|
4388dc7a
|
2008-06-11T23:23:05
|
|
provide for us poor sobs who don't have a native strndup implementation
|
|
c5bcad81
|
2008-06-09T12:57:41
|
|
* Add my copyright information to files I modified significantly.
* Add the Unicode, Inc. to the list of authors because of FTUnicode.h.
|
|
|
4a3dd0f5
|
2008-06-09T11:47:42
|
|
* Code simplifications in FTContour.
|
|
|
291c70ed
|
2008-06-09T11:47:14
|
|
* Add "const" qualifier to FTPoint's scalar multiplication operator.
|
|
|
a132a5e4
|
2008-06-09T11:45:52
|
|
* Document and reorganise FTBuffer. It is now clean enough.
|
|
|
d7c9fc14
|
2008-06-09T10:13:09
|
|
* Document missing function parameters.
|
|
|
0b40c371
|
2008-06-09T10:12:21
|
|
* Better FTContour::ComputeOutsetPoint() documentation.
|
|
|
14e3e435
|
2008-06-09T10:11:58
|
|
* Create FTFont::GlyphLoadFlags() to vary FT_Load_Glyph() calls according
to the font type. Now we no longer load vector information when not
needed.
|
|
|
6c2a187c
|
2008-06-08T15:56:35
|
|
* Revert the FTFont::Advance() and FTGlyph::Advance() improvements. After
discussion, I was convinced they were not worth the backwards compatibility
breakage. They now return float again, instead of FTPoint.
|
|
|
632c866d
|
2008-06-08T15:55:19
|
|
* Add a comment to FTBufferFont to not forget about bugs in it.
|
|
|
b173cbee
|
2008-06-03T12:24:37
|
|
* Fix a crash in FTExtrudeGlyph caused by uninitialised members.
|
|
|
7e2d6252
|
2008-06-03T12:23:25
|
|
* FTBufferFont: use strndup() and, when available, wcsdup().
|
|
|
cdc9b560
|
2008-06-03T12:23:07
|
|
* Coding style: remove tabs and trailing spaces, wrap long lines.
|
|
|
dd224b27
|
2008-05-24T19:04:02
|
|
from BZFlag r17264: rewrite StringCopy. wcscpy is not a standard function; we have logic for doing it ourselves already (minus the len = 0 case), so handle len = 0 and do it all ourselves.
|
|
|
6e4ad4cf
|
2008-05-23T16:45:20
|
|
Copyright notices as requested by sam
|
|
|
e1e19b26
|
2008-05-23T00:56:25
|
|
include wctype.h to make gcc happy
|
|
|
484a90e1
|
2008-05-23T00:53:23
|
|
no need to shadow variables
|
|
|
84869ec7
|
2008-05-23T00:46:08
|
|
* Provide a helper class for walking potentially-multibyte unicode strings.
* Provide support for multibyte encodings (UTF-8, UTF-16) in FTFont and derived classes, and in FTSimpleLayout.
* Put a few UTF-8 strings in non-latin codeplanes in FTGLDemo (toggle at compile-time) for testing.
* FTSimpleLayout should be tested extensively before release. I would be surprised if I didn't break at least one unusual use case.
|
|
|
95183df2
|
2008-05-23T00:16:54
|
|
* Start protecting FTBuffer members using getters and setters.
|
|
|
5de43fe6
|
2008-05-22T15:31:32
|
|
* Empty the FTBufferFont cache when changing the face size.
|
|
|
d1bea035
|
2008-05-22T15:31:07
|
|
* Fix a visual bug in FTBufferFont caused by overlapping glyphs.
|
|
|
b73f2e81
|
2008-05-22T12:39:48
|
|
* Implement caching in FTBufferFont. To avoid unnecessary texture uploads,
each font object keeps the last 16 strings in the graphic card.
|
|
|
865b9043
|
2008-05-22T12:39:12
|
|
* Fix an off-by-one error in FTBufferGlyph::Render.
|
|
|
3d018d80
|
2008-05-21T16:37:54
|
|
* Fix the quad vertex order in FTBufferFont.
|
|
|
0bc4a552
|
2008-05-21T16:36:34
|
|
* Fix a bug in the FTFont::BBox calculation: the first glyph bbox was not
relative to the position argument.
|
|
|
901ee76c
|
2008-05-21T15:45:11
|
|
* Fix a texture coordinate bug caused by our next-power-of-two ceiling.
|
|
|
fa5631df
|
2008-05-21T15:43:52
|
|
* Add proper glPushAttrib() and glPushClientAttrib() calls to the
rendering methods that need them.
|
|
|
68a99036
|
2008-05-21T11:39:31
|
|
* Enable GL_TEXTURE_2D in FTFont::Render variations that need it.
|
|
|
f2011b55
|
2008-05-21T00:39:04
|
|
quell warning
|
|
|
ca7c7a81
|
2008-05-20T23:49:25
|
|
* First try at the FTBufferFont/FTBufferGlyph implementation.
|
|
|
8932eba0
|
2008-05-20T23:48:56
|
|
* Rename FTTextureGlyphImpl::pos to FTTextureGlyphImpl::corner because pos
was misleading.
|
|
|
f7d00178
|
2008-05-19T15:45:47
|
|
* Starting the buffer font class revival. For now, it just consists in empty
FTBuffer, FTBufferGlyph and FTBufferFont classes.
|
|
|
7abb7502
|
2008-05-19T10:06:49
|
|
* Do not crash if the face has no charmaps. Bug found using zzuf:
xvfb-run zzuf -q -F5 -r0.0000001:0.1 -s0:10000 -c CTest EunjinNakseo.ttf
|
|
|
9b33da91
|
2008-05-19T10:06:25
|
|
* Simplified contour parity check routine, formula courtesy of Guillaume
Bittoun.
|
|
|
13adc2c1
|
2008-05-19T10:06:07
|
|
* Add parentheses around && within || to please gcc.
|
|
|
612c3756
|
2008-05-11T21:43:57
|
|
* Simplify the C bindings. 70 lines gained.
|
|
|
b70df3d7
|
2008-05-11T21:43:35
|
|
* Get rid of FTSimpleLayout::RenderSpace(). It's still useful in
FTSimpleLayoutImpl, but as a public method FTFont::Render() is just
as powerful.
|
|
|
55dc91e0
|
2008-05-11T21:43:10
|
|
* Added optional position and string length to the Layout methods.
|
|
|
a90225b5
|
2008-05-11T11:29:39
|
|
* Get rid of all methods in FTLayoutImpl that were accessing FTFontImpl
internals, since FTFont now has all the proper public methods for that.
|
|
|
32ef6d1e
|
2008-05-11T11:29:17
|
|
* Get rid of FTFontImpl::DoRender(), one of the oldest TODOs.
|
|
|
47e20607
|
2008-05-11T11:28:53
|
|
* Re-add FTFont::BBox() implementations with the old prototype, in case
old projects use them.
|
|
|
5f23d804
|
2008-05-11T11:28:38
|
|
* Honour the pen's Z coordinate when rendering glyphs (except the raster
ones, where the Z coordinate makes no sense).
|
|
|
20ac99b1
|
2008-05-11T11:28:08
|
|
* FTFont::Advance(), FTFont::Render() and FTFont::BBox() are now far
more powerful, allowing for substring display and extra spacing between
characters.
|
|
|
d39c7838
|
2008-05-11T11:26:12
|
|
* Kill 180 lines of code by removing duplicate *::Render() functions and
giving a default value to the renderMode parameter.
|
|
|
68738f93
|
2008-05-09T10:02:46
|
|
* Put a few wrapper functions into extern "C++" braces because they return
references to C++ objects.
|
|
|
17f84259
|
2008-05-09T10:01:49
|
|
* Minor cosmetic fix (remove tab).
|
|
|
65ba4d60
|
2008-05-08T23:12:47
|
|
and more fun with float/double consistency
|
|
|
d33c1e50
|
2008-05-08T23:12:29
|
|
more fun with size_t consistency
|
|
|
e9260bbf
|
2008-05-08T22:35:57
|
|
* Implement C bindings for FTGlyph and FTFont subclassing.
* Add subclassing to the simple C demo to show how to do similar stuff as
in the C++ demo.
|
|
|
d5adb85b
|
2008-05-08T17:07:07
|
|
* When a glyph is created, check that all its contours have the proper
clockwise/counterclockwise orientation. This fixes a nasty display bug
with some badly encoded fonts.
|
|
|
dccd536a
|
2008-05-08T17:06:24
|
|
* During contour creation, compute whether the contour is clockwise or
anti-clockwise. This will be needed later to correct fonts that do not
abide to the even-odd and non-zero winding number conventions, thus
breaking our nice outset glyphs.
|
|
|
2ecbe452
|
2008-05-08T17:05:26
|
|
* Add scalar product to the FTPoint operators.
|
|
|
f2348fe5
|
2008-05-08T17:04:07
|
|
* Cosmetic fixes in the FTLayout C bindings.
|
|
|
55bf1ff0
|
2008-05-08T17:03:07
|
|
* Small optimisation in FTContour::ComputeOutsetPoint().
|
|
|
92b95cde
|
2008-05-08T17:01:45
|
|
* Made most FTFont, FTGlyph and FTLayout function virtual. It's true that
most of them use private members of the pImpl class and thus are not
easily replaced, but intercepting the information may be useful in
subclassing, too.
|
|
|
45c2269f
|
2008-05-07T16:10:28
|
|
* Some code cleanup here and there, mostly in FTGlyphContainer.
|
|
|
1deacbc3
|
2008-05-07T15:07:38
|
|
* Documentation updates.
|
|
|
48b68f55
|
2008-05-07T15:06:19
|
|
* Replace FTBBox::Move() with the += operator, to make it clearer that the
object is modified in the process.
|
|
|
8863eeeb
|
2008-05-07T15:03:48
|
|
* Change the += operator for bounding boxes to |=, which better represents
what is happening, and avoids future confusion with "FTBBox + FTPoint"
constructs.
|
|
|
4f44ec81
|
2008-05-07T15:01:47
|
|
* Make all BBox functions return an FTBBox object instead of doing
countless conversions to floats or arrays of floats.
|
|
|
f58ddca2
|
2008-05-07T14:59:03
|
|
* Fix FTBBox::SetDepth() behaviour with negative depth values (as done in
the FTExtrudeGlyph class).
|
|
|
28cc198b
|
2008-05-07T14:58:09
|
|
* Reimplement all FTFont::BBox() variants using the same FTFontImpl::BBox()
common method.
|
|
|
3823a30a
|
2008-05-07T14:56:57
|
|
* Add an overload of FTFont::BBox that returns an FTBBox object. It will
save us a lot of code later.
|
|
|
10cdd9a4
|
2008-05-07T14:55:41
|
|
* Allow to create an FTBBox using two FTPoint objects.
* Allow the operands to FTPoint's "+" and "-" operators to be const.
|
|
|
a065f004
|
2008-05-06T21:19:42
|
|
* Refactor FTContour::ComputeOutsetPoint so that it's twice as short, even
with the additional comments.
|
|
|
134711d7
|
2008-05-06T21:19:29
|
|
* Make FTContour:FTContour comply with the FreeType specification and
remove duplicate points in the Bézier curves. Fixes weird rendering
errors with some fonts.
|
|
|
4ed52c55
|
2008-05-06T10:00:48
|
|
* Fix an unsigned int / size_t mismatch in FTVectoriser.
|
|
|
15012d0f
|
2008-05-06T07:08:55
|
|
* Fix indentation by replacing a few tabs with spaces.
|
|
|
948e3629
|
2008-05-06T06:54:51
|
|
l != 1, depending on your font of course
|
|
|
eb45c699
|
2008-05-06T06:38:37
|
|
VC build fixes from bzflag revs 17848-17852.
* size_t consistency
* avoid coercing from int to bool
* make casts from double to float explicit rather than implicit, mostly by way of a few new getter functions in FTPoint, or avoid if possible.
|
|
|
f7b79437
|
2008-05-05T14:55:11
|
|
* Refactor FTGlyph, FTFont and FTLayout so that client applications can
hopefully subclass them.
|
|
|
68bcaf26
|
2008-05-05T14:52:48
|
|
* Rename FTTextureFontImpl::MakeGlyph to FTTextureFontImpl::MakeGlyphImpl to
avoid confusion.
|
|
|
7d3852f9
|
2008-05-05T14:52:17
|
|
* Rename FTFontImpl::base to FTFontImpl::intf.
|
|
4a459132
|
2008-05-05T13:22:30
|
|
quell warnings, reorder initializations
|
|
|
0a28e5d3
|
2008-05-04T19:39:34
|
|
* Put MakeGlyph back into FT*Font classes instead of FT*FontImpl, and make
it use as few FT*FontImpl members as possible so that external application
may actually have a chance to properly subclass us.
|
|
|
316a7285
|
2008-05-04T19:38:55
|
|
* The FTFont<->FTFontImpl bridge is now complete.
|
|
|
89585b8a
|
2008-05-04T19:38:41
|
|
* Put my name and Sean's in the public headers so that people know who to
contact.
|
|
|
54c815bb
|
2008-05-04T19:38:07
|
|
* Cast strings to unsigned char * before handling them to our internal
methods, because the chars may be cast directly to int, causing crashes
with 8-bit strings.
|
|
|
02817b9c
|
2008-05-04T06:23:04
|
|
quell compilation warnings about the last enum having a comma
|
|
|
b7c7e131
|
2008-05-04T02:58:40
|
|
refer to local/private headers with local path inclusion, otherwise the search include paths are wrong
|
|
|
1b658504
|
2008-05-04T02:57:13
|
|
list the libs as libs instead of flags so they get passed through as dependencies in the libtool archive
|
|
|
552366af
|
2008-05-02T14:52:54
|
|
* Add a virtual destructor to FTLayoutImpl to make sure derived classes
have their destructors called.
* Make all FTLayoutImpl members protected. Only its derived classes and
FTLayout need to access them.
|
|
|
e0e8e479
|
2008-05-02T13:27:11
|
|
* Add an implicit cast to FTGL::TextAlignment in the FTLayout::SetAlignment
C wrapper.
|
|
|
9b3a3a8a
|
2008-05-02T13:18:09
|
|
* Started documenting the FTLayout C bindings.
|
|
|
c33efa0d
|
2008-05-02T12:45:21
|
|
* Use a smaller logo on the User Guide's front page.
|
|
|
7f2bb3f1
|
2008-05-02T12:43:57
|
|
* Document the C bindings for FTGlyph.
|
|
|
d30f9095
|
2008-05-02T12:43:22
|
|
* Fix a potential memory leak in the FTGlyph C bindings error handler.
|
|
|
92238082
|
2008-05-02T09:15:38
|
|
* Change a few function names in the FTFont C bindings to avoid confusion
with other classes.
* Document the C bindings. Since it's almost copypasta from the C++
documentation, I put the constructors back in their original files.
|
|
|
d9cdef45
|
2008-05-02T09:13:14
|
|
* Add RENDER_ALL to the enum used in Font::Render() so that client
applications need not worry about future extensions.
|
|
|
66e53f95
|
2008-05-02T07:21:57
|
|
* Fix a memory leak in ~FTExtrudeGlyph: only one of the three display lists
was being freed.
|
|
|
80e809f1
|
2008-05-02T07:21:09
|
|
* Fix brown-paper-bag bug in the vector product computation: the indices
were completely messed up. Thanks to valgrind for spotting it for me.
|
|
|
dfa07fc3
|
2008-05-02T07:20:51
|
|
* Fix a small memory leak in the C bindings destructors.
|
|
|
e8c876bc
|
2008-05-01T19:31:06
|
|
* Make our base classes' destructors public. We want to prevent accidental
instantiation, but direct destruction is perfectly legal.
|
|
|
54e3e0a6
|
2008-05-01T18:46:22
|
|
* Proper FTFace member initialisation to prevent destructor-time crashes.
|
|
|
fbcf2a88
|
2008-05-01T18:33:44
|
|
* Optimise FTFace::KernAdvance() so that kerning for font indices < 128 is
precomputed during FTFace() instantiation to avoid calling FT_Get_Kerning()
too often. Patch by Sean Morrison, taken from bzflag commit r14652,
reworked for safety and performance by me.
|
|
|
b9df86a4
|
2008-05-01T18:31:35
|
|
* Optimize FTCharmap::FontIndex() so that font indices < 128 are precomputed
during FTCharmap() instantiation to avoid repeated tt_cmap4_char_index()
calls. Performance patch by Sean Morrison, taken from bzflag commit r14644.
|
|
|
822d0f8f
|
2008-05-01T18:30:29
|
|
* Avoid crashing when the texture size is so small that its integer size
becomes zero. At the same time, round many floats instead of simply
flooring them to int. First part by Sean Morrison from bzflag commit
r14590.
|
|
|
40644e3b
|
2008-05-01T14:19:34
|
|
* Patches by Sean Morrison, from BzFlag commits r15755 and r14843:
+ Only delete textures if there is at least one of them.
+ Ensure that the FTTextureFont data members get set for all cases.
* Added Sean to the AUTHORS file.
|
|
|
129a4733
|
2008-05-01T13:42:49
|
|
* Store FTPoints in the BBox object instead of floats. This was a todo
from Henry.
|
|
|
f0705dc2
|
2008-05-01T10:54:28
|
|
* Inline most FTPoint methods and operators. This will probably make the
code smaller instead of bigger, because most of what they do will be
optimised out by the compiler.
* Get rid of the weird GetNormal() method and reimplement it using the ^
operator (vector product) and Normalise() method instead.
|