I agree with Cos that a suffix makes it easier to find.
But I would lean toward "-bigtop" since "-asf" may be vague in some
cases (ie. what if Apache Hadoop starts to also name its own packages
hadoop-asf?).
Using "-bigtop" also makes the brand/project more recognizable and
easier to look for in search engines.
Thanks,
Bruno
On 09/23/2013 06:16 PM, Konstantin Boudnik wrote:
Sorta what I suggested with 'spark-asf'. It'd better than 'apache-' prefix,
because it would be easier to find, sorta...
Cos
On Mon, Sep 23, 2013 at 04:37PM, Peter Linnell wrote:
Putting on my Linux distro hat... I've long been involved with
openSUSE...
I'd say apache-spark would be best, but this adds minor baggage to
packaging.
Thoughts ?
Thanks,
Peter
On Mon, 23 Sep 2013 16:01:32 -0700
Konstantin Boudnik <[email protected]> wrote:
On Mon, Sep 23, 2013 at 06:24PM, Jay Vyas wrote:
Why not call it bigtop-spark .. ?
Well, because it isn't bigtop's spark, to start with ;)
Seems to me like nobody but the root project itself should
monopolize the hadoop distro primary names since there are so many
builds out there... Right?
What's is root project, sorry?
On Sep 23, 2013, at 4:11 PM, Konstantin Boudnik <[email protected]>
wrote:
I'd say Fedora needs to deal with it on their end, considering
that they are coming into this post-Bigtop. Thoughts?
On Mon, Sep 23, 2013 at 11:43AM, Sean Mackrory wrote:
I've been worried about this for a while, and I think we need to
come up with a strategy for dealing with it in Bigtop. In this
case, it's just a coincidence that two separate projects are
named Spark. However I recently submitted a patch for packaging
Avro as a separate component in Bigtop and had to deal with the
fact that a minority of the systems we support already packaged
part of Avro in a package named 'avro', so I had to use
'avro-libs', 'avro-tools' and 'avro-doc'. Fedora's also going to
start releasing packages for Hadoop, so this problem of name
conflicts is only going to get worse.
On Mon, Sep 23, 2013 at 10:52 AM, Konstantin Boudnik
<[email protected]>wrote:
Let's call it spark-asf, perhaps?
On Mon, Sep 23, 2013 at 09:57AM, Roman Shaposhnik wrote:
Hi!
While testing Spark as part of the Bigtop I've come
across the following conflict on Debian-based systems:
$ apt-cache show spark
Package: spark
Priority: optional
Section: universe/devel
Description-en: SPARK programming language toolset
SPARK is a formally-defined computer programming language
based on the Ada programming language, intended to be secure
and to support the development of high integrity software used
in applications and systems where predictable and highly
reliable operation is essential either for reasons of safety
or for business integrity. .
This package contains the tools necessary for checking if
programs
adhere
to the SPARK rules and the tools to show freedom of runtime
exceptions
in
those programs. To compile SPARK programs use any
standards-compliant
Ada
compiler, such as GNAT.
Homepage:
http://libre.adacore.com/libre/tools/spark-gpl-edition/
This has an obvious ramifications for Bigtop packaging, but
it may also have ramifications for PODLINGNAMESEARCH
for Apache Spark (incubating).
Speaking of which -- has anybody else started PODLINGNAMESEARCH
already?
Thanks,
Roman.
