[Hypercontent-users] Re: Can hypercontent handle servlets and JSP?
Brought to you by:
alexvigdor
Menu
▾
▴
[Hypercontent-users] Re: Can hypercontent handle servlets and JSP?
|
From: Alex V. <av...@co...> - 2005-10-31 18:07:39
|
Hi Mike,
You could host a website in HC2, but for your average site it =
makes=20
more sense to publish it out to be hosted by a dedicated web server,=20
which is usually Apache. uPortal is not a web server, but a portal=20
server which can aggregate information and applications from multiple=20
back-end sources, including Apache or HyperContent. As such, uPortal=20
is not well suited to hosting complete web sites, but rather "bite=20
size" bits of information and functionality that can be arranged in a=20
configureable way based upon a user's rights and responsibilities. You=20=
might manage some of these discrete portlets in HC2 and serve them=20
directly to uPortal, for example if you want to require authentication=20=
to view that particular information, since you can configure single=20
sign-on between uPortal and HC2. You could also publish static portlet=20=
content from HC2 to an Apache server, which uPortal could then pull=20
form. HC2 has a flexible workflow for delivering content; the "build"=20=
and "publish" model is presented by default for maximum backwards=20
compatibility with HC1.4, but at Columbia we're actually migrating most=20=
sites to a one-step publish, where the site as it appears in HC2 is=20
rendered directly into the public web space. You could also set up=20
multiple "build" or staging steps, which would primarily useful for=20
very sensitive information that might need to go through several steps=20=
of vetting.
The crashing sounds a bit strange; one possible solution would =
be to=20
boost the JVM memory in the ant "build.xml" script - you'll see it's=20
currently set to 'maxmemory=3D"256m"', which really should be adequate=20=
for the sample. Make suer to check the "hypercontent.log" for error=20
messages. Another probably more useful idea would be to download the=20
latest code from CVS; while those don't remind me of any particular=20
bugs that have been fixed since the 2.0b4 release, there have been a=20
ton of overall fixes resulting from our production migration these last=20=
few weeks. The current code in CVS is exactly what we're running in=20
production; you can find info on how to download it from the web site. =20=
The CVS files are in the exact same structure as the distribution, but=20=
lack the sample project, which you can add in yourself.
There are two ways to add users in HC2; if you use an external=20
authentication provider such as CAS, the users are automatically added=20=
to the system once they have logged in the first time. To add a user=20
manually to HC2, you need to create a file under the "users" directory.=20=
You'll see an example of where these go if you click the "Admin User"=20=
link after logging in using the sample admin user. Basically, the file=20=
has to live beneath a couple of directories that represent the first,=20
then the first and second letters of the user name. E.g.
http://localhost:8080/users/m/my/myuser/user.xml
You would add this by starting at
http://localhost:8080/users/
If the first letter of the username is not already in the list, use the=20=
"add" menu to add "First letter". It will then appear in the list;=20
click it, then add "First two letters". Click that directory, then=20
add "User" and enter the username. This will bring you to the VCard=20
editor where you can search LDAP for the user's info, upload a .vcf=20
file from your PIM, or manually enter it by clicking "Start from=20
scratch". The main things to worry about in the VCard editor are the=20
Full Name, which is displayed when the user logs in, and the email=20
address for workflow notifications. You'll see a second editor called=20=
"Change Password" where you can enter a password for the user. Once=20
you have done this, save the user file, and you are ready to log in as=20=
that user.
Once you have users in the system, in order for them to do anything you=20=
will have to add them to groups in the groups file (/config/groups.xml=20=
in each project) and/or assign permissions to them or their groups in=20
the permissions files (/config/permissions.xml). I think there's more=20=
discussion of the groups and permissions in the mailing list archives.
Cheers,
Alex
On Oct 31, 2005, at 11:10 AM, michael ayodele wrote:
> I think I am a little confused with the roles of hypercontent and=20
> uportal. Let me see if I got this right. Hypercontent is a web=20
> application platform that allows groups of people to work together in=20=
> order to build a consistent web application. Hypercontent just creates=20=
> the website, and does not host the website. Once a website is complete=20=
> it must be built and then published to a host server(uportal) where=20=
> it is actually used by end users. And uportal is a serverside=20
> framework for handling accesses by many users=A0for university =
websites.=20
> Hopefully you can help me fill in gaps and misunderstandings.
> =A0
> Also things I found using hypercontent
> - whenever I try to=A0build the sample project=A0it builds half of it =
then=20
> crashes=A0
> - When I try to download the entire sample project zip(in workflow)=A0it=
=20
> crashes
> =A0
> Other Questions
> - How do I create new users and their passwords and permissions?
> =A0
> Thanks Alex
> =A0
> On 10/28/05, Alex Vigdor <av...@co...> wrote: HI Mike,
>> =A0=A0=A0=A0=A0=A0 There isn't currently a plugin for rendering JSPs, =
but it=20
>> should be
>> possible.=A0=A0Similarly, there is no built-in database access, but =
you
>> should be able to accommodate that via plugins.
>>
>> There are 2 APIs you should look at:
>>
>> org.hypercontent.project.engine.IPipelineStage
>> =A0=A0=A0=A0=A0=A0 - sample implementations under=20
>> org.hypercontent.project.engine.stage
>> =A0=A0=A0=A0=A0=A0 - can be added into any view by binding into a =
pipeline; see
>> "/config/pipelines.xml" in the bootstrap project for how all the
>> built-in screens are configured
>> =A0=A0=A0=A0=A0=A0 - this is where you would query the database to =
pre-populate a=20
>> form.
>> You could render the information into XML and follow with an XSL =
stage
>> for presentation,
>> =A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 or bind the values into =
java objects and place them in=20
>> a
>> VelocityContext for presentation with a Velocity template (see
>> VelocityStage.getContext())
>> =A0=A0=A0=A0=A0=A0 - you could develop a stage which would render =
with a JSP,=20
>> though this
>> would probably take more work
>> =A0=A0=A0=A0=A0=A0 - the stage "process" method gives you a Task =
which has a=20
>> buffer of
>> type "IWriteableAsset".=A0=A0You control what comes out of the stage =
by
>> setting the
>> =A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 IDataLoader on the buffer; =
this is simply an object=20
>> that returns an
>> InputStream with your data (e.g. your XML rendered information).=A0=A0I=
f
>> you're using velocity,
>> =A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 all you have to do is add =
your objects into the=20
>> context and follow up
>> with a VelocityStage that runs through a .vm template that references
>> those objects.
>> =A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 If you were to develop a =
JSP stage, you would create=20
>> an IDataLoader
>> implementation that presents the output of the JSP as an InputStream.
>>
>> org.hypercontent.form.IFormProcessor
>> =A0=A0=A0=A0=A0=A0 - this is where you collect data from user form =
inputs
>> =A0=A0=A0=A0=A0=A0 - you could write a form processor to store the =
form inputs in=20
>> a
>> database
>> =A0=A0=A0=A0=A0=A0 - see samples under "org.hypercontent.form"
>> =A0=A0=A0=A0=A0=A0 - these are bound into views in the pipelines file =
once again;=20
>> you
>> assign them a key like "my-form", and assign individual form inputs =
to
>> it like so:
>> =A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 XSL:
>> =A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 =
<input name=3D"{$my-form}name"/>
>> =A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 Velocity
>> =A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 =
<input name=3D"${my-form}name"/>
>> =A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 Your processor will see it =
with the name "name": the=20
>> prefix is used
>> by the framework to route individual inputs to individual processors.
>> =A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 In this way, you could =
have different inputs in the=20
>> same form routed
>> to separate logical processors, e.g.
>> =A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 =
<input name=3D"${my-form}name"/>
>> =A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0=A0 =
<input name=3D"${my-other-form}name"/>
>>
>> All the built in application screens are developed with these two=20
>> APIs,
>> so there are plenty of examples.=A0=A0There are some screen-specific
>> stages, like "ManagePrepStage", and other commonly reused stages like
>> "XSLTransformStage", "VelocityStage" and "TrimStage".=A0=A0You can =
chain
>> any number of stages in any way, so the possibilties are endless!
>>
>> -Alex
>>
>>
>> On Oct 28, 2005, at 12:14 PM, michael ayodele wrote:
>>
>> > How would one go about querying and storing information in=20
>> databases?
>> > Thanks again for all your help.
>> >
>> > Michael Ayodele
|
