[g-a-devel] Need ATK experienced help

["Quiring, Sam" <Sam Quiring windriver com>]

Greetings,

 

I'm adding ATK support to a Python platform that implements its own graphics library from scratch.  I'm using Papi, which has been really excellent.

 

I've just finished implementing some of the ATK support.  ATK objects (actually, papi.atk.AtkObject's) are created on demand.  When an ATK request comes in for the Nth child of an ATK node (via the ATK callback AtkObjectClass.ref_child), I examine the associated GUI node's children, find the Nth one, and create a corresponding ATK node for it, which I pass back to the caller.

 

One client I'm using to test this is an AT-SPI dump program I wrote a while ago.  It dumps the entire AT-SPI tree to a file.

 

Here's my problem:

 

I bring up the Python platform with the initial screen GUI.  I run my AT-SPI dump client program.  The dump is beautiful.  I run it again.  The Python ATK support crashes with errors in the Python log like this:

 

** (process:2857): CRITICAL **: get_atkobject_from_servant: assertion `ATK_IS_OBJECT(object->gobj)' failed

** (process:2857): CRITICAL **: impl_accessibility_accessible_get_name: assertion `object != NULL' failed

** (process:2857): CRITICAL **: get_atkobject_from_servant: assertion `ATK_IS_OBJECT(object->gobj)' failed

** (process:2857): CRITICAL **: impl_accessibility_accessible_get_description: assertion `object != NULL' failed

** (process:2857): CRITICAL **: get_atkobject_from_servant: assertion `ATK_IS_OBJECT(object->gobj)' failed

The first time I run the client, my ATK Python support is building the ATK nodes on demand.  Those nodes are saved and the second time I run the client all the ATK nodes should exist.  And they do.  But somewhere in between, the first 12 bytes of memory in some of the ATK nodes is getting trashed.  In this case I'm talking about the real ATK node, the one pointed to by the papi.atk.AtkObject.

 

Here are some odd facts:

 

I wrote a routine in Python to run though the entire papi.atk.AtkObject tree and dump each node and it's parent (parent obtained via ATK call). I dumped the tree before and after the memory got trashed. In the before dump, every node's parent displays.  In the after dump, the leaf nodes cannot find their parent as a result of memory being trashed.  But interior nodes are ok.

 

To make ATK do stuff, my Python program must call a Papi routine (papi.atk.iterate()) over and over in the main GUI loop.  I put some checking code around this call examining one particular node which I knew would get trashed.  Before one iteration step the node is fine, afterwards, the node is trashed.  Is there any way I can figure out what action was performed by that iteration step?

 

Does this describe a situation familiar to anyone?  It has some symptoms of garbage collection, as if the leaf ATK nodes don't have the proper ref count.  But the "fact" that the memory seems to be trashed by an iteration of ATK main loop doesn't seem consistent with garbage collection.

 

Any suggestions would be appreciated.

 

-Sam