If next character was not part of valid token and not whitespace we
would end up spinning in the loop indefinitely. Fix by using do..while
loop which always skips at least one character.
Fixes testcase hangs/73b42cd3b6d05e2ddb5c0fe5888459bc
When reading the STL header for a "COLOR=rgb" part, the bytes were treated as signed chars, when in fact they can range from 0-255. This meant that any value greater than 127 would overflow, leading to an incorrect colour.
This change fixes the issue by treating the header as unsigned chars.
For consistency, use a light gray default diffuse color for STL files.
This provides consistency with other geometric types, like PLY, that use a
light-gray default material.
The text "solid" is the general STL header indicator that the file is in fact
an ascii STL archive. Allows spaces to precede the "solid" keyword, providing
compatibility with exporters that insert such (non-compliant )spaces.
Previously, the STL loader relied on a string check at the start of the
buffer to determine if the STL file was of ASCII representation. This led to
certain false-positives for binary STL files with the magic string present in
their header. This patch addresses the issue by adding a proper binary check
to the representation detection.
- pull in IOhannes' patch to set the gcc default visibility for all symbols to NO and to mark ASSIMP_API with __attribute__ ((visibility("default"))).
- drop unneeded ASSIMP_API from most internal classes in /code, we just need to keep some exports on Windows to keep AssimpView alive.
git-svn-id: https://assimp.svn.sourceforge.net/svnroot/assimp/trunk@1066 67173fc5-114c-0410-ac8e-9d2fd5bffc1f
- FIX build error on vc8/release-noboost x64, ConvertUTF.c had wrong PCH settings.
- ADD knowext and listext verbs to assimp_cmd, add some raw docs.
- Update unit tests to reflect these changes. Currently I keep getting failures in some tests, this needs to be resolved *urgently*.
git-svn-id: https://assimp.svn.sourceforge.net/svnroot/assimp/trunk@567 67173fc5-114c-0410-ac8e-9d2fd5bffc1f
IrrXML receives memmapped UTF-8 input data now, it's own (faulty) conversion is not used anymore.
aiString's are explicitly UTF-8 now.
Slight refactorings and improvements.
Adding UTF-8/UTF-16 text files for ASE,obj,collada,ac3d. These contain various japanese/chinese character sequences.
Changing assimp_view's node view to display UTF-8 multibyte sequences correctly.
git-svn-id: https://assimp.svn.sourceforge.net/svnroot/assimp/trunk@469 67173fc5-114c-0410-ac8e-9d2fd5bffc1f
