revisions 846:897 into trunk, with approval and peer review of manager (Luca).
git-svn-id: svn://svn.gna.org/svn/sgpemv2/trunk@898 3ecf2c5c-341e-0410-92b4-d18e462d057c
with a *factor* of ~640x480. That means ~64*48*10^6*3 bytes = a ~1.1 Gb RAM pixmap.
The problem didn't show in versions of cairo >= 1.2 because it automagically
clips (yeah, cairo programmers *are* smart :-))
It is temporary fixed for the moment, will be made better asap.
git-svn-id: svn://svn.gna.org/svn/sgpemv2/trunk@892 3ecf2c5c-341e-0410-92b4-d18e462d057c
to removing all of its resources and all of its processes, but
faster)
git-svn-id: svn://svn.gna.org/svn/sgpemv2/trunk@887 3ecf2c5c-341e-0410-92b4-d18e462d057c
the user may now add plugin directories,
policy directories, and set the simulation
speed.
- Added some menu voice to the main window.
git-svn-id: svn://svn.gna.org/svn/sgpemv2/trunk@872 3ecf2c5c-341e-0410-92b4-d18e462d057c
- Now CPUPolicy has a callback method for scripting languages, but it is up to derived classes to take
mutexes and set the value when needed (maybe we can improve this?)
git-svn-id: svn://svn.gna.org/svn/sgpemv2/trunk@862 3ecf2c5c-341e-0410-92b4-d18e462d057c
- Drop the old SchedulablesWidget in favor of a simple Gtk::TreeView
- Update the GuiBuilder consequentially
- FIXME: in cairo_widget.cc : have we to scale the context before or after
drawing on it?
git-svn-id: svn://svn.gna.org/svn/sgpemv2/trunk@845 3ecf2c5c-341e-0410-92b4-d18e462d057c
- Added a wizard just for show, but there is no
interesting cpu-scheduling policy to test it with
- Better textual output.
git-svn-id: svn://svn.gna.org/svn/sgpemv2/trunk@842 3ecf2c5c-341e-0410-92b4-d18e462d057c
NOTE: to use it _SG_SCHEDULABLES_TREE_WIDGET must be defined, which is obvioulsy not, by default
git-svn-id: svn://svn.gna.org/svn/sgpemv2/trunk@841 3ecf2c5c-341e-0410-92b4-d18e462d057c
imlpementation, so this version should not be considered definitive.
please give me an other day to check it.
git-svn-id: svn://svn.gna.org/svn/sgpemv2/trunk@840 3ecf2c5c-341e-0410-92b4-d18e462d057c
- Added a base class for cpu policy exceptions to make simpler their catching
- Implemented all numeric fields in dialogs with spinboxes, with bounds checking
git-svn-id: svn://svn.gna.org/svn/sgpemv2/trunk@838 3ecf2c5c-341e-0410-92b4-d18e462d057c
- Changed the way syntactically incorrect python policies are handled, we no more exit abruptly
git-svn-id: svn://svn.gna.org/svn/sgpemv2/trunk@829 3ecf2c5c-341e-0410-92b4-d18e462d057c
that the size of the history was the coordinate of the instant
to add.
- Corrected a bug in step forward, which was incorrectly assuming
that all processes started at instant 0.
- Added a wizard to show a bug in step forward, which thinks that
the simulation is terminated while it isn't since some processes
are still future.
- Commented some unused code in step forward (extendThread()).
- Still to extract methods from Extender and provide them to
step forward in a more suitable (static) way.
git-svn-id: svn://svn.gna.org/svn/sgpemv2/trunk@828 3ecf2c5c-341e-0410-92b4-d18e462d057c
- Fix throwing of exceptions from sgpem::Module
- **THOU SHALL ALWAYS UPDATE THY MAKEFILES, OR THE WRATH OF GOD
WILL DESCEND UPON THEE**
git-svn-id: svn://svn.gna.org/svn/sgpemv2/trunk@827 3ecf2c5c-341e-0410-92b4-d18e462d057c
- Setup makefile for xmlsave plugin. It still can`t be activated. why?
- Added a temporary command SAVE to commandline interface to try serialization
git-svn-id: svn://svn.gna.org/svn/sgpemv2/trunk@826 3ecf2c5c-341e-0410-92b4-d18e462d057c
at all. Compiles and runs, but the output is not
correct. Anyway, now it is just a matter of logics :P
- Corrected some minor bugs in step forward.
- Step forward is now rather messy, but still better than
four hours ago.. altough this can be hard to believe.
- Added some wizards that help finding bugs.
- Rewritten dyn process get_state as it was guilty of
premature optimization :)
git-svn-id: svn://svn.gna.org/svn/sgpemv2/trunk@825 3ecf2c5c-341e-0410-92b4-d18e462d057c
copy constructor and by the add_resource and remove_resource methods
found in ConcreteHistory.
- Scheduler now adds the requests in the queue when appropriate, and
removes them when exhausted.
- Still to implement the management of the state of requests depending
on their position in the queue
- Still to implement the way threads block depending on the state of
their requests
- step_forward now reuses some bunch of code taken from the prototype
git-svn-id: svn://svn.gna.org/svn/sgpemv2/trunk@824 3ecf2c5c-341e-0410-92b4-d18e462d057c
derived widget
- Add History::set_notify_enabled() for usage by the frontend
(for example, when loading from file, you'll want to call
set_notify_enabled(false) before starting)
git-svn-id: svn://svn.gna.org/svn/sgpemv2/trunk@823 3ecf2c5c-341e-0410-92b4-d18e462d057c
an abstract base class, and I'll start working out the
adjustement / resizing problems (they're not easy to work out
correctly).
git-svn-id: svn://svn.gna.org/svn/sgpemv2/trunk@822 3ecf2c5c-341e-0410-92b4-d18e462d057c
- Added exception specifications to the methods of the serializer interface and xml plugin
git-svn-id: svn://svn.gna.org/svn/sgpemv2/trunk@821 3ecf2c5c-341e-0410-92b4-d18e462d057c
- Make test-history run by commenting some "dangerous" lines
- Added a couple of comments to SchedulablesWidget::update() in the hope the may be useful to others in finding where the bug resides...
git-svn-id: svn://svn.gna.org/svn/sgpemv2/trunk@818 3ecf2c5c-341e-0410-92b4-d18e462d057c
their elapsed time
- requests are now behaving nicely, since they are raised at the correct
instant, and terminated at the right instant
- the ready queues are still not ready, tough :(, anyway, now there is
really nothing else to do before them.
- stepforward is looking much more uglier (if possible).
git-svn-id: svn://svn.gna.org/svn/sgpemv2/trunk@816 3ecf2c5c-341e-0410-92b4-d18e462d057c
when that thread had just been given the cpu.
fixed a bug involving text-simulation displaying the wrong timestamp for
each snaphot. Now the first snapshot printed is correctly tagged "-1".
added (for the sake of lazyness) a method to readyqueue letting anyone
delete the first element on the queue. Not having this would imply
rebuilding the whole queue each time a thread was selected to run.
Small commits avoid complex mergings.
git-svn-id: svn://svn.gna.org/svn/sgpemv2/trunk@814 3ecf2c5c-341e-0410-92b4-d18e462d057c
tchernobog
elvez
matrevis