Date: Sun, 27 Mar 2005 12:18:46 +0200 From: Alexander Leidinger <Alexander@Leidinger.net> To: Kirill Ponomarew <krion@voodoo.oberon.net> Cc: Kris Kennaway <kris@obsecurity.org> Subject: Re: linux-* pkg-plist errors Message-ID: <20050327121846.518f0772@Magellan.Leidinger.net> In-Reply-To: <20050327095541.GB16502@voodoo.oberon.net> References: <20050327022418.GA92678@xor.obsecurity.org> <20050327112422.680b25a3@Magellan.Leidinger.net> <20050327092943.GA16502@voodoo.oberon.net> <20050327115142.2cf6e3b3@Magellan.Leidinger.net> <20050327095541.GB16502@voodoo.oberon.net>
next in thread | previous in thread | raw e-mail | index | archive | help
On Sun, 27 Mar 2005 11:55:41 +0200 Kirill Ponomarew <krion@voodoo.oberon.net> wrote: > Uuuh, I think you know that we don't commit bsd.port.mk without > running it on cluster first. I think it can be delayed and tested > after the freeze. So far the dependency list is broken since some ports don't register their dependency on the default (or an alternative) linux_base (for a list of affected ports have a look at http://www.leidinger.net/FreeBSD/USE_LINUX2.txt). So the question is: do we want to ship 5.4-release with a broken dependency information? If not we need to backout the commit in question. When we backout Trevor's commit, the users of an alternative linux_base may get an inconsistent pkgdb (there's no linux_base port to register dependencies to). This can be fixed by the patch to bsd.port.mk. I know the actual verification process involves a full run on the cluster. But when you look at the patch you will notice that it only will be considered to affect anything, when USE_LINUX in a port is used (lang/icc still contains the USE_LINUX, so you can use it to test the patch). And it only affects the USE_LINUX case, if OVERRIDE_DEFAULT_LINUX_BASE is defined, so for the default case (of not having OVERRIDE_DEFAULT_LINUX_BASE defined, like on the ports build cluster) it does nothing. Yes, this isn't a formal prove, but as long as our make isn't broken, it's more than a run on the cluster is able to verify. To verify the remaining case (defining OVERRIDE_DEFAULT_LINUX_BASE to a valid and an invalid value) you just need to cd to lang/icc and run: make -V USE_LINUX -V RUN_DEPENDS make -V USE_LINUX -V RUN_DEPENDS OVERRIDE_DEFAULT_LINUX_BASE=foobar make -V USE_LINUX -V RUN_DEPENDS OVERRIDE_DEFAULT_LINUX_BASE=rh-9 (it also works to define it in make.conf). I want to work on this regardless of the decision of portmgr, so if you decide to keep the actual broken dependency list for the release, I will come back to you after the freeze. If you approve the backout but not the bsd.port.mk change, I will show up after the freeze with the bsd.port.mk patch. Regardless of everything above: what about the plist problems Kris' pointed out on emulation@? Do you want to see patches or do I have implicit approval to fix those plist issues? Bye, Alexander. -- Where do you think you're going today? http://www.Leidinger.net Alexander @ Leidinger.net GPG fingerprint = C518 BC70 E67F 143F BE91 3365 79E2 9C60 B006 3FE7
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20050327121846.518f0772>