Just remarks

Marc Heerdink marc at linuxfromscratch.org
Mon May 27 04:52:31 PDT 2002

Op ma 27-05-2002, om 00:22 schreef Mark Hymers:
> On Sun, 26, May, 2002 at 02:37:23PM -0400, Gerard Beekmans spoke thus..
> > Actually the times are the average of a few systems (533 mhz, 750 mhz, 1ghz
> > and another one). They are going to be redone very soon: as soon as the
> > current batch of waiting package upgrades are going into the book (most
> > notably the gcc-3.1 upgrade) we'll redo all of the times and disk space
> > requirement, and this time list what systems are used. People with
> > different configurations can then submit their own times to be taken into
> > account.
> Actually, I'm still interested in the concept of what used to be called
> "ed-times" but would now more logically be "static-bash-times".  I.e.
> instead of giving a specific time, make it a multiple of this.
> Did we ever look to see if this produced consistent results?  If not,
> next time I do a build I'll calculate it for me and then we can try
> asking a few people to submit them and see if it works.

Hmmz I doubt that there's a need for build times anyway. There's
probably nobody who gets the same build times as the book if it's an
avarage of many systems. Relative build times seem a lot more useful to
me, but can't we just say "glibc takes a looooooong time to compile"
where the number of o's is a measure for the time required? :o) j/k of
course, but build times are so machine-dependant (even two identical
machines probably produce different results) that they can be considered
non-information imo.

  / |    ||         ..."Have you mooed today?"...
 *  /\---/\   Marc Heerdink <marc at linuxfromscratch.org>
    ~~   ~~

Unsubscribe: send email to listar at linuxfromscratch.org
and put 'unsubscribe lfs-dev' in the subject header of the message

More information about the lfs-dev mailing list