Re: [gentoo-user] Re: Openoffice being replaced?

[Alex Schuster]

Dale writes:

> Alex Schuster wrote:
> > Dale writes:

> >> But if you emerge something and it has to be fetched first, is that
> >> counted in the time genlop shows or not?  That is the question.  I
> >> don't think it is counted but I'm not sure.
> >
> > That's what I thought, too, but then I simply tried to be sure.
> > Download time _is_ counted.
>
> Now we know.  If I was on dial-up again, I could sure test that theory.
> 3KBs/sec would certainly make a difference.  :-(  Pardon me if I refuse
> to go back tho.  I like youtube to much.

It's easier than that, I simply emerged vanilla-sources-3.0 after deleting 
the tarball. I did not use my digital wrist-watch which I could have done, 
instead I looked into emerge.log. The long numbers to the left are seconds 
since 1970, the difference is what genlop uses. The only question was 
whether it uses the line 'emerge (x of y) category/package-version to /' or 
'(x of y) Compiling/Packaging ...' to determine the start time.

> >> I set mine to fetch in the
> >> background so most of the time the fetch is done after a couple
> >> packages gets compiled.
> >
> > What about parallel emerges? I guess genlop will not take this into
> > account.
>
> I would think not.  As long as the tarball is downloaded before emerge
> gets to it to compile.  I doubt it would even know how long it took to
> download either.

I wasn't talking about the download time here, but about the total time. If 
I emerge two independent packages A and B, which take one hour each to 
build, what does genlop say if I use emerge -j and they build in parallel? 
This would take about two hours for each. And indeed, that's what genlop 
says. So genlop -t is inaccurate when you are emerging with the -j option.

	Wonko