Hi,first of all, you should not run a profiler on a production
environment if you don't have problems that only occure in production.
However, you should use a JMX Monitor to log memory usage and such and
if you find yourself in trouble that you can't solve in another
environment, then push the profiler to the environment. Don't monitor
unless you need to.Over the years, I have taken a like in YourKit. It
seems to be the one that has the least impact on the system. Some years
ago, we got the err...
It depends...If you just need to pass a selection, like "show med the
cities in the country I selected" - use request parameter or Sling
Suffix. This is always the default choice imho.If it is some sort of
wizard, like a registration, then use the session or a cookie. That is
if you need to be able to resume the flow or go back and forth.
Otherwise, use request parameters./Ove
I think you are looking at the problem from a slightly wrong angle.Your
angle is: I need to have ONE service that can do A or B depending on the
time.My angle would be: I need a service that can create a task that
runs depending on the date-time-pattern.This means that I would create a
service factory, and have the instances run depending on a time pattern.
Take a look at the Workflow purger. The purger service does not do
anything unless you create an instance of a purger task, that is run
depe...
This is a common misunderstanding that you are updating the same node on
the publisher.What is actually happening is that you are creating a new
version of the node when you activate it and on the publisher, you then
insert the new node instead of the old one. If you look at the
jcr:content of your node in the Publisher, then you will notice the
properties jcr:baseVersion, jcr:predecessors and jcr:versionHistory that
are all of the type Reference (or Reference array). Take a look at them
nodes a...
In the bundle that holds the class com.book.rental.modal.Book, make sure
that it is exported in the OSGi-manifest. Otherwise, there is no bundle
that will present this class to the system./Ove
This is the CORS-problem (Cross-Origin Resource Sharing)Basically, there
are two ways of doing this.1. Have the external domain use a JSONP and
implement that interface. (See
http://stackoverflow.com/questions/13893361/access-control-allow-origin-localhost).2.
You can manipulate on the Access-Control-Allow-Origin header in your
response. (See
https://developer.mozilla.org/en-US/docs/HTTP/Access_control_CORS?redirectlocale=en-US&redirectslug=HTTP_access_control)Notice
that this is not fully suppo...
dcollie wrote... You need to either "scope" your own jQuery UI CSS, or
scope the CQ version. If you choose to scope the CQ version, you need to
change "\etc\clientlibs\granite\jquery-ui\css\jquery-ui.css" file. Add
the scope of ".cq5jqui" to all the classes. If you continue to have
problems, please log a Daycare ticket. Wouldn't it then also be
necessary to add that css scope to all the places where AEM is using
jquery-ui dialogs? IMHO, that is a bit to intrusive on the author system
and I think...
That is exactly what the Sling Post Servlet can do.Attached is a small
package I've done for test and education purposes (change the ext to
zip). It adds a node /content/geometrixx-media/en/propertytest and 8
subpages that all has the tag Music but only half of the pages has the
"wrongly set" Movies (named hastag#). The acceptance criteria is to use
a POST to remove the Movies tag. I used curl to post my content, but you
can use whatever POST-tool you want.The curl command to remove a tag
is:cur...
What you need to do is sending the following information in the request
to the metadata to remove is, here removing the tag year/2011:./cq:tags
-year/2011 ./cq:tags@Patch true ./cq:tags@TypeHint String[]The magic is
in the @TypeHint String[] that tells the SlingPost servlet that it
should handle the value as a String array. Without it, it tries to
default to just String, since that is what you provide for it./Ove
Easiest in this case would actually be to create a rudimentary version
handling system using the content system. What you can do is that you
create a workflow that whenever a new version of a page is created, it
also creates a subpage with the same content and numbering it according
to some schedule like incrementaly increasing integer or timestamp. What
you then should do is to have the node internally go for the "current"
version when you access the page.As I see it, the requirement is not
cor...
Hi,when you say "to the same location", have you unpacked the zip,
edited, and then re-packed it? Some text editors automatically add .txt
when you save so make sure that you have the correct suffix on your
files.One thing I have notices is that it perfectly ok to unpack the
files and edit it, but if you are on a Windows based system, you need to
drop the __MACOSX folder before you repack the zip and I had to touch
all the files so that they get "windowsified"./Ove
What that error tells you is quite clear. The file index.html or
index.htm needs to be in the root of the zip file. It can't be named
anything else and it can't be in a substructure.What does your zip look
like?/Ove
SQL is the one that most developers knows and is most likely to get
correct. But, some things are not doable.JCR-SQL2 has a lot of power an
the ability to do whatever you want in the JCR tree.JQOM is just a Java
representation of the JCR-SQL2 and eventually boils down to a SQL2
query. I actually like this one, since I don't have mess around with a
lot or String concatenation.Just my $0.02. /Ove
Is this for analysis tool like Coremetrix?The best approach, IMHO, would
be to create a workflow that is triggerd on modify (create, update and
delete) of a node, check if the change is on a node of the type Page and
then write it to the database according to specification. Of course, you
need to have a second workflow or service that sends all the pages to
the workflow initially as a workflow payload.The question I instantly
have to ask is "what is a page"? Is "A page node in AEM" or is it "A U...
What we do is that we have a tag that we quard with the WCM and check if
the runmode is a test environment.This tag displays information about
the bundles we have deployed on the site, what runmodes that are enabled
on the site, if it is an author or publish. All the build information is
generated and inserted in the bundles. Then you will see it on the pages
and not in the quickstart toggler. Even if the server is started as a
service.If I get any spare time, I will add this feature to the ACS ...
Jörg Hoh wrote... Why do you want to process the journal data? There is
no API to access this data from outside of Jackrabbit/CRX, it's an
implementation detail. Jörg I ask the same question as Jörg. Why do you
want to analyze the journal? Data masturbation?Ove
Hi,no, the JCR query language does not have any randomizer, but the
result set is a collection and you can always use the Java
Collections.shuffle(List).A better solution, if you are going to
collect from a named path, is to use the
http://dev.day.com/docs/en/cq/current/javadoc/org/apache/sling/api/resource/ResourceWrapper.html#getChildren%28%29./Ove
You should, as much as possible, use relative linking and relative
resolving. It should not matter if it is on admin.mysite.com,
acc-admin.mysite.com or www.mysite.com or acc-www.mysite.com. Everyone
should resolve to the same node. So IMHO, you are doing it the right
way. /Ove
Hi Scott,no, I am just trying to make the life easier for my authors. We
have ended up in creating two components instead of one and inlineing
the dialogue that holds the tags so that when you double click on the
"tag defining" component, it replaces the text that represents it. So
instead of my "card" having the properties of both tags and values, we
split it so that the card now consists of card-tag and card-value
components./O
Hi,I am stuck on a problem where I try to create an inline editing of a
tag page property (xtype = tag). Is it even possible to do that so that
the editor don't have to go into the page properties to be able to set
the tag. To display it, that is easy, but to actually inline edit it??
/Ove
The allowedParents property is an array of paths to the components where
defining component can be used. In my understanding, it is not the
component name in the page that is defined but the component type. If
you do not have a component type that is called
/apps/something/something/article then I don't think that is would
work.What we have done, when we want paragraps that has restricted set
of allowed children is one of two things.1. Define the allowedChildren
in the parent type definition suc...
LinearGradient wrote... Thanks for the comment Ove. Just one question:
is there anywhere officially indicated that Predicates are converted to
SQL 2 queries? Because a colleague of mine thought the reverse is true:
that SQL 2 queries are converted to predicates in the backend. Well...
converting structured code to SQL2 is easy. Converting text to code... a
bit harder.Empirical studies. When I use Predicates, if fires a query,
sometimes XPATH, sometimes SQL./O
Short answer: No.Longer answer: No, since the Search Predicates
generates SQL2 queries in the backend there are, in theory, no
difference. However, my experience is that developers tend to be better
in using Predicates in Java code and get better preformance out of those
queries than when they use SQL2 directly. My quess is that this is
because the fellows that has been writing the Predicates actually know
what they are doing and optimize the queries and most of us does not
understand SQL2 well ...
LinearGradient wrote... Thanks for the reply Ove. But what's the
difference between "unlimited with intention" "the default unlimited".
And are they documented anywhere? Thanks. Hi,it is more of a way of
thinking when you create query languages. There might be limitations set
by a connector or something that then honors the default limit notation
(0) and if you want to intentionally get everything, regardless of any
limitations, then you use the unlimited flag (-1).I have myself done
quite a lot...
The -1 is used for "unlimited with intention" and the 0 is "the default
unlimited". If you use p.limit=, that is the limit
of the anwswers you get.That is why only -1 is documented. /Ove
Hi,no, it turned out to be a "I did more than one change at the time
without knowing about it"-error. The faults we got when we changed the
maven-scr-plugin was actually due to a Java 7 error. Our mocking
framwork needed some extra parameters. Not it seems to be working./Ove
Hi,is there anyone that has used the maven-scr-plugin and the
org.apache.felix.scr.annotations higher than 1.7.4 with
AEM5.6.1?Whenever I try to step the version of the plugin or the
annotation, I get the error "Bad type on operand stack in arraylength".
/Ove
Hi,what is "mid-sized"? Number of components? Number of nodes? Number of
authors? How will the rest of the system look like? Will you have the
need for many author instances? Will there be need for multi-server-hall
distributions? Will there be a need for 10 second catastrofy recovery?To
many things that are unknown./Ove
