Skip to main content

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [hyades-dev] Commitment candidates for Data Collection team.

I forgot one of the catagories:

Common Base Event
The common base event API's have been expanded for the following function.

1. Addition of the ability to deal with 'long' time representations.  The
previous time code only dealt with times that were of teh same format as
the event schema.  This was unwieldy to assume users to deal with.  We
added that ability to deal with UNC time formats.

2. Added APi's to the IExtendedDataElement interface to support native
types directly.

Richard K. Duggan
Problem Determination Enablement
IBM Toronto Laboratory
External: 905-413-2396
Internal: 969-2396



|---------+---------------------------->
|         |           Richard          |
|         |           Duggan/Toronto/IB|
|         |           M@IBMCA          |
|         |           Sent by:         |
|         |           hyades-dev-admin@|
|         |           eclipse.org      |
|         |                            |
|         |                            |
|         |           05/12/2003 10:43 |
|         |           AM               |
|         |           Please respond to|
|         |           hyades-dev       |
|         |                            |
|---------+---------------------------->
  >--------------------------------------------------------------------------------------------------------------------------------------------|
  |                                                                                                                                            |
  |       To:       Hyades-dev@xxxxxxxxxxx                                                                                                     |
  |       cc:                                                                                                                                  |
  |       Subject:  [hyades-dev] Commitment candidates for Data Collection team.                                                               |
  |                                                                                                                                            |
  |                                                                                                                                            |
  >--------------------------------------------------------------------------------------------------------------------------------------------|



The data collection team has the following candidates for commitment this
week.


Profiler
The following changes are being added to the profiler.
1.  Addition of the following information on class definitions:
      a.  Interfaces it implements
      b.  Superclass of the class
      c.  Classloader of the class.
      d.  Access modifiers of the methods it defines  (private, protected,
public)
      e.  Other modifiers of the methods that it defines (synchronized,
native, abstact, static)
      f.  Exceptions that each of the methods that it defines throw.

2.  Resolved an issue with primatives.  We have added the behaviour of
expressing each of the primatives as a class.  This is useful to keep a
consistant view when expressing array objects.  If we have an array object
that holds a collection of primatives we need to be able to define its
contained type.

3.  Resolved an issue with array objects.  The Sun JVMPI docs specify that
all array objects have a class id of '0'.  This made it impossible to
determine what class the array was intended to hold.  Therefore, the old
code used to blindly stick all arrays of Objects into the generic container
[Object.    As it turns out, some JVM's actually do give us the proper
class id of the type the array holds.  We have modified our code to
exploite this information when it is available.

4.  Added the ability to distinguish between objects which we saw allocated
verses those which we are defining.  When you start profiling an
application after time '0' some objects will already exist in the
application heap.  When these objects are refered to we must generate a
objectAlloc event for the object.  We now have the capability to inform the
user of whether the object was really allocated at this point or it is a
definition of an existing object.


RAServer
1.  Fixed a bug with the missing \config directory.




Richard K. Duggan
Problem Determination Enablement
IBM Toronto Laboratory
External: 905-413-2396
Internal: 969-2396

_______________________________________________
hyades-dev mailing list
hyades-dev@xxxxxxxxxxx
http://dev.eclipse.org/mailman/listinfo/hyades-dev





Back to the top