On 11/10/11 4:09 PM, Pedro Giffuni wrote:
Hello Jürgen;
I value your feedback on this issue, and I will explain:
First of all agg as it is/was is not an IP threat and, in
general, the rules on how to deal with all the other
dependencies were not in place when I did the update anyways.
1) I updated it to version 2.4: this is the last version
under a BSD license. It was supported by the internal build
system so I doubt anyone complains about having it up to
date.
my main concern here is why you haven't updated the tar file with a
newer version and used the same mechanism as for all other 3rd party libs.
2) I disabled it by default simply because it's not really
used in the build. I doubt anyone complains about having off
by default something that is not used.
I also learned about, and killed, a header that supported
the GPC extension which is not AL2 compatible,
I cannot say this has brought any benefit at all but there's
nothing counterproductive as agg was never really productive.
ok, drop counterproductive but i still don't understand why you have
checked in it at all. The update if necessary could have been done at a
later time as well.
Now about --enable-system-agg; this option is a no-op as all
linux/BSD distributions, that I know of, carry version 2.5
of agg which is explicitly prohibited in OOo. This was done
by SUN, not by me, perhaps because it's GPL'd now or maybe
do to API changes, but it doesn't look like it.
why not analyzing if possible to use it? As it is optional
(default=disabled) anyway it would be much easier.
I kept agg around because I think it's important to keep the
last BSD-licensed version in SVN (if we remove it we can bring
it back anytime) and because it may find some uses elsewhere
(anyone in need of a C++ rendering engine, like for SVG? ;-) ).
Further cleanage of the configure script (which I hate to
manipulate to tell you the truth) or even removing agg is
relatively easy.
you can always ask if others can help. I can think also about much more
interesting stuff but some things have to done at the moment ;-)
Juergen
Pedro.
--- On Thu, 11/10/11, Jürgen Schmidt<jogischm...@googlemail.com> wrote:
wrote:
Hi Maho;
I personally didn't plan to remove agg. I like it as
option and license wise it's OK so I have no pland to
remove it, at least for now.
well the question really is why you have checked it in this
way and has
disabled it directly afterwards. That doesn't make any
sense and it was
counterproductive from my pov.
It was initially handled as all other 3rd party modules and
we could
have dropped it and could have kept the option to use it
(if somebody
wants) with the option --with-system-agg. Default would be
to disable it.
I think it would make sense if we follow all the same
rules.
Juergen
I did notice it's still getting built on our port and
I have to look at why. I think when I attempt to
build
AOOo from the tarball it doesn't get built but there
are other ugly issues with icu there.
About epm I don't know, I guess we can remove that
directory now.
Pedro.
--- On Wed, 11/9/11, Maho NAKATA<m...@apache.org>
wrote:
From: Maho NAKATA<m...@apache.org>
Subject: agg and epm are still in svn repo.
To: ooo-dev@incubator.apache.org
Cc: p...@apache.org
Date: Wednesday, November 9, 2011, 7:47 PM
Hi,
while porting AOOo to FreeBSD, pgf@ noticed that
agg and epm are still in svn repo.
is it correct? Should we remove them?
thanks
Nakata Maho
