svn+ssh://bugman@.../svn/relax/1.3
................
r11811 | bugman | 2010-12-14 11:11:03 +0100 (Tue, 14 Dec 2010) | 8 lines
Added a system test for bug #17276.
This bug (https://gna.org/bugs/index.php?17276) was reported by Leanne Minall <lm7902 att bristol
dot ac dott uk>.
The test passes though as the problem is not in this test data.
................
r11812 | bugman | 2010-12-14 11:23:31 +0100 (Tue, 14 Dec 2010) | 3 lines
Bug #17276 is now triggered in the system test.
................
r11817 | bugman | 2010-12-14 18:22:33 +0100 (Tue, 14 Dec 2010) | 6 lines
Removed some checks in the generic_fns.mol_res_spin.molecule_loop() function which was triggering bugs.
The old checking code is of no use, as all types of molecule IDs should be accepted! This only
recently became an issue for some unknown reason?!?!
................
r11820 | bugman | 2010-12-14 23:47:05 +0100 (Tue, 14 Dec 2010) | 7 lines
An execution lock has been added to the status object.
This allows parts of relax to prevent others from running, which is especially useful in certain
UIs. A special Exec_lock class object has been created to simulate a proper threading.Lock object,
but with certain additional features.
................
r11821 | bugman | 2010-12-14 23:47:33 +0100 (Tue, 14 Dec 2010) | 3 lines
All of the auto-analyses now lock relax's execution while they are in operation.
................
r11826 | bugman | 2010-12-15 01:33:13 +0100 (Wed, 15 Dec 2010) | 5 lines
The Mf.bug_14941_local_tm_global_selection system test now prints out some status info.
This is to test the status container.
................
r11827 | bugman | 2010-12-15 01:35:28 +0100 (Wed, 15 Dec 2010) | 5 lines
Fixes for the status object by renaming __init__() to setup() and calling it only once.
The object was not acting as a singleton because of the presence of __init__(). This is now fixed.
................
r11828 | bugman | 2010-12-15 01:36:05 +0100 (Wed, 15 Dec 2010) | 3 lines
The status object is being initialised and setup before execution of the main relax class.
................
r11829 | bugman | 2010-12-15 01:36:48 +0100 (Wed, 15 Dec 2010) | 3 lines
The status object is not being initialised on import, to allow controlled setup upon start up.
................
r11830 | bugman | 2010-12-15 01:45:19 +0100 (Wed, 15 Dec 2010) | 6 lines
Security fix - execfile() is no longer used to run relax scripts.
Instead a custom function exec_script() is being used to call runpy.run_module(). exec_script()
also locks and releases the relax execution lock before and after operation.
................
r11831 | bugman | 2010-12-15 09:41:27 +0100 (Wed, 15 Dec 2010) | 3 lines
The execution lock is released by the interpreter in all Error situations.
................
r11837 | bugman | 2010-12-16 11:52:56 +0100 (Thu, 16 Dec 2010) | 6 lines
The debugging flag now disables the execution locking mechanism.
The acquire() and release() calls are also recorded in the 'lock.log' file to help with lock
debugging.
................
r11838 | bugman | 2010-12-16 11:53:51 +0100 (Thu, 16 Dec 2010) | 3 lines
Missing module import.
................
r11839 | bugman | 2010-12-16 11:56:02 +0100 (Thu, 16 Dec 2010) | 3 lines
Shifted the initialisation of the status object to be after the __main__.debug initialisation.
................
r11840 | bugman | 2010-12-16 12:17:36 +0100 (Thu, 16 Dec 2010) | 5 lines
The system and unit tests are placing their names into the status object.
This will be used for debugging.
................
r11841 | bugman | 2010-12-16 12:18:14 +0100 (Thu, 16 Dec 2010) | 3 lines
The unit/system test name is being placed into the locking log when debugging.
................
r11842 | bugman | 2010-12-16 12:37:36 +0100 (Thu, 16 Dec 2010) | 3 lines
The execution lock mechanism no longer blocks scripts from running within scripts.
................
r11843 | bugman | 2010-12-16 12:49:07 +0100 (Thu, 16 Dec 2010) | 3 lines
Nested scripts no longer acquire the execution lock.
................
r11844 | bugman | 2010-12-16 15:19:14 +0100 (Thu, 16 Dec 2010) | 5 lines
Better locking support for nested scripts.
The execution lock directly handles the locking of the nesting of scripts.
................
r11848 | bugman | 2010-12-17 10:19:50 +0100 (Fri, 17 Dec 2010) | 3 lines
The 'scons clean' target now removes temporary relax save files.
................
r11849 | bugman | 2010-12-17 10:36:56 +0100 (Fri, 17 Dec 2010) | 3 lines
Improved execution locking with support for auto analysis and script launching from within a script.
................
r11854 | bugman | 2010-12-17 12:05:30 +0100 (Fri, 17 Dec 2010) | 3 lines
Initialized merge tracking via "svnmerge" with revisions "1-11813" from
svn+ssh://bugman@.../svn/relax/branches/peak_list_handling
................
r11855 | bugman | 2010-12-17 12:10:03 +0100 (Fri, 17 Dec 2010) | 58 lines
Merged revisions 11815-11816,11819,11832-11835,11847,11850-11853 via svnmerge from
svn+ssh://bugman@.../svn/relax/branches/peak_list_handling
........
r11815 | bugman | 2010-12-14 17:51:09 +0100 (Tue, 14 Dec 2010) | 3 lines
Modified the bug_17276 system test data so that there is some noise simulated.
........
r11816 | bugman | 2010-12-14 17:52:06 +0100 (Tue, 14 Dec 2010) | 3 lines
Peak intensity data is now stored as a dictionary with the spectrum_id arg as keys.
........
r11819 | bugman | 2010-12-14 19:02:59 +0100 (Tue, 14 Dec 2010) | 5 lines
Fix for the Relax_fit.test_read_sparky system test.
This was solely a problem within the test and its data.
........
r11832 | bugman | 2010-12-15 12:17:28 +0100 (Wed, 15 Dec 2010) | 3 lines
Switched __errors_repl() to use dictionaries rather than lists.
........
r11833 | bugman | 2010-12-15 18:33:22 +0100 (Wed, 15 Dec 2010) | 3 lines
Converted the specific relaxation curve fitting code to use the dictionary data structures.
........
r11834 | bugman | 2010-12-15 18:33:44 +0100 (Wed, 15 Dec 2010) | 3 lines
The Monte Carlo simulation code can now handle dictionary type data.
........
r11835 | bugman | 2010-12-15 18:51:02 +0100 (Wed, 15 Dec 2010) | 3 lines
The Grace code can now handle dictionary type data.
........
r11847 | bugman | 2010-12-17 10:03:04 +0100 (Fri, 17 Dec 2010) | 3 lines
Fixes for the system tests - the data is now in dictionaries.
........
r11850 | bugman | 2010-12-17 10:48:25 +0100 (Fri, 17 Dec 2010) | 3 lines
Redesigned the specific API common method _return_value_general() to handle dictionaries.
........
r11851 | bugman | 2010-12-17 11:12:52 +0100 (Fri, 17 Dec 2010) | 3 lines
Fix for the API base method _return_value_general() - not all values have errors associated with them.
........
r11852 | bugman | 2010-12-17 11:43:27 +0100 (Fri, 17 Dec 2010) | 3 lines
The NOE specific analysis now uses the dictionary type structure for the intensity data structures.
........
r11853 | bugman | 2010-12-17 11:44:35 +0100 (Fri, 17 Dec 2010) | 6 lines
The list-type data can also be a numpy array, so the Monte Carlo code can now handle this.
Previously the data was being converted from an ndarray to a list, but now this is explicitly
handled.
........
................
r11856 | bugman | 2010-12-17 12:10:25 +0100 (Fri, 17 Dec 2010) | 3 lines
Removed merge tracking for "svnmerge" for
svn+ssh://bugman@.../svn/relax/branches/peak_list_handling
................
r11858 | bugman | 2010-12-17 12:46:00 +0100 (Fri, 17 Dec 2010) | 5 lines
Removed all execfile() calls by switching to the script() user function.
This is for security and for handling the execution locking mechanism properly.
................
r11859 | bugman | 2010-12-17 12:57:49 +0100 (Fri, 17 Dec 2010) | 3 lines
Renamed all the frame order scripts so that they have valid module names and can be imported.
................
r11860 | bugman | 2010-12-17 13:00:51 +0100 (Fri, 17 Dec 2010) | 3 lines
Fix for running scripts - the '.' character is now disallowed as module import cannot work.
................
r11861 | bugman | 2010-12-17 13:04:44 +0100 (Fri, 17 Dec 2010) | 3 lines
More informative nesting information in the execution lock debugging.
................
r11862 | bugman | 2010-12-17 13:23:24 +0100 (Fri, 17 Dec 2010) | 3 lines
Modification of the execution lock debugging print outs.
................
r11863 | bugman | 2010-12-17 14:11:23 +0100 (Fri, 17 Dec 2010) | 5 lines
Bug fix - the script path is now being removed from sys.path.
This was causing incorrect scripts to be loaded if exec_script() is run multiple times.
................
r11864 | bugman | 2010-12-17 14:51:20 +0100 (Fri, 17 Dec 2010) | 6 lines
Fix for the execution locking when an auto-analysis is launched from a script.
The lock is not released and a new one acquired. Instead the mechanism is script, just as it is
for nested scripting.
................
r11865 | bugman | 2010-12-17 15:03:00 +0100 (Fri, 17 Dec 2010) | 3 lines
Added a system test for nested scripting.
................
r11866 | bugman | 2010-12-17 15:04:02 +0100 (Fri, 17 Dec 2010) | 5 lines
Fix for the Generic.test_nested_scripting() system test.
A data pipe is needed.
................
r11867 | bugman | 2010-12-17 15:13:51 +0100 (Fri, 17 Dec 2010) | 5 lines
Nested scripting is now better supported.
The full path of the nested script is not needed if is in the same directory as the parent script.
................
