* Changed the AGGTextRenderer to use the new UTF8 handling

* Added more char codes to is_white_space(), should be all I think

Sorry if I stepped on your toes Stephan, but I wanted these changes flushed before I leave for holidays :-).

git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@16501 a95241bf-73f2-0310-859d-f6bbb57e9c96
This commit is contained in:
Michael Lotz 2006-02-23 09:04:26 +00:00
parent 6990418519
commit 20a4e7265c
4 changed files with 179 additions and 218 deletions

View File

@ -159,10 +159,8 @@ UTF8ToCharCode(const char **bytes)
}
if ((*bytes)[0] == 0) {
/* We do not advance beyond the terminating 0. Just pad any further
request with spaces. */
result += 0x20;
return result;
/* We do not advance beyond the terminating 0. */
return 0x00;
}
result += (*bytes)[0];

View File

@ -81,11 +81,21 @@ CubicToFunc(FT_Vector *control1, FT_Vector *control2, FT_Vector *to, void *user)
inline bool
is_white_space(uint16 glyph)
is_white_space(uint32 charCode)
{
// TODO: handle them all!
if (glyph == ' ' || glyph == B_TAB)
return true;
switch (charCode) {
case 0x0009: /* tab */
case 0x000b: /* vertical tab */
case 0x000c: /* form feed */
case 0x0020: /* space */
case 0x00a0: /* non breaking space */
case 0x000a: /* line feed */
case 0x000d: /* carriage return */
case 0x2028: /* line separator */
case 0x2029: /* paragraph separator */
return true;
}
return false;
}
@ -428,10 +438,9 @@ ServerFont::GetEscapements(const char charArray[], int32 numChars,
const char *string = charArray;
for (int i = 0; i < numChars; i++) {
// Do this first as UTF8ToCharCode advances string.
escapementArray[i].x = is_white_space(*string) ? delta.space : delta.nonspace;
FT_Load_Char(face, UTF8ToCharCode(&string), FT_LOAD_NO_BITMAP);
uint32 charCode = UTF8ToCharCode(&string);
FT_Load_Char(face, charCode, FT_LOAD_NO_BITMAP);
escapementArray[i].x = is_white_space(charCode) ? delta.space : delta.nonspace;
escapementArray[i].x += float(face->glyph->advance.x) / 64;
escapementArray[i].y = -float(face->glyph->advance.y) / 64;
escapementArray[i].x /= fSize;
@ -466,10 +475,9 @@ ServerFont::GetEscapements(const char charArray[], int32 numChars,
const char *string = charArray;
for (int i = 0; i < numChars; i++) {
// Do this first as UTF8ToCharCode advances string.
widthArray[i] = is_white_space(*string) ? delta.space : delta.nonspace;
FT_Load_Char(face, UTF8ToCharCode(&string), FT_LOAD_NO_BITMAP);
uint32 charCode = UTF8ToCharCode(&string);
FT_Load_Char(face, charCode, FT_LOAD_NO_BITMAP);
widthArray[i] = is_white_space(charCode) ? delta.space : delta.nonspace;
widthArray[i] += float(face->glyph->metrics.horiAdvance) / 64.0;
widthArray[i] /= fSize;
}
@ -493,14 +501,15 @@ ServerFont::GetBoundingBoxesAsString(const char charArray[], int32 numChars,
const char *string = charArray;
for (int i = 0; i < numChars; i++) {
uint32 charCode = UTF8ToCharCode(&string);
if (stringEscapement) {
if (i > 0)
rectArray[i].OffsetBy(is_white_space(*string) ? delta.space / 2.0 : delta.nonspace / 2.0, 0.0);
rectArray[i].OffsetBy(is_white_space(charCode) ? delta.space / 2.0 : delta.nonspace / 2.0, 0.0);
rectArray[i].OffsetBy(is_white_space(*string) ? delta.space / 2.0 : delta.nonspace / 2.0, 0.0);
rectArray[i].OffsetBy(is_white_space(charCode) ? delta.space / 2.0 : delta.nonspace / 2.0, 0.0);
}
FT_Load_Char(face, UTF8ToCharCode(&string), FT_LOAD_NO_BITMAP);
FT_Load_Char(face, charCode, FT_LOAD_NO_BITMAP);
if (i < numChars - 1) {
rectArray[i + 1].left = rectArray[i + 1].right = rectArray[i].left
+ face->glyph->metrics.horiAdvance / 64.0;

View File

@ -20,6 +20,7 @@
#include <agg_trans_affine.h>
#include "AGGTextRenderer.h"
#include "moreUTF8.h"
#define FLIP_Y false
@ -42,13 +43,22 @@ rect_to_int(BRect r,
}
// is_white_space
inline bool
is_white_space(uint16 glyph)
is_white_space(uint32 charCode)
{
// TODO: handle them all!
if (glyph == ' ' || glyph == B_TAB)
return true;
switch (charCode) {
case 0x0009: /* tab */
case 0x000b: /* vertical tab */
case 0x000c: /* form feed */
case 0x0020: /* space */
case 0x00a0: /* non breaking space */
case 0x000a: /* line feed */
case 0x000d: /* carriage return */
case 0x2028: /* line separator */
case 0x2029: /* paragraph separator */
return true;
}
return false;
}
@ -156,7 +166,7 @@ AGGTextRenderer::Unset()
// TODO ? release some kind of reference count on the ServerFont?
}
// RenderString
BRect
AGGTextRenderer::RenderString(const char* string,
uint32 length,
@ -178,121 +188,121 @@ AGGTextRenderer::RenderString(const char* string,
Transformable transform(fEmbeddedTransformation);
transform.TranslateBy(baseLine);
uint32 glyphCount;
fCurves.approximation_scale(transform.scale());
if (_PrepareUnicodeBuffer(string, length, &glyphCount) >= B_OK) {
fCurves.approximation_scale(transform.scale());
// use a transformation behind the curves
// (only if glyph->data_type == agg::glyph_data_outline)
// in the pipeline for the rasterizer
typedef agg::conv_transform<conv_font_curve_type, agg::trans_affine>
conv_font_trans_type;
conv_font_trans_type transformedOutline(fCurves, transform);
// use a transformation behind the curves
// (only if glyph->data_type == agg::glyph_data_outline)
// in the pipeline for the rasterizer
typedef agg::conv_transform<conv_font_curve_type, agg::trans_affine>
conv_font_trans_type;
conv_font_trans_type transformedOutline(fCurves, transform);
double x = 0.0;
double y0 = 0.0;
double y = y0;
uint16* p = (uint16*)fUnicodeBuffer;
double advanceX = 0.0;
double advanceY = 0.0;
bool firstLoop = true;
double x = 0.0;
double y0 = 0.0;
double y = y0;
// for when we bypass the transformation pipeline
BPoint transformOffset(0.0, 0.0);
transform.Transform(&transformOffset);
double advanceX = 0.0;
double advanceY = 0.0;
uint32 charCode;
while ((charCode = UTF8ToCharCode(&string)) > 0) {
// line break (not supported by R5)
/*if (charCode == '\n') {
y0 += LineOffset();
x = 0.0;
y = y0;
advanceX = 0.0;
advanceY = 0.0;
continue;
}*/
// for when we bypass the transformation pipeline
BPoint transformOffset(0.0, 0.0);
transform.Transform(&transformOffset);
const agg::glyph_cache* glyph = fFontCache.glyph(charCode);
for (uint32 i = 0; i < glyphCount; i++) {
/* // line break (not supported by R5)
if (*p == '\n') {
y0 += LineOffset();
x = 0.0;
y = y0;
advanceX = 0.0;
advanceY = 0.0;
++p;
continue;
}*/
if (!glyph) {
if (charCode >= 0x20 && charCode < 0x80)
fprintf(stderr, "failed to load glyph for '%c'\n", (char)charCode);
else
fprintf(stderr, "failed to load glyph for 0x%04x\n", charCode);
continue;
}
const agg::glyph_cache* glyph = fFontCache.glyph(*p);
if (!firstLoop && fKerning)
fFontCache.add_kerning(&advanceX, &advanceY);
if (glyph) {
if (i > 0 && fKerning)
fFontCache.add_kerning(&advanceX, &advanceY);
x += advanceX;
y += advanceY;
x += advanceX;
y += advanceY;
if (delta)
x += is_white_space(charCode) ? delta->space : delta->nonspace;
if (delta)
x += is_white_space(*p) ? delta->space : delta->nonspace;
// "glyphBounds" is the bounds of the glyph transformed
// by the x y location of the glyph along the base line,
// it is therefor yet "untransformed".
const agg::rect& r = glyph->bounds;
BRect glyphBounds(r.x1 + x, r.y1 + y - 1, r.x2 + x + 1, r.y2 + y + 1);
// NOTE: "-1"/"+1" converts the glyph bounding box from pixel
// indices to pixel area coordinates
// "glyphBounds" is the bounds of the glyph transformed
// by the x y location of the glyph along the base line,
// it is therefor yet "untransformed".
const agg::rect& r = glyph->bounds;
BRect glyphBounds(r.x1 + x, r.y1 + y - 1, r.x2 + x + 1, r.y2 + y + 1);
// NOTE: "-1"/"+ 1" converts the glyph bounding box from pixel
// indices to pixel area coordinates
// track bounding box
if (glyphBounds.IsValid())
bounds = bounds.IsValid() ? bounds | glyphBounds : glyphBounds;
// track bounding box
if (glyphBounds.IsValid())
bounds = bounds.IsValid() ? bounds | glyphBounds : glyphBounds;
// render the glyph if this is not a dry run
if (!dryRun) {
// init the fontmanager's embedded adaptors
// NOTE: The initialization for the "location" of
// the glyph is different depending on wether we
// deal with non-(rotated/sheared) text, in which
// case we have a native FT bitmap. For rotated or
// sheared text, we use AGG vector outlines and
// a transformation pipeline, which will be applied
// _after_ we retrieve the outline, and that's why
// we simply pass x and y, which are untransformed.
// render the glyph if this is not a dry run
if (!dryRun) {
// init the fontmanager's embedded adaptors
// NOTE: The initialization for the "location" of
// the glyph is different depending on wether we
// deal with non-(rotated/sheared) text, in which
// case we have a native FT bitmap. For rotated or
// sheared text, we use AGG vector outlines and
// a transformation pipeline, which will be applied
// _after_ we retrieve the outline, and that's why
// we simply pass x and y, which are untransformed.
// "glyphBounds" is now transformed into screen coords
// in order to stop drawing when we are already outside
// of the clipping frame
if (glyph->data_type != agg::glyph_data_outline) {
// we cannot use the transformation pipeline
double transformedX = x + transformOffset.x;
double transformedY = y + transformOffset.y;
fFontCache.init_embedded_adaptors(glyph,
transformedX, transformedY);
glyphBounds.OffsetBy(transformOffset);
} else {
fFontCache.init_embedded_adaptors(glyph, x, y);
glyphBounds = transform.TransformBounds(glyphBounds);
}
// "glyphBounds" is now transformed into screen coords
// in order to stop drawing when we are already outside
// of the clipping frame
if (glyph->data_type != agg::glyph_data_outline) {
// we cannot use the transformation pipeline
double transformedX = x + transformOffset.x;
double transformedY = y + transformOffset.y;
fFontCache.init_embedded_adaptors(glyph,
transformedX,
transformedY);
glyphBounds.OffsetBy(transformOffset);
} else {
fFontCache.init_embedded_adaptors(glyph, x, y);
glyphBounds = transform.TransformBounds(glyphBounds);
}
if (clippingFrame.Intersects(glyphBounds)) {
switch (glyph->data_type) {
case agg::glyph_data_mono:
agg::render_scanlines(fFontCache.mono_adaptor(),
fFontCache.mono_scanline(), *binRenderer);
break;
if (clippingFrame.Intersects(glyphBounds)) {
switch (glyph->data_type) {
case agg::glyph_data_mono:
agg::render_scanlines(fFontCache.mono_adaptor(),
fFontCache.mono_scanline(),
*binRenderer);
break;
case agg::glyph_data_gray8:
agg::render_scanlines(fFontCache.gray8_adaptor(),
fFontCache.gray8_scanline(), *solidRenderer);
break;
case agg::glyph_data_gray8:
agg::render_scanlines(fFontCache.gray8_adaptor(),
fFontCache.gray8_scanline(),
*solidRenderer);
break;
case agg::glyph_data_outline: {
fRasterizer.reset();
// NOTE: this function can be easily extended to handle
// conversion to contours, so that's why there is a lot of
// commented out code, I leave it here because I think it
// will be needed again.
case agg::glyph_data_outline: {
fRasterizer.reset();
// NOTE: this function can be easily extended to handle
// conversion to contours, so that's why there is a lot of
// commented out code, I leave it here because I think it
// will be needed again.
//if(fabs(0.0) <= 0.01) {
// For the sake of efficiency skip the
// contour converter if the weight is about zero.
// if(fabs(0.0) <= 0.01) {
// For the sake of efficiency skip the
// contour converter if the weight is about zero.
//-----------------------
fRasterizer.add_path(transformedOutline);
fRasterizer.add_path(transformedOutline);
#if SHOW_GLYPH_BOUNDS
agg::path_storage p;
p.move_to(glyphBounds.left + 0.5, glyphBounds.top + 0.5);
@ -304,51 +314,42 @@ AGGTextRenderer::RenderString(const char* string,
ps.width(1.0);
fRasterizer.add_path(ps);
#endif
/* } else {
// fRasterizer.add_path(fContour);
fRasterizer.add_path(transformedOutline);
}*/
agg::render_scanlines(fRasterizer, fScanline, *solidRenderer);
break;
}
default:
break;
}
}
}
/*} else {
//fRasterizer.add_path(fContour);
fRasterizer.add_path(transformedOutline);
}*/
// increment pen position
advanceX = glyph->advance_x;
advanceY = glyph->advance_y;
} else {
if (*p < 128) {
char c[2];
c[0] = (uint8)*p;
c[1] = 0;
fprintf(stderr, "failed to load glyph for '%s'\n", c);
} else {
fprintf(stderr, "failed to load glyph for %d\n", *p);
agg::render_scanlines(fRasterizer, fScanline, *solidRenderer);
break;
}
default:
break;
}
}
++p;
}
// put pen location behind rendered text
// (at the baseline of the virtual next glyph)
if (nextCharPos) {
nextCharPos->x = x + advanceX;
nextCharPos->y = y + advanceY;
transform.Transform(nextCharPos);
}
// increment pen position
advanceX = glyph->advance_x;
advanceY = glyph->advance_y;
firstLoop = false;
};
// put pen location behind rendered text
// (at the baseline of the virtual next glyph)
if (nextCharPos) {
nextCharPos->x = x + advanceX;
nextCharPos->y = y + advanceY;
transform.Transform(nextCharPos);
}
return transform.TransformBounds(bounds);
}
// StringWidth
double
AGGTextRenderer::StringWidth(const char* utf8String, uint32 length)
AGGTextRenderer::StringWidth(const char* string, uint32 length)
{
// NOTE: The implementation does not take font rotation (or shear)
// into account. Just like on R5. Should it ever be desirable to
@ -361,69 +362,25 @@ AGGTextRenderer::StringWidth(const char* utf8String, uint32 length)
// Note that shear will not have any influence on the baseline though.
double width = 0.0;
uint32 glyphCount;
if (_PrepareUnicodeBuffer(utf8String, length, &glyphCount) >= B_OK) {
uint16* p = (uint16*)fUnicodeBuffer;
uint32 charCode;
double y = 0.0;
const agg::glyph_cache* glyph;
bool firstLoop = true;
double y = 0.0;
const agg::glyph_cache* glyph;
for (uint32 i = 0; i < glyphCount; i++) {
if ((glyph = fFontCache.glyph(*p))) {
if (i > 0 && fKerning)
fFontCache.add_kerning(&width, &y);
width += glyph->advance_x;
}
++p;
while ((charCode = UTF8ToCharCode(&string)) > 0) {
glyph = fFontCache.glyph(charCode);
if (glyph) {
if (!firstLoop && fKerning)
fFontCache.add_kerning(&width, &y);
width += glyph->advance_x;
}
}
firstLoop = false;
};
return width;
}
// _PrepareUnicodeBuffer
status_t
AGGTextRenderer::_PrepareUnicodeBuffer(const char* utf8String,
uint32 length, uint32* glyphCount)
{
if (length == 0) {
*glyphCount = 0;
return B_OK;
}
int32 srcLength = length;
int32 dstLength = srcLength * 4;
// take care of adjusting buffer size
if (dstLength > fUnicodeBufferSize) {
fUnicodeBufferSize = dstLength;
fUnicodeBuffer = (char*)realloc((void*)fUnicodeBuffer, fUnicodeBufferSize);
}
status_t ret;
if (!fUnicodeBuffer) {
ret = B_NO_MEMORY;
} else {
int32 state = 0;
ret = convert_from_utf8(B_UNICODE_CONVERSION,
utf8String, &srcLength,
fUnicodeBuffer, &dstLength,
&state, B_SUBSTITUTE);
}
if (ret >= B_OK) {
*glyphCount = (uint32)(dstLength / 2);
ret = swap_data(B_INT16_TYPE, fUnicodeBuffer, dstLength,
B_SWAP_BENDIAN_TO_HOST);
} else {
*glyphCount = 0;
fprintf(stderr, "AGGTextRenderer::_PrepareUnicodeBuffer() - UTF8 -> Unicode conversion failed: %s\n", strerror(ret));
}
return ret;
}
// _UpdateSizeAndHinting
void
AGGTextRenderer::_UpdateSizeAndHinting(float size, bool hinted, bool force)

View File

@ -48,9 +48,6 @@ class AGGTextRenderer {
uint32 length);
private:
status_t _PrepareUnicodeBuffer(const char* utf8String,
uint32 length,
uint32* glyphCount);
void _UpdateSizeAndHinting(float size, bool hinted,
bool force = false);