On 08/29/2014 02:52 AM, Yair Zaslavsky wrote:
Thanks everyone for your thoughts.
I would like to sum things up (as I understood from this thread) -
a. We will defer the move to commons collections4.
b. We should introduce some class renaming, not have LinqXXX
I think those renaming and general refactoring of compat should be a
series of bugzilla's we should communicate as
entry-level,low-hanging-fruits,you-name-it
for new comers to ovirt.
c. Later on we can shift to an "already maintained" package.
I wonder how close we are to moving to java8 where all these
dependencies (commonsX, LinqWhatever) could go to the waste bin.
probably a mix of jboss/rhel/gwt.
----- Original Message -----
From: "Vojtech Szocs" <[email protected]>
To: "Greg Sheremeta" <[email protected]>
Cc: "Yair Zaslavsky" <[email protected]>, [email protected]
Sent: Thursday, August 28, 2014 4:49:57 PM
Subject: Re: [ovirt-devel] [ENGINE] thoughts about LinqUtils
----- Original Message -----
From: "Greg Sheremeta" <[email protected]>
To: "Yair Zaslavsky" <[email protected]>
Cc: [email protected]
Sent: Monday, August 25, 2014 1:52:44 PM
Subject: Re: [ovirt-devel] [ENGINE] thoughts about LinqUtils
----- Original Message -----
From: "Yair Zaslavsky" <[email protected]>
To: "Alon Bar-Lev" <[email protected]>
Cc: "Greg Sheremeta" <[email protected]>, "Itamar Heim"
<[email protected]>, [email protected]
Sent: Sunday, August 24, 2014 9:51:31 PM
Subject: Re: [ovirt-devel] [ENGINE] thoughts about LinqUtils
----- Original Message -----
From: "Alon Bar-Lev" <[email protected]>
To: "Greg Sheremeta" <[email protected]>
Cc: "Yair Zaslavsky" <[email protected]>, "Itamar Heim"
<[email protected]>, [email protected]
Sent: Friday, August 22, 2014 11:36:31 PM
Subject: Re: [ovirt-devel] [ENGINE] thoughts about LinqUtils
----- Original Message -----
From: "Greg Sheremeta" <[email protected]>
To: "Alon Bar-Lev" <[email protected]>
Cc: "Yair Zaslavsky" <[email protected]>, "Itamar Heim"
<[email protected]>, [email protected]
Sent: Friday, August 22, 2014 11:33:09 PM
Subject: Re: [ovirt-devel] [ENGINE] thoughts about LinqUtils
----- Original Message -----
From: "Alon Bar-Lev" <[email protected]>
To: "Greg Sheremeta" <[email protected]>
Cc: "Yair Zaslavsky" <[email protected]>, "Itamar Heim"
<[email protected]>, [email protected]
Sent: Friday, August 22, 2014 4:20:19 PM
Subject: Re: [ovirt-devel] [ENGINE] thoughts about LinqUtils
----- Original Message -----
From: "Greg Sheremeta" <[email protected]>
To: "Yair Zaslavsky" <[email protected]>
Cc: "Itamar Heim" <[email protected]>, [email protected], "Alon
Bar-Lev"
<[email protected]>
Sent: Friday, August 22, 2014 8:39:54 PM
Subject: Re: [ovirt-devel] [ENGINE] thoughts about LinqUtils
----- Original Message -----
From: "Yair Zaslavsky" <[email protected]>
To: "Itamar Heim" <[email protected]>
Cc: [email protected]
Sent: Thursday, August 21, 2014 8:39:31 PM
Subject: Re: [ovirt-devel] [ENGINE] thoughts about LinqUtils
----- Original Message -----
From: "Itamar Heim" <[email protected]>
To: "Yair Zaslavsky" <[email protected]>, "Yevgeny
Zaspitsky"
<[email protected]>
Cc: [email protected]
Sent: Friday, August 22, 2014 12:25:52 AM
Subject: Re: [ovirt-devel] [ENGINE] thoughts about LinqUtils
On 08/21/2014 09:55 AM, Yair Zaslavsky wrote:
----- Original Message -----
From: "Yevgeny Zaspitsky" <[email protected]>
To: "Yair Zaslavsky" <[email protected]>
Cc: "Moti Asayag" <[email protected]>, "Allon Mureinik"
<[email protected]>, [email protected]
Sent: Thursday, August 21, 2014 4:35:33 PM
Subject: Re: [ovirt-devel] [ENGINE] thoughts about
LinqUtils
On 21/08/14 12:08, Yair Zaslavsky wrote:
----- Original Message -----
From: "Yevgeny Zaspitsky" <[email protected]>
To: "Moti Asayag" <[email protected]>
Cc: "Yair Zaslavsky" <[email protected]>, "Allon
Mureinik"
<[email protected]>, [email protected]
Sent: Thursday, August 21, 2014 11:26:40 AM
Subject: Re: [ovirt-devel] [ENGINE] thoughts about
LinqUtils
It seems like we can try moving to common-collections4.
Yum
on
my
Fedora20
computer finds apache-commons-collections4 package.
Fortunately
somebody
packed the jar into for a rpm for us. :-)
What about RHEL 6.5? Can you please run a quick check?
Unfortunately my happiness was too hasty. Only Fedora
people
care
to
be
in the forward of the technology... The RHEL ones do not
care
about
that...
This is what I remembered. When you responded to the email
for
the
first
time , I had a strong deja vu that you tried addressing
this
issue
yourself in the past (commons-collectios4) - due to
different
reason.
is there a specific conflict or problem (or a huge chain of
dependencies)
?
To me it seems the answer to both is no -
This is the requirement list -
java >= 1.5
jpackage-utils
rpmlib(CompressedFileNames) <= 3.0.4-1
rpmlib(FileDigests) <= 4.6.0-1
rpmlib(PayloadFilesHavePrefix) <= 4.0-1
rpmlib(PayloadIsXz) <= 5.2-1
Probably a matter of packaging?
IIRC, Alon was the one who replied, and the issue was that Jboss
included
an
old version (and we don't have classpath isolation, I guess)
Greg
We would like to avoid maintaining and package components that are
not
provider either by el6 or jboss distribution.
But based on other threads, it seems that I am the only one who
remained
trying to push compliance to the old ways, people feel that can
maintain
anything anywhere with no effort.
Regards,
Alon
Alon, I disagree with your comment (about the "you're the only one" part
:)
)
+1
We have three (four if you include PatternFly) ongoing threads about
dependency
issues at the moment, and I hope we all realize that Alon is trying to do
what's best for our project. I certainly empathize with him. He has a tough
role, and there are a lot of us young'uns who want 'shiny new things'
brought
into the project. I certainly don't have the experience to know about all
the
long term costs of bringing in dependencies into an enterprise project like
this
-- but I'm learning :)
I don't think that motivation to introduce new dependencies is driven by
desire to have "shiny new things" (we're not kids, right?) - I think that
motivation is driven by actual needs, backed by potential value that might
be broght in. For example, better/easier code due to newer version of
library.
I agree that we should avoid maintaining packages ourselves as much as we
can,
I think that everyone's in agreement with Alon on that.
As I wrote - I had a strong deja-vu about that the issue was already
brought
up.
Now that you reminded , I don't think you're the only person who feels
this
way.
I would also like to understand more what it means before jumping to
conclusions and upgrading to collections4.
At past I had some issues with another commons project
(commons-configuration) that had different versions upstream and
downstream.
I think collections4 is a nonstarter because it's not packaged for EL,
IIUC.
I am sure the changes include not just
"move to generics" and should carefully be considered.
If I may clarify, there would be at least two stipulations for
introducing
collections4.
1. someone else packages it and maintains it, available in Fedora and
EL,
long term. Quality package.
this is what missing, us maintaining a new package just to have more
beautiful code is something that can be deferred for now.
2. JBoss has proper classloader isolation so that, even though JBoss
uses
collections3, a webapp can use collections4.
should not be a problem to use both.
I don't know the answer to either question :)
Seems like minimal gain to me, though.
Greg
_______________________________________________
Devel mailing list
[email protected]
http://lists.ovirt.org/mailman/listinfo/devel
_______________________________________________
Devel mailing list
[email protected]
http://lists.ovirt.org/mailman/listinfo/devel
_______________________________________________
Devel mailing list
[email protected]
http://lists.ovirt.org/mailman/listinfo/devel
