[r15404]
by
bugman
Import fix for the recently deleted IO classes.
|
2012-02-29 10:13:17
|
Tree
|
|
|
2012-02-29 10:06:20
|
Tree
|
[r15402]
by
bugman
Fix for the Immediate_result_queue class.
The __init__() method was broken!
|
2012-02-29 10:03:48
|
Tree
|
[r15401]
by
bugman
Clean up of the processor IO module.
The now unused IO_filter and PrependStringIO classes have been eliminated and the imports all
removed.
|
2012-02-29 09:54:11
|
Tree
|
[r15400]
by
bugman
Eliminated all usage of sys.__stdout__ and sys.__stderr__ in the multi-processor package.
This returns full control of IO streams to the parent program.
|
2012-02-29 09:49:46
|
Tree
|
[r15399]
by
bugman
Merged revisions 15380,15385,15388-15395 via svnmerge from
svn+ssh://bugman@.../svn/relax/1.3
........
r15380 | bugman | 2012-02-27 17:46:41 +0100 (Mon, 27 Feb 2012) | 8 lines
Fix for the storage of numpy data structures in the relax save files.
If special structures are created by the user, these were not correctly stored. The precision was
often too low (printed at the numpy default of a precision of 8), and if one dimension had more than
1000 elements, ellipses would be stored instead. The ellipses would cause the save file to be
unreadable by relax.
........
r15385 | bugman | 2012-02-28 13:23:04 +0100 (Tue, 28 Feb 2012) | 3 lines
Fix for the Mf.test_write_results system test for the higher precision numpy numbers in results files.
........
r15388 | bugman | 2012-02-28 14:43:44 +0100 (Tue, 28 Feb 2012) | 6 lines
Bug fix for the multi-processor package.
The 2to3 conversion had created code which was not functional on Python2. This should not have
happened!
........
r15389 | bugman | 2012-02-28 15:50:30 +0100 (Tue, 28 Feb 2012) | 3 lines
Implemented the flush() method for the multi-processor IO streams.
........
r15390 | bugman | 2012-02-28 15:59:58 +0100 (Tue, 28 Feb 2012) | 3 lines
The flush() methods for the multi-processor IO streams now are dummy functions.
........
r15391 | bugman | 2012-02-28 16:17:23 +0100 (Tue, 28 Feb 2012) | 5 lines
Completion of a TODO in the multi-processor package.
The stdio_capture argument to the Processor classes has been eliminated.
........
r15392 | bugman | 2012-02-28 16:23:20 +0100 (Tue, 28 Feb 2012) | 3 lines
Eliminated the excessive setup_stdio_capture() method from the Processor classes.
........
r15393 | bugman | 2012-02-28 16:24:12 +0100 (Tue, 28 Feb 2012) | 3 lines
Removed a no longer relevant FIXME comment.
........
r15394 | bugman | 2012-02-28 16:30:03 +0100 (Tue, 28 Feb 2012) | 5 lines
The relax prompt mode now prints the intro message to sys.stdout.
This fixes this mode with the logging option or in the mpi4py multiprocessor mode.
........
r15395 | bugman | 2012-02-28 18:51:09 +0100 (Tue, 28 Feb 2012) | 14 lines
Big fix for the IO streams when running in the multi-processor mpi4py mode!
This affects all multi-processor implementations, excluding the standard uni-processor. The entire
stream capture design has been overhauled. The IO streams of the master processor are no longer
touched, whereas those of the slaves are captured, stored in a list of lists where the original
'S |' and 'S E|' tokens are prepended to each line as well as a number identifying the stream,
stored in the results command object, and unpacked and sent to sys.stdout and sys.stderr in the
correct order by the master.
This fixes the implementation for both running on the GUI (the slaves are no longer dumping text to
the terminal where the GUI was launched from), and on clusters where the IO is dumped on the
processing node rather than back on the master node.
........
|
2012-02-29 09:45:42
|
Tree
|
[r15398]
by
bugman
Merged revisions 15380,15385,15388-15395 via svnmerge from
svn+ssh://bugman@.../svn/relax/1.3
........
r15380 | bugman | 2012-02-27 17:46:41 +0100 (Mon, 27 Feb 2012) | 8 lines
Fix for the storage of numpy data structures in the relax save files.
If special structures are created by the user, these were not correctly stored. The precision was
often too low (printed at the numpy default of a precision of 8), and if one dimension had more than
1000 elements, ellipses would be stored instead. The ellipses would cause the save file to be
unreadable by relax.
........
r15385 | bugman | 2012-02-28 13:23:04 +0100 (Tue, 28 Feb 2012) | 3 lines
Fix for the Mf.test_write_results system test for the higher precision numpy numbers in results files.
........
r15388 | bugman | 2012-02-28 14:43:44 +0100 (Tue, 28 Feb 2012) | 6 lines
Bug fix for the multi-processor package.
The 2to3 conversion had created code which was not functional on Python2. This should not have
happened!
........
r15389 | bugman | 2012-02-28 15:50:30 +0100 (Tue, 28 Feb 2012) | 3 lines
Implemented the flush() method for the multi-processor IO streams.
........
r15390 | bugman | 2012-02-28 15:59:58 +0100 (Tue, 28 Feb 2012) | 3 lines
The flush() methods for the multi-processor IO streams now are dummy functions.
........
r15391 | bugman | 2012-02-28 16:17:23 +0100 (Tue, 28 Feb 2012) | 5 lines
Completion of a TODO in the multi-processor package.
The stdio_capture argument to the Processor classes has been eliminated.
........
r15392 | bugman | 2012-02-28 16:23:20 +0100 (Tue, 28 Feb 2012) | 3 lines
Eliminated the excessive setup_stdio_capture() method from the Processor classes.
........
r15393 | bugman | 2012-02-28 16:24:12 +0100 (Tue, 28 Feb 2012) | 3 lines
Removed a no longer relevant FIXME comment.
........
r15394 | bugman | 2012-02-28 16:30:03 +0100 (Tue, 28 Feb 2012) | 5 lines
The relax prompt mode now prints the intro message to sys.stdout.
This fixes this mode with the logging option or in the mpi4py multiprocessor mode.
........
r15395 | bugman | 2012-02-28 18:51:09 +0100 (Tue, 28 Feb 2012) | 14 lines
Big fix for the IO streams when running in the multi-processor mpi4py mode!
This affects all multi-processor implementations, excluding the standard uni-processor. The entire
stream capture design has been overhauled. The IO streams of the master processor are no longer
touched, whereas those of the slaves are captured, stored in a list of lists where the original
'S |' and 'S E|' tokens are prepended to each line as well as a number identifying the stream,
stored in the results command object, and unpacked and sent to sys.stdout and sys.stderr in the
correct order by the master.
This fixes the implementation for both running on the GUI (the slaves are no longer dumping text to
the terminal where the GUI was launched from), and on clusters where the IO is dumped on the
processing node rather than back on the master node.
........
|
2012-02-29 09:45:09
|
Tree
|
[r15397]
by
bugman
Merged revisions 15388-15395 via svnmerge from
svn+ssh://bugman@.../svn/relax/1.3
........
r15388 | bugman | 2012-02-28 14:43:44 +0100 (Tue, 28 Feb 2012) | 6 lines
Bug fix for the multi-processor package.
The 2to3 conversion had created code which was not functional on Python2. This should not have
happened!
........
r15389 | bugman | 2012-02-28 15:50:30 +0100 (Tue, 28 Feb 2012) | 3 lines
Implemented the flush() method for the multi-processor IO streams.
........
r15390 | bugman | 2012-02-28 15:59:58 +0100 (Tue, 28 Feb 2012) | 3 lines
The flush() methods for the multi-processor IO streams now are dummy functions.
........
r15391 | bugman | 2012-02-28 16:17:23 +0100 (Tue, 28 Feb 2012) | 5 lines
Completion of a TODO in the multi-processor package.
The stdio_capture argument to the Processor classes has been eliminated.
........
r15392 | bugman | 2012-02-28 16:23:20 +0100 (Tue, 28 Feb 2012) | 3 lines
Eliminated the excessive setup_stdio_capture() method from the Processor classes.
........
r15393 | bugman | 2012-02-28 16:24:12 +0100 (Tue, 28 Feb 2012) | 3 lines
Removed a no longer relevant FIXME comment.
........
r15394 | bugman | 2012-02-28 16:30:03 +0100 (Tue, 28 Feb 2012) | 5 lines
The relax prompt mode now prints the intro message to sys.stdout.
This fixes this mode with the logging option or in the mpi4py multiprocessor mode.
........
r15395 | bugman | 2012-02-28 18:51:09 +0100 (Tue, 28 Feb 2012) | 14 lines
Big fix for the IO streams when running in the multi-processor mpi4py mode!
This affects all multi-processor implementations, excluding the standard uni-processor. The entire
stream capture design has been overhauled. The IO streams of the master processor are no longer
touched, whereas those of the slaves are captured, stored in a list of lists where the original
'S |' and 'S E|' tokens are prepended to each line as well as a number identifying the stream,
stored in the results command object, and unpacked and sent to sys.stdout and sys.stderr in the
correct order by the master.
This fixes the implementation for both running on the GUI (the slaves are no longer dumping text to
the terminal where the GUI was launched from), and on clusters where the IO is dumped on the
processing node rather than back on the master node.
........
|
2012-02-29 09:44:28
|
Tree
|
[r15396]
by
bugman
Merged revisions 15380,15385,15388-15395 via svnmerge from
svn+ssh://bugman@.../svn/relax/1.3
........
r15380 | bugman | 2012-02-27 17:46:41 +0100 (Mon, 27 Feb 2012) | 8 lines
Fix for the storage of numpy data structures in the relax save files.
If special structures are created by the user, these were not correctly stored. The precision was
often too low (printed at the numpy default of a precision of 8), and if one dimension had more than
1000 elements, ellipses would be stored instead. The ellipses would cause the save file to be
unreadable by relax.
........
r15385 | bugman | 2012-02-28 13:23:04 +0100 (Tue, 28 Feb 2012) | 3 lines
Fix for the Mf.test_write_results system test for the higher precision numpy numbers in results files.
........
r15388 | bugman | 2012-02-28 14:43:44 +0100 (Tue, 28 Feb 2012) | 6 lines
Bug fix for the multi-processor package.
The 2to3 conversion had created code which was not functional on Python2. This should not have
happened!
........
r15389 | bugman | 2012-02-28 15:50:30 +0100 (Tue, 28 Feb 2012) | 3 lines
Implemented the flush() method for the multi-processor IO streams.
........
r15390 | bugman | 2012-02-28 15:59:58 +0100 (Tue, 28 Feb 2012) | 3 lines
The flush() methods for the multi-processor IO streams now are dummy functions.
........
r15391 | bugman | 2012-02-28 16:17:23 +0100 (Tue, 28 Feb 2012) | 5 lines
Completion of a TODO in the multi-processor package.
The stdio_capture argument to the Processor classes has been eliminated.
........
r15392 | bugman | 2012-02-28 16:23:20 +0100 (Tue, 28 Feb 2012) | 3 lines
Eliminated the excessive setup_stdio_capture() method from the Processor classes.
........
r15393 | bugman | 2012-02-28 16:24:12 +0100 (Tue, 28 Feb 2012) | 3 lines
Removed a no longer relevant FIXME comment.
........
r15394 | bugman | 2012-02-28 16:30:03 +0100 (Tue, 28 Feb 2012) | 5 lines
The relax prompt mode now prints the intro message to sys.stdout.
This fixes this mode with the logging option or in the mpi4py multiprocessor mode.
........
r15395 | bugman | 2012-02-28 18:51:09 +0100 (Tue, 28 Feb 2012) | 14 lines
Big fix for the IO streams when running in the multi-processor mpi4py mode!
This affects all multi-processor implementations, excluding the standard uni-processor. The entire
stream capture design has been overhauled. The IO streams of the master processor are no longer
touched, whereas those of the slaves are captured, stored in a list of lists where the original
'S |' and 'S E|' tokens are prepended to each line as well as a number identifying the stream,
stored in the results command object, and unpacked and sent to sys.stdout and sys.stderr in the
correct order by the master.
This fixes the implementation for both running on the GUI (the slaves are no longer dumping text to
the terminal where the GUI was launched from), and on clusters where the IO is dumped on the
processing node rather than back on the master node.
........
|
2012-02-29 09:43:54
|
Tree
|
[r15395]
by
bugman
Big fix for the IO streams when running in the multi-processor mpi4py mode!
This affects all multi-processor implementations, excluding the standard uni-processor. The entire
stream capture design has been overhauled. The IO streams of the master processor are no longer
touched, whereas those of the slaves are captured, stored in a list of lists where the original
'S |' and 'S E|' tokens are prepended to each line as well as a number identifying the stream,
stored in the results command object, and unpacked and sent to sys.stdout and sys.stderr in the
correct order by the master.
This fixes the implementation for both running on the GUI (the slaves are no longer dumping text to
the terminal where the GUI was launched from), and on clusters where the IO is dumped on the
processing node rather than back on the master node.
|
2012-02-28 17:51:09
|
Tree
|
