complete aborts of= the build process due to items with no maintainer. I was quite surprised = to see that "ncurses" was one of these, as that seems to be a critical piec= e of many things. In any case, "portupgrade -af" got me to where I needed to go, with only mi= nimal cleanup required this morning after a full rebuild. Thanks, Andrew. ________________________________ From: Benjamin Kaduk <kaduk@mit.edu> Sent: Monday, July 31, 2017 10:04:06 PM To: Andrew Hamilton-Wright Cc: freebsd-doc@FreeBSD.org Subject: Re: why recommend "portmaster -af" for full port rebuild? On Mon, Jul 31, 2017 at 02:55:31PM +0000, Andrew Hamilton-Wright wrote: > > Dear FreeBSD Docs, > > On the upgrading/updating page: > https://www.freebsd.org/doc/handbook/updating-upgrading-freebsdupdate= .html > there is a line that recommends using > portmaster -af > to rebuild all of the ports. There is even a recommendation that '-G' be= added to avoid interaction involving configuration screens. > > This, however, ignores the many other interactions that portmaster still = will require. > > Is there no way to rebuild the ports without interaction? It seems that = even with > portmaster -afG --no-confirm -y > I still get a long parade of interaction requests -- far more than I woul= d if I used portupgrade. > > I will note that Doug Barton, the author of portmaster, seems to advise a= gainst using it in this form, as noted in this thread: > https://lists.freebsd.org/pipermail/freebsd-stable/2009-August/051623= .html > > > Why is portmaster being recommended on this page, and if this is the best= option, why is there no discussion of the many interactions that will be r= equired (in spite of the implication of the discussion of the '-G' option)? Does this hold even when one picks one of -D or -d? (I mostly assume that it is implicitly expected that any portmaster user will get so frustrated by those prompts that they set an appropriate configuration entry to make them automatically.) Regardless, it may be worth considering documenting poudriere over portmaster or other alternatives, as I'm given to understand that it presents a smoother experience. -Ben
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?YTXPR01MB02701E165DC7E066CCF87CEDC4B30>