Am 04.07.2018 um 22:46 schrieb Kay Schenk:
On Wed, Jul 4, 2018 at 1:00 PM, Marcus <marcus.m...@wtnet.de> wrote:
Am 04.07.2018 um 08:23 schrieb Peter Kovacs:
I think Jim is referring to the gstreamer situation, where we decided
that we skip CentOS6 more or less for 4.2.0.And one argument was, if they
want something they should support us. This is not showing sympathy for a
small user group that uses very old software for 2 more years until they
have to move to CentOS 7. I personally think that the gstreamer Topic can
be solved after we have released a beta version. Damjan and I have pointed
out a lot of possible ways to deal with the issue. Just for now I think we
have other problems then gstreamer in 4.2.0. I think it is my fault that I
put that argument so much in the front line, but that stuck for me.
In the last months we had a drop in activity. And more then one topic
received not the attention it deserved. I would not conclude that anyone
has stopped caring at this point in time.
Let us conclude for now:
4.1.x is still in maintenance. And in my opinion we could think of
maintaining it until 2020 when CentOS6 drops out of maintenance. Some
support from CentOS6 side would be nice. But we need to search someone for
this.
I have that on my todo list, but did not manage to follow it up.
incl. gstreamer 0.1.0 that is now within the 4.1.x code.
PS:
CentOS 6 will be supported until Nov 2020; which means another ~2.5 years.
4.2.0 has I think 3 bugs we know about and that blocks a beta release.
Current target for building with gstreamer is CentOS7. Building without
gstreamer could be done on CentOS6. We should keep the code in trunc CentOS
6 compatible where ever we can for now. That will make it easy to back port
patches to 4.1.x if we decide to maintain 4.1.x until EOL of CentOS6.
In 4.2.0 we can still keep gstreamer 0.1.0 or update to something newer.
To be honest, I don't care *about this special topic*.
And it is only relevant on Linux, right?
IMHO more relevant is the baseline: When we increase the CentOS version we
also raise the sysreq for Linux kernel, glibc, etc. This has a much bigger
impact for our users.
You are absolutely correct about this, Marcus. Monitoring the 32-bit Linux
downloads might help here. It does seem like AOO could be moving away from
32-bit for Linux and other operating systems. I don't know what impact this
will have overall though.
I don't remember exactly, does the gstreamer 0.1.0 vs. 1.0.0 discussion
is also connected to the Linux 32-bit builds? If so, a solution could be
indeed to drop the 32-bit builds. From SF.net stats I get the following
(2018-01-01 until today).
BTW:
Very likely it's the used OS the download is started from. And not the
OS where OpenOffice should be installed on.
OS %
-----------------------
Windows 86,1165
Macintosh 7,8424
Unknown 4,9012
Linux 1,0621
Android 0,0762
BSD 0,0011
Solaris 0,0006
But even then, I'm sure the most downloads from resp. for Linux will be
for 64-bit.
Has anybody more exact numbers - or an idea how to get them?
Marcus
On 03.07.2018 23:50, Matthias Seidel wrote:
What impact has Ant 1.10.x exactly on older machines?
It is no problem for me to build the Windows version with Ant 1.9.12. As
long as we use Java 8.
But again, I just did a personal build to test AOO 4.1.x with Java 8.
Nothing else.
To be more precise: I was the only one who cared. No response from other
members!
Am 03.07.2018 um 23:19 schrieb Jim Jagielski:
The above made it appear that Ant 1.9.x was no longer supported plus
had some sort of security related issue making it unsuited for AOO... ie,
we *needed* to use Ant 1.10 not just that we now *can* use it.
How about showing some sympathy and understanding for those who may be
stuck w/ older machines? After all, let's be real, our continued support
for "older" systems is the only real thing we have going for us... It's
these little things that make significant ripples in our eco-system and we
seem to not really care about that anymore.
On Jul 3, 2018, at 4:02 PM, Matthias Seidel <matthias.sei...@hamburg.de>
wrote:
Am 03.07.2018 um 21:45 schrieb Jim Jagielski:
On Jul 1, 2018, at 11:27 AM, Peter Kovacs <pe...@apache.org> wrote:
Hi everbody.
I would like to bring a 4.1.6 Release on the way in July. Even if we
manage to get 4.2.0 ready it will only be a beta. And we have some stuff to
get out to the people.
Matthias has created a suggestion for a 4.1.6 release on security.
Containing some security fixes, plus
- Java 8 Update 172
- Apache Ant 1.10.3
What is wrong w/ Apache Ant 1.9.12? Why the need for 1.10.x?
What is wrong with Ant 1.10.x? If we build with Java 8 we can use
it... ;-)
My test build was just a Proof-of-Concept what can be done with AOO
4.1.x.
But of course we can build with 1.9.x if that is wanted?
Regards,
Matthias
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org
For additional commands, e-mail: dev-h...@openoffice.apache.org
