<html>
<head>
<meta content="text/html; charset=UTF-8" http-equiv="Content-Type">
</head>
<body bgcolor="#FFFFFF" text="#000000">
Hi Fred and all JBDS team!<br>
<br>
The idea of "merge data from separate yaml files" doesn't worked and
we (JDF team) didn't put any more effort on it.<br>
<br>
The original idea is that we could join the information (appending a
new content to the original stacks.yaml) and use its YAML anchors.
To clarify what I'm saying, look the matrix.yaml file
(<a class="moz-txt-link-freetext" href="https://github.com/jboss-jdf/jdf-stack/blob/1.0.0.Final/matrix.yaml">https://github.com/jboss-jdf/jdf-stack/blob/1.0.0.Final/matrix.yaml</a>)
it has some references to BOMs and archtypes that is defined on
original stacks.yaml.<br>
<br>
The main problem is that the new generarate yaml file is not
parsable via StacksClient (because it expects the original 1.0.0
format only) and a new "client" should be built to each new format
extension. This could make us loose the control since any data
change on stacks.uaml could impact all extensions that uses it. So
in fact: It was not a good idea to have this extensions on separate
files. <br>
<br>
The next "Stacks 1.0" planned evolution is
<a class="moz-txt-link-freetext" href="https://issues.jboss.org/browse/JDF-222">https://issues.jboss.org/browse/JDF-222</a> -
<meta charset="utf-8">
Stacks should permit a "Early Access" Runtimes - And because of the
file format, this "Early Acccess" will be a new "label" but I have
afraid that consumers must know that "allAvailableRuntimes" has
"Released/Final" and "Early Access" runtimes mixed. They should
query the labels to distinguish them.<br>
<br>
That's my 2 cents.<br>
<br>
<br>
<div class="moz-cite-prefix">Em 26/03/13 10:25, Fred Bricon
escreveu:<br>
</div>
<blockquote cite="mid:5151A1DD.2050009@redhat.com" type="cite">Hey
Rafael,
<br>
<br>
what's the status on the "merge data from separate yaml files"? I
know you worked on that at some point.
<br>
We'll most certainly need to combine JBoss Tools specific runtimes
or infos with the existing JDF ones.
<br>
<br>
Fred
<br>
<br>
Le 02/03/2013 08:52, Rob Stryker a écrit :
<br>
<blockquote type="cite">license, size, and disclaimer can all be
added in the labels section of
<br>
the yaml.
<br>
<br>
Max has already made it clear that having three things that do
the same
<br>
thing is wasteful and confusing, so the question isn't *if* we
will
<br>
unify them, but rather *how*.
<br>
<br>
On 03/01/2013 09:29 PM, Snjezana Peco wrote:
<br>
<blockquote type="cite">I think we need to keep all of those
options (since they already
<br>
exist). Each of them has its advantages/disadvantages.
<br>
I agree with the proposed changes and emphasize that stacks
don't have
<br>
the following properties:
<br>
<br>
- license
<br>
- size
<br>
- requireSso (see
<a class="moz-txt-link-freetext" href="https://github.com/jbosstools/jbosstools-base/pull/50">https://github.com/jbosstools/jbosstools-base/pull/50</a>)
<br>
- disclaimer
<br>
<br>
Snjeza
<br>
<br>
On 3/1/2013 10:22 AM, Rob Stryker wrote:
<br>
<blockquote type="cite">I would like to hear some feedback
here from Max and Snjezana.
<br>
<br>
On 02/28/2013 08:05 PM, Fred Bricon wrote:
<br>
<blockquote type="cite">I mostly agree with the changes you
described. Here's my 0.02€ :
<br>
- I strongly believe runtimes should be split into
different stacks
<br>
descriptors, but I don't like the idea of having to
maintain a fork of
<br>
the JDF one. We should only add extensions (old AS'es,
seam, ESB ...).
<br>
- I believe the merge all stacks descriptors in one
metamodel should be
<br>
done in stacks client. I know Raphael kinda started
working on that a
<br>
few months back, he probably can give us his insight.
<br>
- runtimes in stacks could list their managing JBT plugins
in the
<br>
labels property : i.e if ESB runtime can be downloaded,
but no ESB
<br>
plugin is installed, the we'd be able to discover both the
runtime AND
<br>
its associated plugin
<br>
<br>
Fred
<br>
<br>
Le mercredi 27 février 2013 18:22:35, Rob Stryker a écrit
:
<br>
<blockquote type="cite">Regarding Stacks, Runtimes, and
Remote Descriptors
<br>
<br>
Hi All:
<br>
<br>
This email is to try to begin discussion on some recent
duplication of
<br>
code and responsibilities, which should probably be
fixed before
<br>
things get too comfortable. I'm speaking specifically
about the role
<br>
of discovering runtimes to download, where that's done,
how that's
<br>
done, and which responsibility belongs to who. Forgive
me if the email
<br>
is long, as I am trying to be thorough.
<br>
<br>
Currently, there are three places from which runtimes to
download may
<br>
be discovered.
<br>
<br>
1) base/runtimes has an extension point named
downloadRuntimes, which
<br>
is used by AS Tools and Seam Tools (and perhaps others).
<br>
<br>
2) a remote descriptor file which acts as a second arm
of 1) and is
<br>
basically an xml form of 1) used to dynamically add new
runtimes as
<br>
they become available
<br>
<br>
3) The new Stacks methodology, currently stored in
<br>
jbosstools-central/maven/plugins/org.jboss.tools.maven.project.examples
<br>
<br>
<br>
We should begin unifying these three locations into one,
but the goal
<br>
is to do it correctly. So, I would first like to list
the benefits of
<br>
each.
<br>
<br>
a) downloadable runtimes provided through the extension
point cannot
<br>
be removed without a maintenance or major release of
some type, and
<br>
for this reason are semi-permanent
<br>
<br>
b) downloadable runtimes available via the remote
descriptor file may
<br>
be added OR removed at will. This provides flexibility
and
<br>
post-release updates are easy.
<br>
<br>
c) The new stacks section has a more robust model
capable of providing
<br>
more information than the downloadable runtimes does.
However, the
<br>
plugin requires several libraries and is currently
placed in the
<br>
jboss-central module, where others may not make use of
it.
<br>
<br>
d) the Stacks yaml file does not provide a place to
access the file
<br>
size for the download, however it does provide a
'labels' section,
<br>
which seems extendable to add whatever properties you
may want to add.
<br>
<br>
At first glance, it seems that Stacks is the superior
framework. It is
<br>
extensible, it can have unlimited labels (aka
properties) if desired,
<br>
and it already provides more information which is usable
to others who
<br>
may want it. To make use of stacks inside Runtimes,
however, we'd
<br>
either need to:
<br>
a) Expand the API in runtimes to allow other
plugins (like in
<br>
central) to provide downloadable runtimes, or,
<br>
b) Push 'stacks' out of central and down into
runtimes, as its
<br>
own
<br>
plugin upon which runtimes.core and runtime.ui can
depend.
<br>
<br>
The main negative of pushing stacks into base/runtimes,
in my opinion,
<br>
is that there are a significant number of libraries
required. It's not
<br>
too much, by far, but it is about 7 jars totalling about
1 megabyte.
<br>
Whether these jars belong in base/runtimes is debatable,
and currently
<br>
we do not have a "3rd-party dependencies" section in
base where we
<br>
organize common dependencies and versions together so
that each plugin
<br>
doesn't need to bundle their own 3rd party libraries. I
admit, this is
<br>
a debate for another time, but, I just wanted to point
out that
<br>
pushing the stacks logic down into runtimes would be
another example
<br>
of this issue.
<br>
<br>
Even still, I would argue that we should push stacks
into its own
<br>
small plugin below runtimes, deprecate the
"downloadRuntimes"
<br>
extension point, and the online downloadRuntime.xml
(wherever the file
<br>
is, I forget).
<br>
<br>
However, once we do that, there are many more questions.
The first is,
<br>
who's job is it to provide the yaml file from which
stacks are
<br>
generated?
<br>
<br>
Currently there is only one yaml file, and it is
referenced directly
<br>
via a github url. Aside from how (IMO) this is fairly
crazy in itself,
<br>
it causes another problem. The stacks client jar *can*
cache the yaml
<br>
file and only update if the timestamp has changed,
however, when
<br>
checking the timestamp on a github file, there isn't
one...
<br>
<br>
This would seem to imply we should take control of the
yaml file
<br>
ourselves and put it NOT in github but rather in a
release-specific
<br>
online-accessible folder, ex: jbt4.1/stacks.yaml,
jbt5.0/stacks.yaml,
<br>
etc.
<br>
<br>
The problem with this is that we are then taking control
away from the
<br>
jdf team, and once we take the file away, it is our job
to keep it
<br>
updated and in synch. This may cause errors if we are
not very
<br>
careful.
<br>
<br>
Assuming we do this, though, the next question is, do we
add seam and
<br>
esb runtimes to this yaml file, which currently only
provides
<br>
application servers? Remember, the purpose of moving
stacks down would
<br>
be to deprecate the downloadRuntime extension point,
therefore any
<br>
replacement would need to do everything downloadRuntime
does, which
<br>
includes providing seam and esb runtimes for download.
<br>
<br>
Let's assume (for now) that we simply add lines to the
yaml to allow
<br>
it to provide seam and esb runtimes. We may come back to
this point
<br>
later, but for now, assume we do that.
<br>
<br>
<br>
Then which plugin will provide the url to our copied
yaml? Who's
<br>
responsibility is it to point to this yaml file? Let's
look at our
<br>
options:
<br>
<br>
1) The runtimes plugin references the yaml
<br>
2) The central plugin references the yaml
<br>
<br>
Both of these fail after thinking about it. How?
<br>
<br>
1) If the runtimes plugin references the yaml, then the
download
<br>
runtimes dialog will list things (like seam) which may
not be present
<br>
in the installation. Imagine an installation with only
base and server
<br>
plugins installed, and so no seam or esb. A user
clicking 'download
<br>
runtimes' will see esb and seam downloads, but the
plugins which are
<br>
prepared to handle those runtimes after the download are
not present.
<br>
<br>
2) If central is in charge of providing this yaml,
perhaps through a
<br>
new extension point to the base/runtimes/stacks plugin
we add there,
<br>
then an installation including only plugins from base /
server will
<br>
have a BLANK download list. Users who install only
ASTools will not be
<br>
able to download JBoss Application Servers.
<br>
<br>
So both of these fail in their own way. The only
solution as I can
<br>
see, the only way it would work, would be to have
multiple such yaml
<br>
files, one for astools, one for seam, one for esb, etc.
Each of these
<br>
modules would provide their own yaml url to
base/runtimes/stacks via
<br>
an extension point in base/runtimes/stacks, and let
stacks fetch each
<br>
one and build a unified model.
<br>
<br>
Problems:
<br>
a) multiple urls need to be loaded
<br>
b) multiple yaml files need to be kept up to date,
instead of just
<br>
one. Multiply number of contributing plugins by number
of major
<br>
releases
<br>
c) Possibility of duplicates. Once you have
multiple yaml files
<br>
generating models, it's possible some duplication leaks
in. I'm not so
<br>
sure about this one, but Fred listed it as a concern.
<br>
<br>
So, by my analysis, this is the only way I can imagine a
unification
<br>
of these three models. I'll summarize the changes below,
but it does
<br>
seem there would be a bit of work to do.
<br>
<br>
Summary of changes:
<br>
1) Deprecate downloadRuntimes extension point
<br>
2) Create new plugin in runtimes module called
"stacks"
<br>
3) Add extension point to 'stacks' plugin called
stacksProvider
<br>
4) modify runtime.core and runtime.ui to use the
model built in
<br>
'stacks'
<br>
5) Create a web-accessible location for
jbt-release-relevent
<br>
data on
<br>
a per-module basis. For example,
<br>
<a class="moz-txt-link-freetext" href="http://wherever/jbt/4.1.0/stacks/astools.yaml">http://wherever/jbt/4.1.0/stacks/astools.yaml</a>, etc.
<br>
6) Copy the current jdf yaml file to that location
for astools.yaml
<br>
7) Create a new yaml file which can build stacks for
esb, seam, etc
<br>
8) Ensure astools, esb, seam, etc, make use of the
new
<br>
stacksProvider
<br>
extension point
<br>
9) Test the shit out of it
<br>
<br>
There are other benefits to this approach. Currently
there's no really
<br>
good mapping of downloadRuntimes id's to an app-server
id. This is
<br>
done in a hard-coded fashion in astools. This could
instead be added
<br>
to the labels in the astools.yaml file if desired. It
would allow
<br>
dynamic addition or removal of any runtimes, though in
the yaml
<br>
syntax. It would minimize connections and re-downloads
of the yaml
<br>
files, since they'll actually have a timestamp now (as
opposed to in
<br>
github, where they don't). And it could help clean up
some other areas
<br>
that could benefit from a cleanup.
<br>
<br>
I'd really like feedback on this issue from anyone who
knows anything
<br>
about the topic, because I know for sure I'm lacking a
bit in fully
<br>
understanding the entire api. But I'd love at the least
for someone to
<br>
tell me which of the logic here is obviously bad or if
i'm wrong on
<br>
any details.
<br>
<br>
Thanks and look forward to the feedback
<br>
<br>
- Rob Stryker
<br>
I break things, and then put them back together.
<br>
</blockquote>
_______________________________________________
<br>
jbosstools-dev mailing list
<br>
<a class="moz-txt-link-abbreviated" href="mailto:jbosstools-dev@lists.jboss.org">jbosstools-dev@lists.jboss.org</a>
<br>
<a class="moz-txt-link-freetext" href="https://lists.jboss.org/mailman/listinfo/jbosstools-dev">https://lists.jboss.org/mailman/listinfo/jbosstools-dev</a>
<br>
</blockquote>
_______________________________________________
<br>
jbosstools-dev mailing list
<br>
<a class="moz-txt-link-abbreviated" href="mailto:jbosstools-dev@lists.jboss.org">jbosstools-dev@lists.jboss.org</a>
<br>
<a class="moz-txt-link-freetext" href="https://lists.jboss.org/mailman/listinfo/jbosstools-dev">https://lists.jboss.org/mailman/listinfo/jbosstools-dev</a>
<br>
</blockquote>
</blockquote>
_______________________________________________
<br>
jbosstools-dev mailing list
<br>
<a class="moz-txt-link-abbreviated" href="mailto:jbosstools-dev@lists.jboss.org">jbosstools-dev@lists.jboss.org</a>
<br>
<a class="moz-txt-link-freetext" href="https://lists.jboss.org/mailman/listinfo/jbosstools-dev">https://lists.jboss.org/mailman/listinfo/jbosstools-dev</a>
<br>
</blockquote>
<br>
</blockquote>
<br>
</body>
</html>