| Author: nicholas |
Date: 2002-10-11.02:08:49 |
If a Graphic happens to be damaged during a draw
traversal, the server will block waiting for a new
event instead of repairing the damage already
incurred.
See Redisplay in Fresco:
http://www2.fresco.org/docs/papers/redisplay.ps.gz
|
| Author: nicholas |
Date: 2002-10-11.02:11:13 |
More exactly, see under the heading "Redisplay
thread" on page 6.
|
| Author: njs |
Date: 2002-10-11.02:23:24 |
There was also some discussion of this on the
mailing list; it starts at:
http://lists.fresco.org/pipermail/fresco-devel/2002-June/018616.html
and continues in:
http://lists.fresco.org/pipermail/fresco-devel/2002-July/018620.html
http://lists.fresco.org/pipermail/fresco-devel/2002-July/018622.html
http://lists.fresco.org/pipermail/fresco-devel/2002-July/018625.html
|
| Author: nicholas |
Date: 2002-10-12.14:24:46 |
patch adds "has_event()" test to Console and
implements it for GGI and SDL. Does not change
event loop.
|
| Author: nicholas |
Date: 2002-10-30.01:00:14 |
Stefan approved the change that does what I
describe my patch as doing. I'm going to want to
re-read it before I commit it. (Or someone else
can read it.)
|
| Author: nicholas |
Date: 2002-12-06.01:28:54 |
If the contents of my patch are correct, this is a
pretty accessible project for someone. My patch
just adds the functions necessary to correct the
main-loop logic.
That logic is ScreenManager::run , which is a
short simple little function.
Test it with something like ggiwrapper -r `which
ggi-dots` and see that it responds correctly. Also
test with 'top' to make certain that the server
isn't burning 100% CPU.
|
| Author: nicholas |
Date: 2003-01-04.13:51:37 |
This patch has landed. Those interested in taking
up this bug should know that there is a new
function, _emanager->has_event(), which guarantees
that the Console will not block when you next call
next_event().
|
| Author: nicholas |
Date: 2003-01-12.02:25:11 |
Patch backed out.
Mid-traversal damage will still not be calculated.
Stefan has commented that it really shouldn't be.
A fix to the SDLConsole to make wakeup work as
Stefan explained it to is now checked in.
|
| Author: nicholas |
Date: 2003-01-12.15:25:59 |
If a DrawTraversal is underway and you have a
timer that triggers need_resize during that
traversal, the damage will be stored until the
next event/update, instead of being dealt with at
the next available opportunity.
Mutating this bug.
|
| Author: nicholas |
Date: 2003-01-12.15:51:33 |
The trouble is that Console::wakeup() isn't being
called for mid-Traversal damage as it is for other
damage. Even if it were, the GGIConsole probably
wouldn't handle the case correctly since it's
using select on a pipe ... but that other thread
would be buzy so both threads would lock and
there'd be a deadlock. For the SDLConsole, wakeup
would just push a new event on the SDL event queue.
|
| Author: anonymous |
Date: 2009-11-06.01:23:06 |
eIGAcj <a href="http://ecczewddsauz.com/">ecczewddsauz</a>, [url=http://pfervigtthlh.com/]pfervigtthlh[/url], [link=http://vklwhjiiayvn.com/]vklwhjiiayvn[/link], http://ywhuiasjwdyp.com/ |
|
| Date |
User |
Action |
Args |
| 2002-10-11 02:08:50 | nicholas | create | files: (no value) status: new severity: normal title: redraw loop blocks damage repair nosy:
nicholas topics: (no value) messages:
msg435 priority: normal assignedto: (no value) superseder: (no value) bug_type: behavior components:
Berlin core dependson: (no value) platforms: (no value) resolution: (no value) |
| 2002-10-11 02:11:13 | nicholas | set | messages:
+ msg436 |
| 2002-10-11 02:12:57 | nicholas | set | title: redraw loop blocks for events leaving damage pending |
| 2002-10-11 02:23:24 | njs | set | messages:
+ msg437 title: redraw loop blocks damage repair |
| 2002-10-11 17:09:06 | tobias | set | nosy:
+ tobias |
| 2002-10-12 14:24:47 | nicholas | set | files:
+ event.patch messages:
+ msg443 |
| 2002-10-30 01:00:15 | nicholas | set | messages:
+ msg609 |
| 2002-12-06 01:28:54 | nicholas | set | topics:
+ bitesize, entrylevel messages:
+ msg1010 |
| 2003-01-04 13:51:37 | nicholas | set | messages:
+ msg1259 |
| 2003-01-12 02:25:12 | nicholas | set | resolution: (no value) messages:
+ msg1361 assignedto: (no value) |
| 2003-01-12 02:25:56 | nicholas | set | status: new |
| 2003-01-12 15:26:00 | nicholas | set | status: closed assignedto: nicholas messages:
+ msg1367 resolution: invalid title: redraw loop blocks damage repair |
| 2003-01-12 15:26:37 | nicholas | set | bug_type: behavior |
| 2003-01-12 15:51:33 | nicholas | set | messages:
+ msg1369 |
| 2009-11-06 01:23:18 | anonymous | set | platforms:
+ GNU/Linux, not applicable messages:
+ msg4552 components:
+ Berlin server, C++ clients, python clients, website & tools, tracker test, GGI Console, SDL Console, documentation, daVinci, test framework nosy:
+ sami, plfiorini, redmondp, neiljp, jnf, njs |