Activity for Anti-Grain Geometry Library
-
Marco Aurelio Padilha Cassilha posted a comment on discussion Open Discussion
There is no tar.gz package [r145] to download, this is a big problem, there is no simple way to use svn to download, it makes it difficult to create packages to use in distros. Could you make a tar.gz package available?
-
John Horigan committed [r145] on SVN
Fix some warnings from the Debian build environment:
-
cecilio posted a comment on discussion Open Discussion
When I made my initial comment I hadn't reviewed all the code to see the uses of the tags variable. I just thought that doing an explicit cast on a couple of lines could cause problems if a FreeType method was invoked passing tags as a parameter, since, depending on the version, errors could be generated. I prefer the #if/#endif block solution as it leaves the code cleaner even though I hate this type of macro solutions. Thank you.
-
Update gpc to latest (and final) version, now under MIT license
-
Alternate approach to Freetype 2.13.3 change, use a typedef instead of a cast
-
The compiler complains about implicit casts from char* to unsigned char*, but not explicit casts. This is what set this all in motion. I maintain the Debian package for AGG and Debian is switching to Freetype 2.13.3 from Freetype 2.12.1. I don't know why the Freetype developers think that it is OK to have an API change on a patch release. Explicit casting from unsigned char* to char* or from char* to unsigned char* is legal and does not generate an error. I have tested both directions with both versions...
-
I have not tested but I would expect a compiler complain: error: invalid conversion from ‘unsigned char*’ to ‘char*’ [-fpermissive] or error: invalid conversion from ‘char*’ to ‘unsigned char*’ [-fpermissive] depending on agg/FreeType parameters combination.
-
If agg is linked with a previous version of FreeType there is now a problem. r402 should be reverted and fixed with something as: // Freetype version 2.13.3 and later uses unsigned char for tags #if (FREETYPE_MAJOR > 2) || (FREETYPE_MAJOR == 2 && FREETYPE_MINOR > 13) || (FREETYPE_MAJOR == 2 && FREETYPE_MINOR == 13 && FREETYPE_PATCH >= 3) unsigned char* tags; #else char* tags; #endif
-
I have compiled r142 with Freetype 2.12.1 and Freetype 2.13.3 and it works fine. How does casting to char* cause a problem?
-
If agg is linked with a previous version of FreeType there is now a problem. r402 should be reverted and fixed with something as: ~~~ // Freetype version 2.13.3 and later uses unsigned char for tags if (FREETYPE_MAJOR > 2) || (FREETYPE_MAJOR == 2 && FREETYPE_MINOR > 13) || (FREETYPE_MAJOR == 2 && FREETYPE_MINOR == 13 && FREETYPE_PATCH >= 3) unsigned char* tags; else char* tags; endif
-
FreeType changed a char* to unsigned char*, add appropriate cast
-
Forgot several switch case fallthrough.
-
Fixed a large number of warnings when compiled under C++23 with -Wall
-
Zardshard created ticket #9
Fix member initialization order
-
Why file copying uses modified BSD license, if repository uses MIT license?
-
Adrien Destugues posted a comment on ticket #7
To add to this, the fact that the SVN switched ot a 2.7 release (in r134) but then there never was a formal release of that (the repository has no tags, and there is no released source archive) is also a little confusing. Should we assume r134 is the 2.7.0 release?
-
Questions about the project and the future
-
Jim Barry modified ticket #3
Replace hard-coded math constants with cmath constants
-
Addressed in r136.
-
Successive assignment of variable 'da' in agg_math_stroke.h
-
Patch applied in r136.
-
[HAIKU] PVS 730: initialize m_distance_tolerance_square
-
Fixed in r136.
-
Fix multiple definitions linker error
-
Patch applied (r139). Thanks for submitting!
-
Fixed multiple-definition linker errors (https://sourceforge.net/p/agg/patches/6/).
-
Fix multiple definitions linker error
-
Remove dependence on antigrain.com for the examples
-
Undo previous, incorrect fix.
-
Coconut posted a comment on discussion Open Discussion
The code should be changed to use pod_allocator instead.
-
Some minor fixes requested by users
-
M_SQRT1_2 is from the POSIX standard, not any C or C++ standard. I can update the file with a more precise version of the constant.
-
Fix off-by-one error in mol_view example
-
Ivan Gagis posted a comment on discussion Open Discussion
I'm also maintaining my own fork at https://github.com/cppfw/agg I have done some changes there too (check pull requests).
-
Graeme Geldenhuys modified a comment on discussion Open Discussion
I know of this Git repo which has some commits against the GPL 2.5 version. I think it's just an unofficial fork, but it's something at least. https://github.com/timothytylee/agg/commits/master And this one is listed as 2.6 and has even more commits and more recent changes. https://github.com/ghaerr/agg-2.6/commits/master
-
I know of this Git repo which has some commits against the GPL 2.5 version. I think it's just an unofficial fork, but it's something at least. https://github.com/timothytylee/agg/commits/master
-
nocturn posted a comment on discussion Open Discussion
Does this project accept patches? Or more to the point if I post one in the patches page here, is there someone who could merge it or at least reject it should anything be wrong with?
-
Your understanding is correct. I am not aware of any actively maintained fork of the GPL'd agg-2.5.
-
Sakthi Sakthivelmurugan posted a comment on discussion Open Discussion
Jim, Thanks for the quick response and the link. Just to clarify that I understood correctly -- all the changes made on this repo, for example version 2.7 https://sourceforge.net/p/agg/svn/134/ continue to retain the old permissive license. Do you know if the forked agg-2.5 version matained somewhere else? Thanks, Sakthi
-
The sourceforge project is based on AGG 2.4 and retains the permissive license. Version 2.5 was released with exactly the same code as 2.4 but with a GPL license. No futher work was ever done on version 2.5 to my knowledge. See http://agg.sourceforge.net/antigrain.com/news/index.html
-
Hi, What is the licence on AGG? On sourceforge project page, its listed as MIT license but on the below URl it is listed 2.5 is GPL. http://agg.sourceforge.net/antigrain.com/license/index.html Thanks
-
Update library revision from 2.6.0 to 2.7.0
-
Replicate change from r57 made in pixfmt_alpha_blend_rgba to pixfmt_custom_blend_rgba
-
Norbert Nemec posted a comment on discussion Open Discussion
After having another close look, I must correct myself: Only now I noticed the changed done on the SVN repo here between 2014 and 2018 - seems to be a similar amount of changes to those that happened on github and never made their way back into the main subversion repo. I guess, it would be worth consolidating these changes.
-
The main developer abandoned the project a long time ago, but the quality is still so good that it is worth preserving. Back in 2014, I carefully revised and merged all the free-floating forks of the BSD-licensed version 2.4 into one git repository: https://github.com/NNemec/antigrain At that time, I fixed all compiler warnings and various issues to bring the library to a state where I could successfully use it in a commercial product. To my knowledge, that merge still is the highest-quality fork...
-
Rafał Bernacki created ticket #5
[HAIKU] PVS 730: initialize m_distance_tolerance_square
-
David Hoese posted a comment on discussion Open Discussion
How is the maintainer-ship of this project going? Are people still working on stuff here? Does the project need to be moved to a different platform to get more eyes? I'm mostly concerned by the antigrain.com website which seems to have been squatted and not working.
-
Rob Gill created ticket #4
Successive assignment of variable 'da' in agg_math_stroke.h
-
Replace hard-coded math constants with cmath constants
-
Update version to 2.6.0. Use modern macros for configure.ac
-
Replace all C standard headers with C++ equivalent headers, use std:: namespace everywhere (except examples)
-
Always check the return value of fread()
-
Bug fixes and enhancements requested by matplotlib developers
-
I am just wondering if you guys considered to move the project to github? The last week I could not access agg on sourceforge at all since sourceforge was basically down the whole week. And the patch process is a lot simpler on github (i.e. pull request).
-
Thanks a lot. I will try it out.
-
Jim Barry posted a comment on discussion Open Discussion
If you want to do compositing, the (destination) layers need to be pixfmt_rgba32_pre. You can find a brief explanation in the version 2.2 release notes. By the way, rgba(1, 1, 1, 0) is not valid in premultiplied space, since each rgb component should be <= alpha.
-
Hi, I have a piece of code that was originally written as directly draw lines, fonts, polygons with anti-aliasing on an image. But later on, we decided to use layer approach which is the following create a layer, and draw that layer. we can have multiple layers... and then compose all the layers onto a white background image to get a final image. The problem is, because of the alpha handling (we clear each layer to rgba(255,255,255,0)), it effectively blends the transparency twice, resulting incorrect...
-
I quite agree. By convention, preprocessor symbols should be in uppercase, and introducing such a symbol to represent a very large number seems especially inadvisable when one could just use FLT_MAX (or std::numeric_limits<float>::max). The file gradients_contours.cpp contains more lowercase preprocessor symbols, which I have either renamed to uppercase or replaced with const variables. SVN revision 128. Thanks for reporting the issue.
-
Removed inappropriate use of lowercase preprocessor symbols.
-
When building gradients_ountour.cpp using cmake, I encountered the following error message In file included from /c/dr/svg/agg-2.4/examples/gradients_contour.cpp:25:0: /c/dr/svg/agg-2.4/include/agg_span_gradient_contour.h:37:18: error: expected unqualified-id before numeric constant #define infinity 1E20 My build environment is cygwin64 with $ gcc --version gcc (GCC) 6.4.0 Copyright (C) 2017 Free Software Foundation, Inc. This is free software; see the source for copying conditions. There is NO warranty;...
-
Thomas A Caswell created ticket #2
Bug fixes from matplotlib's vendored copy of Agg
-
sujitha posted a comment on discussion Open Discussion
Hi, I am new to AGG.How to render GIS(Geographic Information System) vector data or GIS shape file using AGG? What are the APi's I need to call to achieve this. Thanks in advance.
-
Improve svg_test SVG viewer
-
Add path_storage::align_path() for fixing slightly misaligned closed paths
-
Renaming configure.in to configure.ac
-
Add missing files to agg-2.4/*/Makefile.am, convert files from DOS line endings to Unix
-
Suppress spurious uninitialized variable warnings from gcc 5.4
-
Remove assumption that ints are 32-bit w.r.t. min/max int
-
Changed some signed arithmetic to unsigned to eliminate possible undefined behavior
-
Removed unused compositing mode parameter for the four compositing adaptors that do not use it.
-
Fix crash in libagg when very, very long lines are drawn
-
Remove a superfluous method that causes build failures on Debian systems using clang
-
SourceForge project name is now "agg".
-
Adding original content from antigrain.com.
-
-
-
Deleting unneeded files.
-
Fixed bug exposed by g++ warning "unused parame...
-
Fixed g++ warning "comparison is always false d...
-
Fixed g++ warning "enumeral and non-enumeral ty...
-
Fixed g++ warning "typedef ‘src_color_type’ loc...
-
Fixed warnings about unused parameters when com...
-
Fixed error in scale adjustment, due to use of ...
-
On 24 April 2015 at 15:53, Graeme Geldenhuys graemeg@users.sf.net wrote: I've just...
-
I've just updated AGG to the latest r106 and compiled in under FreeBSD 10.1. I then...
-
Anti-Grain Geometry Library released /OldFiles/msys-sdk-1.0.1-1.tar.gz
-
Fixed error in "opacity" function (patch from M...
-
Fixed some build errors/warnings in Clang (patc...
-
Switched implementation of sRGB conversion from...
-
Hi Julien, It's a bit of a mess, but for historical reasons the mailing list is at...
-
Julien B aka Artemys posted a comment on discussion Open Discussion
Hello, I'm creating a Buildroot package for your project and I would like to know...
-
Fixed breakage from backing out from rev 99 cha...
-
Backing out from rev 99 changes.
-
Added "explicit" specifier to constructors that...
-
Avoid potential compiler warning about converti...
-
Removed "Step" and "Offset" template parameters...
-
Fixed a typo and compiler warnings about conver...
-
Added missing "plain" pixel formats.
-
Added missing conversions for gray16->rgba32 an...
