Re: compilation of gnome-1.4-rc1

[Gregory Leblanc <gleblanc cu-portland edu>]

On 19 Mar 2001 12:13:28 -0600, Ade Lovett wrote:
> On Mon, Mar 19, 2001 at 09:57:36AM -0800, Gregory Leblanc wrote:
> > Say WHAT?  scrollkeeper is not linux-only by -any- means.  In fact, Nik
> > (FreeBSD docs guy) has been quite helpful in designing scrollkeeper.
> > What ever made you think it was linux-only?
> 
> It is in the placement of it's database.  /var/lib does not exist
> on FreeBSD, and is in violation of the hierarchy specification
> (man 7 hier on any FreeBSD box).  One possible option would be to
> use FreeBSD's /var/db, but there is no method for removing the 'lib'
> component, so it would appear as /var/db/lib/scrollkeeper.

This should be pretty trivial to fix.

> The fact that you can only set the top level location (/var by default),
> and that lib/scrollkeeper is automatically appended is just plain wrong.
> 
> I've been talking to one of the authors about this.  IMO, the correct
> approach would be to not derive things from ${localstatedir} but to
> have an overrideable --with-dbdir=... configure option.  Add a --dbdir
> to scrollkeeper-config, then it simply doesn't matter where the database
> gets installed.

<flame>Well, as far as I'm concerned you haven't contacted the
developers, because I haven't seen anything come through the
scrollkeeper-devel mailing list.</flame>  Please don't be offended, I
beat everybody up who doesn't communicate about this stuff on public
mailing lists. 

> In addition, ignoring all of the above, the port breaks at install
> time assuming that (I believe, it's been a few days) scrollkeeper-config
> is directly in the $PATH which may not be the case - I use chrooted
> environments for each individual port/package build to ensure that
> the application puts things in the right place.
>
> Just because something compiles on FreeBSD, does not automatically mean
> that it works, subject to the constraints that I, and other members
> of the ports team, have to work under, and take very seriously.

Hmm, I wasn't aware of this, and I'm sure that Laszlo and Dan weren't
either.  Are these constraints well documented and published someplace
so that we can read them?

    Greg