169483 - RFE: Add prefs to enable/disable double-buffering

Status: RESOLVED FIXED

(Core :: Web Painting, defect)

Description

[Roland Mainz]

RFE: Add prefs to enable/disable double-buffering instead of using a
compile-time switch (--disable-double-buffer).

Comment 1

[Roland Mainz]

Taking myself...

Comment 2

[Roland Mainz]

Attachment: Patch for 2002-09-14-08-trunk

Comment 3

[Robert O'Callahan (:roc) (email my personal email if necessary)]

Comment on attachment 99685 [details] [diff] [review]
Patch for 2002-09-14-08-trunk

/* Check the prefs we we should do doublebuffering or not... */
doesn't make sense, how about
/* Check the prefs to see whether we should do double buffering or not */

with that, r=roc+moz

Comment 4

[Roland Mainz]

Attachment: New patch to match r=roc+moz

Comment 5

[Robert O'Callahan (:roc) (email my personal email if necessary)]

Comment on attachment 99688 [details] [diff] [review]
New patch to match r=roc+moz

r=roc+moz

Comment 6

[kinmoz]

Comment on attachment 99688 [details] [diff] [review]
New patch to match r=roc+moz

So where are the changes that set the default value? Aren't there some .js
changes that need to be made in modules/libpref? If we checkin without checking
in a default, won't double buffering be disabled on all platforms?

Also note that the result of the following call will also overwrite whatever
the pref says on some platforms like the mac:

aContext->UseBackbuffer(&contextWantsBackBuffer);

Comment 7

[hacker formerly known as [e-mail]]

Comment on attachment 99688 [details] [diff] [review]
New patch to match r=roc+moz

r=cls on the build change.

Comment 8

[Roland Mainz]

kin wrote:
> (From update of attachment 99688 [details] [diff] [review])
> So where are the changes that set the default value? Aren't there some .js
> changes that need to be made in modules/libpref? If we checkin without 
> checking in a default, won't double buffering be disabled on all platforms?

From the patch:
-- snip --
+static
+PRBool DoDoubleBuffering(void)
+{
+  static PRBool got_doublebuffer_prefs = PR_FALSE;
+  static PRBool do_doublebuffering     = PR_TRUE;
-- snip --

|do_doublebuffering| is set to |PR_TRUE| by default - and |GetBoolPref()| does
not alter that value if the matching pref is not present ((nearly) all of our
|GetBoolPref()| callers do that... :)
Therefore we do double-buffering except the prefs says something else.

I did not provide an entry for prefs.js - this switch is more for developers who
want to see how inefficient our code is (BTW: the number of (unneccesary!!) full
redraws and how we render the scrollbars is _shocking_... ;-(((( ) ...

> Also note that the result of the following call will also overwrite whatever
> the pref says on some platforms like the mac:
>
> aContext->UseBackbuffer(&contextWantsBackBuffer);

The new prefs is a replacement for the --disable-double-buffer "configure"
switch (and exactly does the same).
I've looked at the line above but I think it's better not to play with that -
all h*ll may break loose if we try to force non-double-buffered mode in all
cases.

Comment 9

[kinmoz]

Comment on attachment 99688 [details] [diff] [review]
New patch to match r=roc+moz

sr=[e-mail]

Comment 10

[Robert O'Callahan (:roc) (email my personal email if necessary)]

Actually, Roland, you should add an entry to all.js.

Also, when you add the all.js setting to the patch, would you mind changing your
local variables to interCaps? Thanks!

BTW, the UseBackbuffer check can only turn double-buffering OFF, not on.

Comment 11

[Roland Mainz]

Robert O'Callahan wrote:
> Actually, Roland, you should add an entry to all.js.

OK... but I can't do that now, I am away for the rest of the week... - new patch
on sunday or monday... ;-((

> Also, when you add the all.js setting to the patch, would you mind changing 
> your local variables to interCaps?

What are "interCaps" ?!

Comment 12

[Robert O'Callahan (:roc) (email my personal email if necessary)]

not interCaps: do_double_buffering
interCaps: doDoubleBuffering

Comment 13

[Roland Mainz]

Attachment: New patch for 2002-09-18-08-trunk per roc+moz's comments

Comment 14

[Christian :Biesinger (don't email me, ping me on IRC)]

checked in

Comment 15

[Adam D. Moss]

This is only useful/interesting for debugging, right?

Comment 16

[Roland Mainz]

Adam D. Moss wrote:
> This is only useful/interesting for debugging, right?

Yes and no. It speeds-up the rendering performance on some dumb framebuffers
(those which don't have "accerlated" rect copy) therefore it may help endusers
in some rare cases, too... :)

Comment 17

[Roland Mainz]

Usage:
Set 
-- snip --
user_pref("viewmanager.do_doublebuffering", false);
-- snip --
to disable double-buffering.