Moved the GeometryCompressor's free_list_maximum into a member variable of the class.
(+UniqueNumber() commented out since it is not being used anywhere in all of OSRM)
There was absolutely no need for free_list_maximum to be a static variable and it caused major issues in my code.
Short description:
- I ran osrm-prepare's Prepare().Process(..) function in my own code multiple times for multiple files
=> After the first file is done, free_list_maximum is still on the same size, but a new GeometryCompressor will start at a size of 100 again.
=> Indices broken.
f074d94 Use shorter path names in Doxygen doc.
5117d5a Make Doxygen-generated HTML docs work better on small screens.
41607ac Appveyor config: Use lowercase command names everywhere. Filter 7z output.
9e77a39 Remove debug output.
3416c4b Revert "Comment out test to see if there are dependencies between tests."
fa49dde Comment out test to see if there are dependencies between tests.
5add75d Fix typo in appveyor conf: nul.
53b3778 Use different file names for tests.
1091c0b Fix Windows munmap() to return the same as the posix one.
14fdc4e Revert "CMake: Remove unneeded build type setting."
0270087 Workaround for MSVC.
56c2120 Another try...
8e96a6f Trying to fix test on Windows...
11a64c2 CMake workaround: Set CMAKE_CONFIGURATION_TYPES before project(). As per http://www.cmake.org/pipermail/cmake/2012-January/048856.html
3847e7a CMake: Remove unneeded build type setting.
a1bcaf1 CMake: Make CMAKE_CONFIGURATION_TYPES a cache variable.
ddab3e3 Create test file in test_typed_mmap in current dir instead of tmp.
ce1d3bd Build "RelWithDebInfo" on appveyor instead of "Release".
b6db34f Add README showing source of FindGem.cmake.
2c86c55 Better error message and some comments in CMake for sparsetable check.
b16a5a3 Fix ctest on Windows.
2203bc8 Fix FindGem.cmake.
5d1f81d Another try to fix travis build.
ddd3f5e Hopefully fix problem where travis doesn't find osm-testdata.
0e673ea More debug output to find travis problem.
eb01107 Updated data-tests README.
8a971c8 Remove special case of the multipolygon test in travis/appveyor config.
dce792c CMake: Check dependencies of multipolygon test and run as CMake script.
b967677 Add tests for thread pool.
400e9b3 Bugfix: Handle exception in one pool threads properly.
a1ba489 More detailed error reporting from zlib.
95cf621 Disable SparseMemTable if sparsetable size_type is smaller than 8.
7b601b5 Add missing overload to cast_with_assert() function.
60a7d86 Giving up on trying to remove LNK4221 warning.
35ed5df Try another way to get rid of MSV warning...
f9c7d92 Disable a warning about changed behaviour on MSVC.
c3c6b2d CTest: Only output failed tests.
9c99996 Try setting option in a different way for getting rid of LNK4221 on MSVC.
4ac563c Another ssize_t fix for Windows.
11db84f Fix mmap for windows. Convert macro to function.
95d8f75 Fix test on Windows (which doesn't have ssize_t).
df51aa4 Do not write huge files in one system call.
9c4f772 Change mmap() implementation for Windows.
5be817c Rewrote function to work on Windows without warnings.
ea84f73 Fix warnings on Windows.
572d692 Fixed the static_cast_with_assert function and added tests.
89ef86b Use SparseMemArray instead of SparseMemTable in examples.
git-subtree-dir: third_party/libosmium
git-subtree-split: f074d949a5585a81578d682035f2163de971beb3
3b02ca0 add which() method returning zero based index of stored T in Types... for boost::variant() compatibility
c117592 update unit test to match c64c74775a80474f2012c1a49ab2865e3666107a
36f1e12 add get<T>() overloads for when T is stored in recursive_wrapper<T> also makes get<T>() a compile time error where T is not in Types... (ref #24)
7dfdfa2 clean up coverage files in test directory too [skip ci]
89f8a41 add coverage report
git-subtree-dir: third_party/variant
git-subtree-split: 3b02ca0e3ab1a36dd6ec9138e7f93eb3176ae5f7
- refactor facade::GetEscapeName() into get_name_for_id() call that is implemented in subclasses
- remove dead code
- fix failing tests where names got double-escaped
- fixes https://github.com/Project-OSRM/node-osrm/issues/83
Previously, the request parser was created on demand...but this could
cause requests to be flagged as bad requests if the string length was
large enough to cause the string to be split up into chunks by boost's
read_some method.
