Activity for Mihael Hategan
-
Mihael Hategan posted a comment on ticket #752
True, Swing uses double-bufferring. Swing also uses a "paint everything in one thread" model. The problem is that repaint requests for the 3d view can only be serviced between invocations of PlanComponent.paintComponent(). The patch moves the heavy work in paintComponent() off the AWT thread, so that the 3d view is updated as soon as the GPU completes it. The PlanComponent view is also updated as soon as it completes, so there's no penalty for PlanComponent. I'm not sure if the buffer size changes...
-
3D view performance
-
added method to check if node has ancestor of a...
-
cancel timeout checks when channel is closed (r...
-
set polling intervals in passive mode
-
fixed some score calculations
-
fixed scheduler property parsing error(s) (swif...
-
fixing future notification loops: each relevant...
-
removed reference to tc.data
-
avoid passing internal attributes to service
-
added hidden option to argument parser and made...
-
changed environment to a map rather than a channel
-
added public method to calculate the site TScore
-
made class visible since a public method return...
-
changed probe interval to 60 seconds
-
not used
-
not used any more
-
resources is not needed any more
-
retired setup
-
removed graph editor dependency
-
removed logging of job details; it is now done ...
-
use the new configuration id scheme (supporting...
-
support for multiple configurations on the same...
-
moved message to debug
-
only output diagnostic messages if at least one...
-
added option to specify concurrency when using ...
-
retired graph editor
-
added ability to remove service attributes
-
added jobspec bulk environment variable setting...
-
added jobspec bulk environment variable setting...
-
changed BoundContact.host to BoundContact.name;...
-
utility functions to calculate throttling param...
-
no need to keep stack frame count if starting a...
-
removed breakpoint-catching code
-
unclevered statement
-
fixed space before stderr redirect
-
fixed space before stderr redirect
-
properly redirect stdout and stderr from ssh wh...
-
properly redirect stdout and stderr from ssh wh...
-
shell provider
-
removed WS provider
-
TaskHandler.getName()
-
TaskHandler.getName()
-
TaskHandler.getName()
-
TaskHandler.getName()
-
TaskHandler.getName()
-
added a name property to task handlers and make...
-
fixed FileLocation.toString()
-
fixed bug where multiple stageout notifications...
-
fix for bug #1197 (port from 4.1.10)
-
an improved stack frame allocation scheme and m...
-
save some memory by using a constant status for...
-
misc improvements
-
a thread set for when the number of threads is ...
-
more efficient memory using a separate implemen...
-
allow null service contact
-
some memory optimizations
-
patch to use compression in GSS streams when 1....
-
deal better with errors in the lead job when a ...
-
better handling of errors inside job subprocess
-
fixed npe on instantiation of SSH (and others) ...
-
allow sub-classes to get the method used for in...
-
don't enable stream compression by default (bug...
-
wildcard stageouts for local provider
-
use new permission stuff
-
wildcard stageouts for coasters
-
use new permission stuff
-
compress GSS channel communications
-
added some convenience classes for path expansion
-
constructor to allow duplication of a RemoteFile
-
added filtered list method to file resources
-
more efficient permissions objects
-
re-organized some functions a bit and added joi...
-
fixed premature removal of cached entries for c...
-
fixed manually specified port causing exception...
-
fix for bug #1197
-
don't instantiate possibly unnecessary integers
-
(potentially) fixed generation of autoCA proxie...
-
moved memory logging code to a place that makes...
-
added code to regularly log memory usage of the...
-
added code to regularly log memory usage of the...
-
fixed thread creation and throw exceptions if a...
-
release lock before calling pthread_join
-
Don't try to lock soft image things when no sof...
-
fixed the miriad of soft image locking issues
-
make the soft image use normal staging mechanis...
1