Commit Graph

659 Commits

Author SHA1 Message Date
9657c04eed extent: aspect should return float 2014-12-19 18:29:12 +11:00
7d5eac3f2f io: move mapped_file to platform subfiles 2014-12-19 18:28:50 +11:00
bce2f81d2f point: add trivial std::hash specialisation 2014-12-19 15:00:33 +11:00
a248d14d54 point: instance ostream for size_t templates 2014-12-19 15:00:04 +11:00
58aeeb1786 point: add equality operator 2014-12-19 14:59:50 +11:00
1bb4c752f1 point: templatise distance functions 2014-12-19 14:59:16 +11:00
67b03fd10f nocopy: explicitly delete move construct/assign 2014-12-19 14:58:13 +11:00
d4a6c115bc point: add point2u typedef for point<size_t,2> 2014-12-18 19:43:13 +11:00
c1c74a14f0 io: always enable mapped_file 2014-12-18 12:57:04 +11:00
d81149b6e1 vector: remove explicit qualifier for init_list 2014-12-17 14:43:03 +11:00
3b1226ee9c vector: add 4d and 4f convenience typedefs 2014-12-16 15:30:25 +11:00
f3733b1cbc types: remove internal make_unique implementation
we can use the c++1y implementation instead which is a lot more robust.
2014-12-16 15:29:49 +11:00
7c8562407f rand: remove narrowing cast warning in test 2014-12-16 15:29:26 +11:00
557842441f matrix: remove unneeded cast warning in test 2014-12-16 15:28:59 +11:00
ede1203ff7 matrix: add affine scale implementation 2014-12-16 15:28:40 +11:00
af3adb9ff1 coord: add more named element accessors 2014-12-16 15:28:02 +11:00
49a467aae8 build: remove unneeded use of RANLIB 2014-12-16 15:27:32 +11:00
c28545fe20 build: enable c++1y support 2014-12-16 15:27:12 +11:00
6f23e39459 colour: use floats for values
double is way too much precision for almost every application. we can
always add more templates or other insanity later.
2014-12-16 15:12:44 +11:00
5956d0421e parameterise point, vector, region, matrix on type 2014-12-15 20:10:56 +11:00
2fc9073901 matrix: add affine definitions and translate impl 2014-12-15 13:45:50 +11:00
d51caee0c7 matrix: add perspective matrices 2014-12-15 13:45:50 +11:00
dcdbbe938d matrix: add scalar multiply 2014-12-15 13:45:50 +11:00
82f318c8ad matrix: more readable temps for matrix mul 2014-12-15 13:43:22 +11:00
253be704cb matrix: add transpose functions 2014-12-15 13:42:44 +11:00
a198aa1cba matrix: add trivial matrix-mul test 2014-12-15 13:32:17 +11:00
6a6933ea70 build: update test to use only static libraries 2014-12-15 13:32:17 +11:00
c3cd865c16 coord: add generic dot function 2014-12-15 13:30:02 +11:00
29c2457549 coord: add index accessor operators 2014-12-15 13:29:57 +11:00
5ebb777200 build: tune code generation options 2014-12-10 13:58:43 +11:00
6af493ba46 build: disable LTO for MinGW
Avoids a probably compiler error whereby LTO static libraries cannot be
linked to applications.
2014-12-10 13:53:13 +11:00
de353c9297 build: switch from LT libraries to pure static
We only ever supported static libraries, and this greatly simplifies
argument management compiling/linking for MinGW
2014-12-10 13:51:54 +11:00
b56426ff16 stringid: use boost threads for windows support 2014-12-05 13:22:48 +11:00
976a4b7330 socket: explain and relax type casting requirement
The recv/send functions (and the assertions) were only tested under
32bit systems. They need relaxing for 64bit.
2014-12-05 13:21:40 +11:00
b8ba95f365 io: use appropriate string types for windows 2014-12-05 13:20:44 +11:00
331ea58ec4 fixed: add accessor for raw int values 2014-12-05 13:20:23 +11:00
c2d608410b fixed: move to util namespace 2014-12-05 13:20:05 +11:00
a340473502 debug: print test message when installing console 2014-12-05 13:19:01 +11:00
189d1f7363 build: dont redundant SUBST pkg-config variables 2014-12-05 13:18:16 +11:00
82ae659fb5 build: remove allow-multiple-definition flag
Silently ignoring multiple definitions hides linking issues under
Windows in particular (eg, linking to multiple runtimes).

We don't seem to be hitting the boost bug which triggered this anymore.
2014-12-05 13:14:48 +11:00
0d32b07b2d build: delete references to old COMMON variables
COMMON_CXXFLAGS and COMMOND_LDFLAGS were the old flag accumulators
before we used AX_APPEND. Obviously not needed anymore.
2014-12-05 13:12:18 +11:00
733f16d1c8 build: add json test data to dist 2014-10-26 18:55:24 +11:00
cf6fc87c9f options: use range based for in some tests 2014-10-21 21:48:13 +11:00
31326ef576 options: instantiate valueoptions privately
By not providing each translation unit with the general form code we can
more easily force people to use the specialised code for types like
bool.
2014-10-21 21:47:10 +11:00
7278eff60a build: explicitly enable leak checking 2014-10-21 19:38:40 +11:00
9ffaa5782b rational: add invert, inverse, div, and mul 2014-10-20 12:56:28 +11:00
3fd3b805af adler: fix c-style cast without const error 2014-10-20 12:55:59 +11:00
23b0a7b146 remove annotations header
The annotations were never particularly robust, and weren't used in many
locations (or consistently). And we should be migrating to c++14
annotation syntax anyway. By focusing on writing them all locally we
will be a little more standards compliant, and the attributes should be
more visible.
2014-10-17 19:24:53 +11:00
abfe4daac3 io: default mapped_file to read only
This is vastly safer compared to write by default. No more unexpectedly
changing files.
2014-09-29 16:19:16 +10:00
84ca008305 copyright headers and date bumps 2014-09-17 18:20:28 +10:00