CVSup and PortUpgrade

martin at rakupottery.org.uk martin at rakupottery.org.uk
Mon Sep 25 22:24:35 BST 2006


James Elstone writes: 

> Hi All! 
> 
> I have been dabbling in FreeBSD for a few months now and have been having 
> the time of my life, having headed into FreeBSD from the embedded route so 
> the novelty of non-Intel hardware has not yet worn off! 
> 
> However, I have a few questions relating to CVSup and PortUpgrade and 
> installing in general though which I am finding difficult to answer, and 
> would like to help. Remember I am a newbie in this at the moment, so 
> excuse the possible silly questions or possibly obvious answers: 
> 
> 1).  I have managed to successfully obtained the Ports collection by using 
> the CVSup method as described in the forever useful handbook.  This took 
> about 30 minutes and seemed to download the source code for all current 
> ports that are available.  Is my assumption that I have just downloaded 
> the source code for the ports correct?

No, you have installed the skeleton of the ports tree, with all the 
makefiles and other little informative items, if you had downloaded all the 
sources for the entire ports tree you might still be doing it now :)
When you run make in the port directory of your choice it will grab the 
source code from a site mentioned in the makefile. 

> 
> 2). When I installed the Portupgrade package it downloaded from the 
> ftp.freebsd.org some other "packages" in a tarball for Ruby.  I installed 
> Portupgrade by using "make install clean" from the 
> /usr/ports/sysutils/portupgrade directory.  If I have downloaded all the 
> source code in the ports structure when I executed CVSup, why does it not 
> use the already downloaded code, and compile from it rather than 
> downloading what seemed to be a packaged tar ball? (It kept mentioning 
> that the /usr/distfiles (or something similar) were not present and that 
> it had to download them...)

see my note above, you havent got all the sources. 


> 
> 3). Is there a difference between a packaged tar ball (.tar.gz) that is 
> downloaded in the above example or source code from the /usr/ports file 
> structure? 
> 
> 4). How once a binary is installed can I reference it at the command line 
> without specifying the full path, e.g. "portupgrade" rather than 
> "/usr/local/etc/portupgrade".  I suppose what I am asking is how do I 
> alter the global PATH setting?

After compiling a port, if you are using csh or tcsh, you need to type 
rehash to get the newly install program recognised, you may also need to 
look at your path definition as well 

> 
> 5). Why when I am in a directory can I not just enter the binaries name, 
> but instead I have to prefix "./" to it before the command is recognised, 
> and is there a way to alter the behaviour so the pwd is part of the path? 
> (I remember from Uni days there is a way to do this, but cannot for the 
> life of me remember how!) 
> 
> 6). The reason I ask about the downloading of tarballs above is that the 
> machine I am working on will eventually only have access to an up_to_date 
> /usr/ports tree (via nfs/ftp to another internal FreeBSD node that has 
> internet access and is on a DMZ?!) and no direct access to the internet.  
> I am trying to understand why the ports collection is not enough on it's 
> own, and how I can overcome this issue!?! 
> 
> 7). When Ruby installed it took all night and some of the day to "make".  
> Is there anyway I can increase the performance on this? The machine in 
> question is currently a PII-633MHz with 128Mb ram with a 20GB HD. Any 
> ideas?

more ram would help, as would a faster processor 

 


-- 
Martin Smith 





More information about the Ukfreebsd mailing list