Re: Question on Java JDK build environments impacting JRE use

David,


David Ostrovsky-3 wrote
> But anyway i can conduct a new and shiny "51" byte code version of LO if
> you would like to try it and see if we have any improvements with JAB
> with it?

I would still very much like to explore this further if you are able to
generate viable Java 1.7 bytecode of master without having to refactor the
entire project to meet stricter code requirements of Java SE 7.

In any case, I hope you've convinced yourself that there really is an issue
with the JRE 1.7u25 & JAB v2.0.3 bridge of  UNO Accessability API  <--> Java
Accessibility API serviced by JRE 7 on Windows OS and tht fdo#58995 remains
a valid accessibility issue for LibreOffice.  Let me know if you still have
doubts.

Also, apologies for that rather confusing set of postings  to the fdo#58995
bug. I really should just start editing outside BZ and then paste into the
dialog.

Setting up just a 32-bit JRE and the associated JavaFerret-32.exe (from the
Java SE 6 Java Access Bridge package) functions as a Graphical accessibility
explorer. It is comparable in function to tools provided in Python based
*Accerciser* for GNOME  AT-SPI, or the *AccProbe* Eclipse Rich-Client
Product for evaluating  IAccessible2 on Windows, or even Microsoft's
Inspect.exe for UIAutomation and MAA accessibility events and roles.

Along with JavaFerret-32, Oracle (via Sun) also provides the JavaMonkey-32
which compliments javaferret with a GUI tree representation of a documents
accessible objects. While JavaFerret will show the attributes of any single
accessible object--the tree representation of JavaMonkey shows hierarchical
relationships of parent and child so they complement each other but
JavaFerret is really the tool of choice when assessing how well AT tools can
be interfaced.

They've not repackaged them as JRE 7 utilities, but the packaging in JAB
v2.0.2 build still functions well enough.

In a perfect world attributes (annotations hard coded into the source) for
both the Ferret and the Monkey representations would be fully populated--as
Oracle has done with the SwingSet2.jar demo.  Reality is, we are doing well
if we have correct AccessibleRole assigned along with parent and child
relationships. 

At some point we will gain access to the Apache OpenOffice work being done
on their IA2 port from IBM Symphony and have the start of an UAA <--> 
IAccessible2 bridge for Windows.  That will reduce dependence on the Java
Accessibility API and JRE using the Java AccessibilityBridge, but as Oracle
support for J2SE 1.5 and Java SE 6 is retracting--we have to do something to
improve function of Java 7 JREs for Windows based users.

Stuart



--
View this message in context: 
http://nabble.documentfoundation.org/Question-on-Java-JDK-build-environments-impacting-JRE-use-tp4062592p4062800.html
Sent from the Dev mailing list archive at Nabble.com.