cvs commit: doc/en_US.ISO8859-1/books/faq book.sgml (fwd)

[Oliver Eikemeier]

Am Samstag den, 12. Juni 2004, um 04:06, schrieb Mark Linimon:

> FYI.  This should be publically available in the next 24 hours.
>
> I do not think it will be controversial so I skipped the pre-review,
> but I won't object if it needs to be reworded.

You should call it "why I should not try to build the INDEX and
don't complain when I'm failing to do so".

Just a joke. Honestly, great job. You might want to mention
downloading the INDEX as an alternative (there is a fetchindex
target in the queue, otherwise you could always do
   fetch -o /usr/ports/INDEX http://www.freebsd.org/ports/INDEX
on -STABLE) and possible problems someone can run into when
installing the ports tree from a release CD and CVSup it afterwards,
   find /usr/ports -name README.html -delete
should help here (not tested, ask lofi for more input).

Also, people should be told to check whether they have an *up-to-date*
*complete* ports collection twice (possibly with the explanation _why_
moved to an extra clause, at least I had problems following it which
may be my fault). The most common problems to look for are
having some language categorie in the CVSup refuse file, which either
results in not having this categorie at all (easy to spot) or this
part of the tree to be out of date and excluded from updating (harder
to spot, because stuff does not fail immediately). I know this is
mentioned in the article, but it can't hurt to mention the most common
case, since this is a FAQ.

Generally one should ask himself why one needs to build the INDEX
(instead of just downloading it), and, if deciding to do so, consider
this to be part of a sanity check of the local ports tree.

Just my 2 ¢ (0.02€)
-Oliver