[Swift-devel] [Swift-commit] r5316 - in branches/release-0.93/src/org/griphyn/vdl: karajan/lib mapping

Fri Nov 25 23:00:58 CST 2011

The test example seems to work in trunk, and in 0.91. Fails in 0.92.1 and in current rev of 0.93.

The behavior of assigning mapped variables is mysterious to me; I need to do more testing.  In many cases the mapping is lost on assignment, but in this test program, the mapping is somehow retained (or visible?) when the array to which mapped variables has been assigned is passed as an argument to an app() function.

- Mike

----- Original Message -----
> From: "Michael Wilde" <wilde at mcs.anl.gov>
> To: hategan at ci.uchicago.edu
> Cc: "Swift Devel" <swift-devel at ci.uchicago.edu>
> Sent: Friday, November 25, 2011 8:43:00 PM
> Subject: Re: [Swift-devel] [Swift-commit] r5316 - in branches/release-0.93/src/org/griphyn/vdl: karajan/lib mapping
> What seems to be happening here is that when a mapped file variable is
> assigned to a variable, the mapping is not assigned with it.
> 
> I need to check, but I think that in older revisions (eg 0.92.1 ?) the
> mapping *was* assigned along with the value (i.e. the state) of the
> file variable.
> 
> Whats happening in this case is that when the array [out,out2] is
> created, the file-typed members of the constructed array have only
> default mappings (ie _concurrent/etc) - they have lost their original
> mappings assigned by the single_file_mapper <"filename">.
> 
> I *think* this is incorrect behavior, but perhaps that needs
> discussion. What do you think?
> 
> - Mike
> 
> ----- Original Message -----
> > From: "Michael Wilde" <wilde at mcs.anl.gov>
> > To: hategan at ci.uchicago.edu
> > Cc: "Swift Devel" <swift-devel at ci.uchicago.edu>
> > Sent: Friday, November 25, 2011 7:59:19 PM
> > Subject: Re: [Swift-devel] [Swift-commit] r5316 - in
> > branches/release-0.93/src/org/griphyn/vdl: karajan/lib mapping
> > Actually, looking at the error message more closely I am not sure my
> > comment below is correct. The message may be complaining about the
> > array object as a whole, and the [2] may be stating the number of
> > elements (the array does indeed have 2 elements) as opposed to a
> > member of the array that it believes is not mapped.
> >
> > I need to experiment further. Do you see an error in the test case?
> >
> > - Mike
> >
> >
> > ----- Original Message -----
> > > From: "Michael Wilde" <wilde at mcs.anl.gov>
> > > To: hategan at ci.uchicago.edu
> > > Cc: "Swift Devel" <swift-devel at ci.uchicago.edu>
> > > Sent: Friday, November 25, 2011 7:52:26 PM
> > > Subject: Re: [Swift-devel] [Swift-commit] r5316 - in
> > > branches/release-0.93/src/org/griphyn/vdl: karajan/lib mapping
> > > Mihael, this does not fix the bug that I reported in bug 644, but
> > > it
> > > does seem to clarify the cause: the test case in bug 644 now gives
> > > the
> > > message:
> > >
> > > ?:file[2] - Closed is not a mapped type
> > >
> > > @filenames() should only be looking at array elements [0] and [1]
> > > -
> > > seems to me that its going beyond the assigned elements of the
> > > array
> > > to complain about [2] not being mapped. If I set element [2] it
> > > complains about element [3] in the error message above.
> > >
> > > Did you try the test case? That case should not generate an error
> > > -
> > > I
> > > believe that its a valid program. if you replace @filenames() with
> > > two
> > > calls to @filename() it works. Ie all elements of the array passed
> > > to
> > > @filenames() are properly mapped.
> > >
> > > - Mike
> > >
> > >
> > >
> > >
> > > ----- Original Message -----
> > > > From: hategan at ci.uchicago.edu
> > > > To: swift-commit at ci.uchicago.edu
> > > > Sent: Thursday, November 24, 2011 3:18:52 PM
> > > > Subject: [Swift-commit] r5316 - in
> > > > branches/release-0.93/src/org/griphyn/vdl: karajan/lib mapping
> > > > Author: hategan
> > > > Date: 2011-11-24 15:18:51 -0600 (Thu, 24 Nov 2011)
> > > > New Revision: 5316
> > > >
> > > > Modified:
> > > > branches/release-0.93/src/org/griphyn/vdl/karajan/lib/VDLFunction.java
> > > > branches/release-0.93/src/org/griphyn/vdl/mapping/AbstractDataNode.java
> > > > branches/release-0.93/src/org/griphyn/vdl/mapping/RootArrayDataNode.java
> > > > branches/release-0.93/src/org/griphyn/vdl/mapping/RootDataNode.java
> > > > Log:
> > > > throw better exception than NPE when @filename is called with a
> > > > non-mapped argument
> > > >
> > > > Modified:
> > > > branches/release-0.93/src/org/griphyn/vdl/karajan/lib/VDLFunction.java
> > > > ===================================================================
> > > > ---
> > > > branches/release-0.93/src/org/griphyn/vdl/karajan/lib/VDLFunction.java
> > > > 2011-11-23 19:33:26 UTC (rev 5315)
> > > > +++
> > > > branches/release-0.93/src/org/griphyn/vdl/karajan/lib/VDLFunction.java
> > > > 2011-11-24 21:18:51 UTC (rev 5316)
> > > > @@ -197,8 +197,17 @@
> > > >
> > > > private static String[] leavesFileNames(DSHandle var) throws
> > > > ExecutionException, HandleOpenException {
> > > > Mapper mapper;
> > > > +
> > > > synchronized (var.getRoot()) {
> > > > - mapper = var.getMapper();
> > > > + if (var instanceof AbstractDataNode) {
> > > > + mapper = ((AbstractDataNode) var).getActualMapper();
> > > > + if (mapper == null) {
> > > > + throw new ExecutionException(var + " is not a mapped type");
> > > > + }
> > > > + }
> > > > + else {
> > > > + mapper = var.getMapper();
> > > > + }
> > > > }
> > > > List<String> l = new ArrayList<String>();
> > > > try {
> > > >
> > > > Modified:
> > > > branches/release-0.93/src/org/griphyn/vdl/mapping/AbstractDataNode.java
> > > > ===================================================================
> > > > ---
> > > > branches/release-0.93/src/org/griphyn/vdl/mapping/AbstractDataNode.java
> > > > 2011-11-23 19:33:26 UTC (rev 5315)
> > > > +++
> > > > branches/release-0.93/src/org/griphyn/vdl/mapping/AbstractDataNode.java
> > > > 2011-11-24 21:18:51 UTC (rev 5316)
> > > > @@ -513,7 +513,7 @@
> > > > }
> > > > }
> > > >
> > > > - protected Mapper getActualMapper() {
> > > > + public Mapper getActualMapper() {
> > > > return null;
> > > > }
> > > >
> > > >
> > > > Modified:
> > > > branches/release-0.93/src/org/griphyn/vdl/mapping/RootArrayDataNode.java
> > > > ===================================================================
> > > > ---
> > > > branches/release-0.93/src/org/griphyn/vdl/mapping/RootArrayDataNode.java
> > > > 2011-11-23 19:33:26 UTC (rev 5315)
> > > > +++
> > > > branches/release-0.93/src/org/griphyn/vdl/mapping/RootArrayDataNode.java
> > > > 2011-11-24 21:18:51 UTC (rev 5316)
> > > > @@ -103,7 +103,7 @@
> > > > throw new
> > > > FutureNotYetAvailable(waitingMapperParam.getFutureWrapper());
> > > > }
> > > >
> > > > - protected Mapper getActualMapper() {
> > > > + public Mapper getActualMapper() {
> > > > return mapper;
> > > > }
> > > >
> > > >
> > > > Modified:
> > > > branches/release-0.93/src/org/griphyn/vdl/mapping/RootDataNode.java
> > > > ===================================================================
> > > > ---
> > > > branches/release-0.93/src/org/griphyn/vdl/mapping/RootDataNode.java
> > > > 2011-11-23 19:33:26 UTC (rev 5315)
> > > > +++
> > > > branches/release-0.93/src/org/griphyn/vdl/mapping/RootDataNode.java
> > > > 2011-11-24 21:18:51 UTC (rev 5316)
> > > > @@ -213,7 +213,7 @@
> > > > throw new
> > > > FutureNotYetAvailable(waitingMapperParam.getFutureWrapper());
> > > > }
> > > >
> > > > - protected Mapper getActualMapper() {
> > > > + public Mapper getActualMapper() {
> > > > return mapper;
> > > > }
> > > >
> > > >
> > > > _______________________________________________
> > > > Swift-commit mailing list
> > > > Swift-commit at ci.uchicago.edu
> > > > https://lists.ci.uchicago.edu/cgi-bin/mailman/listinfo/swift-commit
> > >
> > > --
> > > Michael Wilde
> > > Computation Institute, University of Chicago
> > > Mathematics and Computer Science Division
> > > Argonne National Laboratory
> > >
> > > _______________________________________________
> > > Swift-devel mailing list
> > > Swift-devel at ci.uchicago.edu
> > > https://lists.ci.uchicago.edu/cgi-bin/mailman/listinfo/swift-devel
> >
> > --
> > Michael Wilde
> > Computation Institute, University of Chicago
> > Mathematics and Computer Science Division
> > Argonne National Laboratory
> >
> > _______________________________________________
> > Swift-devel mailing list
> > Swift-devel at ci.uchicago.edu
> > https://lists.ci.uchicago.edu/cgi-bin/mailman/listinfo/swift-devel
> 
> --
> Michael Wilde
> Computation Institute, University of Chicago
> Mathematics and Computer Science Division
> Argonne National Laboratory
> 
> _______________________________________________
> Swift-devel mailing list
> Swift-devel at ci.uchicago.edu
> https://lists.ci.uchicago.edu/cgi-bin/mailman/listinfo/swift-devel

-- 
Michael Wilde
Computation Institute, University of Chicago
Mathematics and Computer Science Division
Argonne National Laboratory