ben at algroup.co.uk
Fri Oct 31 02:40:37 PST 2003
Christopher Vance wrote:
> On Thu, Oct 30, 2003 at 03:19:37PM +0100, Emiel Kollof wrote:
>> that I want to see portupgrade go away. Although it serves a need, it's
>> a need that needs to be serviced by the packaging system. And the
>> packaging system should be able to do a better job than portupgrade does
> Indeed. I use portupgrade on FreeBSD, and it seems to "work", but it
> rankles that you need to install ruby. (Sort of like cvsup and
> Modula-3, I guess.)
> Does portupgrade do all its work in dependency order, so that if A
> depends on B then B is updated before A? I don't believe it does, but
> could be convinced otherwise.
Yes, it does, so long as you give it the right flags.
The trouble with portupgrade is it relies on ports, and they seem to
break it with monotonous regularity. For example, sometime a month or
two ago, something was done to ports that caused portupgrade to blow
away Xfontlib and never put it back. Manual intervention was required.
Similarly, name changes confuse the crap out of it.
Also, never make the mistake of installing anything from source and then
updating with a port. I did this once with X and the result was a
machine so hosed the easiest option was to blow it away and start again.
I am not a fan of ports. portupgrade is a nice idea, but it needs
something better underneath it.
"There is no limit to what a man can do or how far he can go if he
doesn't mind who gets the credit." - Robert Woodruff
More information about the Kernel