Date: Tue, 12 Mar 2002 14:31:03 -0500 (EST) From: Mikhail Teterin <mi@aldan.algebra.com> To: knu@iDaemons.org Cc: roam@ringlet.net, cvs-committers@FreeBSD.org, cvs-all@FreeBSD.org, portmgr@FreeBSD.org Subject: Re: cvs commit: ports/graphics/autotrace Makefile ports/graphics/graphviz Makefile ports/graphics/libafterimage Makefile ports/graphics/librsvg Makefile ports/graphics/libwmf Makefile ports/graphics/sdl_ttf Makefile ports/print/ft2demos Makefile ... Message-ID: <200203121931.g2CJV33b082819@aldan.algebra.com> In-Reply-To: <86u1rlaa42.wl@archon.local.idaemons.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On 13 Mar, Akinori MUSHA wrote: > You are still misunderstanding something. PORTREVISION is not just > for binary packages. As you say, few people do binary upgrading > because there is no good way to do that (even with portupgrade); we > are here talking about ports. [...] > One day one builds and installs port A and then port B which depends > on A, and after a week the port A is updated with a library major > bump. Now, if the committer who updates the port A do PORTREVISION > bumps for the port B and other dependant ports properly, one will be > notified by the /etc/periodic/weekly/400.status-pkg script that an > upgrade of the port A is available and the port B should be upgraded > at the same time, For A to stop working, one has to remove the older libB.so.N. To remove the earlier version of libB, one has to do pkg_delete of the old B. That is when he/she will be told, that port A depends on it, and when he/she will know to delete/upgrade port A as well. If he/she piles up the new port B on top of the old B, port A will continue working since the old libB.so.N will still be there along with the new libB.so.N+1 >> > In the last discussion, some people expressed anxiety that people >> > might neglect bumping PORTREVISIONs because of your changes. Do you >> > remember? >> >> I do. However, I countered, that such bumping is only serving the binary >> upgrades, which can be done without it anyway, and rejected that argument. > > You cannot "reject" what we once introduced after a discussion. > PORTREVISION was introduced for the very purose; to help people know > when they should upgrade which. I'm not rejecting PORTREVISIONs bump. I am saying it is not important enough to justify having to build the latest versions of each port, that a port I actually want depends on. >> I find it much easier to communicate with computers, than with people >> :-) If I knew Python, I'd concentrate on coding the feature instead >> of preaching to people... >> >> Then, the PORTREVISION bump would only be needed if the port itself >> changes. > > Well, do not mistake me... While the feature can and should be > implemented in most tools like the standard pkg_* and portupgrade, > proper PORTREVISION bumping keeps being a duty until then. Ok. But, please, hurry :) -mi To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe cvs-all" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200203121931.g2CJV33b082819>