Enlightenment CVS committal
Author : barbieri
Project : e17
Module : libs/edje
Dir : e17/libs/edje/src/lib
Modified Files:
edje_calc.c
Log Message:
Avoid useless huge implicit memcpy.
The "simple" block:
p3 = p1;
was doing an implicity memcpy() responsible for about 1% of processing
time when no animation happens.
Patch by Cedric BAIL.
===================================================================
RCS file: /cvs/e/e17/libs/edje/src/lib/edje_calc.c,v
retrieving revision 1.115
retrieving revision 1.116
diff -u -3 -r1.115 -r1.116
--- edje_calc.c 21 Feb 2008 18:35:27 -0000 1.115
+++ edje_calc.c 27 Feb 2008 14:27:10 -0000 1.116
@@ -1288,7 +1288,7 @@
static void
_edje_part_recalc(Edje *ed, Edje_Real_Part *ep, int flags)
{
- Edje_Calc_Params p1, p2, p3;
+ Edje_Calc_Params p1, p2, p3, *pf;
Edje_Part_Description *chosen_desc;
double pos = 0.0;
@@ -1421,16 +1421,15 @@
p3.gradient.id = pos > 0.5 ? p2.gradient.id : p1.gradient.id;
p3.gradient.type = pos > 0.5 ? p2.gradient.type : p1.gradient.type;
+
+ pf = &p3;
}
else
{
- p3 = p1;
+ pf = &p1;
}
- ep->req.x = p3.req.x;
- ep->req.y = p3.req.y;
- ep->req.w = p3.req.w;
- ep->req.h = p3.req.h;
+ ep->req = pf->req;
if (ep->drag.need_reset)
{
@@ -1455,18 +1454,18 @@
case EDJE_PART_TYPE_TEXTBLOCK:
case EDJE_PART_TYPE_GRADIENT:
evas_object_color_set(ep->object,
- (p3.color.r * p3.color.a) / 255,
- (p3.color.g * p3.color.a) / 255,
- (p3.color.b * p3.color.a) / 255,
- p3.color.a);
- if (p3.visible) evas_object_show(ep->object);
+ (pf->color.r * pf->color.a) / 255,
+ (pf->color.g * pf->color.a) / 255,
+ (pf->color.b * pf->color.a) / 255,
+ pf->color.a);
+ if (pf->visible) evas_object_show(ep->object);
else evas_object_hide(ep->object);
/* move and resize are needed for all previous object => no break
here. */
case EDJE_PART_TYPE_SWALLOW:
case EDJE_PART_TYPE_GROUP:
/* visibility and color have no meaning on SWALLOW and GROUP
part. */
- evas_object_move(ep->object, ed->x + p3.x, ed->y + p3.y);
- evas_object_resize(ep->object, p3.w, p3.h);
+ evas_object_move(ep->object, ed->x + pf->x, ed->y + pf->y);
+ evas_object_resize(ep->object, pf->w, pf->h);
break;
case EDJE_PART_TYPE_TEXT:
/* This is correctly handle in _edje_text_recalc_apply at the
moment. */
@@ -1477,13 +1476,13 @@
switch (ep->part->type)
{
case EDJE_PART_TYPE_TEXT:
- _edje_text_recalc_apply(ed, ep, &p3, chosen_desc);
+ _edje_text_recalc_apply(ed, ep, pf, chosen_desc);
break;
case EDJE_PART_TYPE_IMAGE:
- _edje_image_recalc_apply(ed, ep, &p3, chosen_desc, pos);
+ _edje_image_recalc_apply(ed, ep, pf, chosen_desc, pos);
break;
case EDJE_PART_TYPE_GRADIENT:
- _edje_gradient_recalc_apply(ed, ep, &p3, chosen_desc);
+ _edje_gradient_recalc_apply(ed, ep, pf, chosen_desc);
break;
case EDJE_PART_TYPE_RECTANGLE:
case EDJE_PART_TYPE_SWALLOW:
@@ -1500,21 +1499,21 @@
//// done via the clipper anyway. this created bugs when objects had their
//// colro set and were swallowed - then had their color changed.
// evas_object_color_set(ep->swallowed_object,
-// (p3.color.r * p3.color.a) / 255,
-// (p3.color.g * p3.color.a) / 255,
-// (p3.color.b * p3.color.a) / 255,
-// p3.color.a);
- evas_object_move(ep->swallowed_object, ed->x + p3.x, ed->y + p3.y);
- evas_object_resize(ep->swallowed_object, p3.w, p3.h);
- if (p3.visible) evas_object_show(ep->swallowed_object);
+// (pf->color.r * pf->color.a) / 255,
+// (pf->color.g * pf->color.a) / 255,
+// (pf->color.b * pf->color.a) / 255,
+// pf->color.a);
+ evas_object_move(ep->swallowed_object, ed->x + pf->x, ed->y +
pf->y);
+ evas_object_resize(ep->swallowed_object, pf->w, pf->h);
+ if (pf->visible) evas_object_show(ep->swallowed_object);
else evas_object_hide(ep->swallowed_object);
}
}
- ep->x = p3.x;
- ep->y = p3.y;
- ep->w = p3.w;
- ep->h = p3.h;
+ ep->x = pf->x;
+ ep->y = pf->y;
+ ep->w = pf->w;
+ ep->h = pf->h;
ep->calculated |= flags;
ep->calculating = FLAG_NONE;
-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
enlightenment-cvs mailing list
enlightenment-cvs@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/enlightenment-cvs
