[drmaa-wg] ACTION REQUIRED: Finishing C Binding

Daniel Templeton Dan.Templeton at Sun.COM
Tue Feb 15 08:28:10 CST 2005


There are some open issues that we should discuss so I can finalize the
C binding spec.  *PLEASE READ THROUGH THE BELOW LIST OF ISSUES!*  Unless
I hear contrary opinions in the next day or two, I will handle the below
issues as indicated.  That means I will have the next version ready by
COB CET on Wednesday.

Tracker #780:
There are two interesting proposals in this tracker, but both are
significant changes.  We simply do not have enough time before Friday to
come to a reasonable conclusion.  Since it is a P3 tracker, I suggest we
save it for later.

Tracker #1094:
The proposal is that either the signed keyword should be removed from
the timeout parameter of the drmaa_wait() and drmaa_synchronize()
functions (since signed is the default behavior), or its presence needs
to be explained.
I'm all for adding a short bit of text that explains why the attribute
is signed.  If no one objects, I will add that to the next version.

Tracker #1127/#1141:
The C-binding document should explicitly define the linkage name of the
DRMAA library and explicitly demand the implementation as shared library.
I consider this a no-brainer.  Unless there are objections, I will add
it to the next version.

Tracker #1143:
The DRMAA_DURATION_HLIMIT and DRMAA_DURATION_SLIMIT constants in drmaa.h
have misspellings in their definitions ("durartion").
No brainer.  I'm adding it to the next version.

Tracker #1144:
There are several errors in the DRMAA C binding example.
Do we want to fix the errors or write a new example?  In the online
howtos at http://gridengine.sunsource.net, there are several *working*
examples which could be used instead.  Thoughts?

Tracker #1173:
The drmaa_get_next_*_*() functions needs their own error code to
indicate that there are no more elements left to traverse.
Since we broke trump on new error codes with the C binding, I think it's
completely reasonable to say that a given binding may need an error code
not forseen by the language independent spec.  Therefore, unless there
are objections, I will add this to the next version.

Tracker #1269:
Developers need a way to get the element count of the DRMAA String
vector types.
The vote on this one was 3 for, 1 against, with 1 abstaining.  Did we
have a quorum?  Is this ratified?  I am going to assume so and add it to
the next version.

Daniel

-- 
***************************************************
*        Daniel Templeton   ERGB01 x60220         *
*       Staff Engineer, Sun N1 Grid Engine        *
***************************************************
* "Roads? Where we're going we don't need roads." *
*                    -Dr. Emmett Brown            *
*                     Back to the Future (1985)   *
***************************************************






More information about the drmaa-wg mailing list