Date: Sun, 20 Mar 2005 00:05:04 -0300 From: Fernan Aguero <fernan@iib.unsam.edu.ar> To: Gerald Pfeifer <gerald@pfeifer.com> Cc: ports@FreeBSD.ORG Subject: Re: problems fetching lang/gcc*: random? Message-ID: <20050320030504.GG79230@iib.unsam.edu.ar> In-Reply-To: <Pine.BSF.4.62.0503192221290.36661@acrux.dbai.tuwien.ac.at> References: <20050317131405.GA31049@iib.unsam.edu.ar> <Pine.BSF.4.62.0503192221290.36661@acrux.dbai.tuwien.ac.at>
next in thread | previous in thread | raw e-mail | index | archive | help
+----[ Gerald Pfeifer <gerald@pfeifer.com> (19.Mar.2005 18:29): | Dear Gerald, | On Thu, 17 Mar 2005, Fernan Aguero wrote: | > I'm having a weird problem trying to fetch lang/gcc* ports | > using the ports system. | | I don't see how the problem you are experiencing should be related | to these specific ports per se... I don't know either, but at least now I know a little bit more ... | > This does not happen with other ports (at least not with the | > ones I tested, and with the latest gnome upgrade, I've been | > fetching lots of them), so I don't believe | > the cause to be my network connectivity or fetch(1). This is | > the first reason for this weirdness. | | ...but perhaps it might be related to the specific servers you are | using when downloading the distfiles for these ports? it might, since I have RANDOMIZE_MASTER_SITES=yes in my /etc/make.conf | > In all cases this is as far as I get: | > | > pi# pwd | > /usr/ports/lang/gcc34 | > pi# make fetch | > Making GCC 3.4.4 for FreeBSD 4.11 target=i386-portbld-freebsd4.11 | > => gcc-core-3.4-20050311.tar.bz2 doesn't seem to exist in /usr/ports/distfiles/. | > | > at the same time CPU usage raises considerably and 75-80% of | > this corresponds to a process named random, which ps(1) shows | > as /usr/games/random. As soon as I Ctrl-C and quit the | > fetch, this process disappears and CPU usage return to | > normal levels. This is the second reason I found this weird. | > | > I'm completely intrigued by this. Anyone can reproduce it, | > or better yet have an idea of what is going on? | | Ah, wait, this might be a very useful piece of information and indeed | triggers at least one idea: try renaming /usr/games/random, run rehash, | and then give "make fetch" another try. Does that make any difference? | | Gerald | | PS: Please keep me Cc:ed on this thread. | +----] yes, renaming /usr/games/random did it. No need to rehash. Now the distfiles are being downloaded. Still I'm intrigued. Why this doesn't happen with other ports? Why /usr/games/random? Well, seems like it's used by bsd.port.mk when you define RANDOMIZE_MASTER_SITES. In this case, the master sites list is piped to '| random -w -f -'. The raise in the CPU usage makes me think that the list might be monstrous. However, I've run a 'make fetch' overnight once and it never got to fetch anything, so it seems like it never finished. Of course, the list in MASTER_SITE_GCC is not that big, so it should be something else. But that something seems to be in the gcc* ports. Any ideas anyone? Fernan
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20050320030504.GG79230>