Updates work only when the game data is at the same location as the binary.
If this is not the case then the game was probably installed differently,
for example via the distribution or with make install. In this case we
cannot do automatic updates but we also want to use a different system path.
This fixes the linux development snapshots and release tarballs. They were
broken in the sense that they didn't find their game data.
The usage of timsort instead of std::sort at this point is twofold. First,
it's faster in our case where the array is already sorted in many cases
(remember this is called at least once a frame). And it's not just a bit
faster either but a lot. I have measured a factor of 7 on my system.
Second, in our Windows autobuilds there is a crash within std::sort which is
very hard to debug because it's hardly reproducible with anything other than
the autobuilds (I tried hard). If the crash goes away with timsort then
great, if not then maybe it's easier to debug since the code is in our tree.
With this change, MSVC will build binaries in ${CMAKE_CURRENT_BINARY_DIR}
without adding any more subdirectories. It will also expect its data in a
directory called "planet" immediately below the binary directory.
Since MSVC allows building multiple configurations from the same input file,
the resulting binaries will be suffixed by the configuration type. An exception
is RelWithDebInfo, which will have no suffix; this was chosen over plain Release
to aid in debugging.
Building OpenClonk will work out of the box for in-source builds, but
out-of-source builds will have to create a symlink or a directory junction.
We consider this an acceptable drawback; it was proposed that if you use the
non-default option of an out-of-tree build, you will also know how to create a
link or a junction to, or copy the planet directory.
This changeset also revives looking for game data in the same directory as the
binary, which was part of c3fc1ee1ec8c [Peter Wortmann].
This time, the relocation code checks for a "System.c4g" in either
the executable path or a "data" folder directly below. CMake makes
sure that this points somewhere sensible for normal builds.
TODO:
* Check whether this actually works under Unixes. Can "ln -sf"
delete important stuff? Is there a safe alternative?
* Further unify with the Mac Os solution. Other platforms might
auto-pack for release builds too, for example - and it might
be a good idea to have a proper data subdirectory in Mac bundles
as well.
This fixes make install, which previously tried to install nonexistant
packed groups from the source directory. Make it use the ones from
the build directory and build them during make all.
The various small utilities do not use the engine Log implementation but
one that simply prints to stdout. Instead of duplicating that one, link a
common one into the utilities.
Although the code already uses boost, boost/uuid hides the sha1
implementation in a deeply nested namespace, which is just too bizarre to
use. Also the name of that namespace suggests that it is just an
implementation detail that could go away without notice.
This replaces the BUILD_TO_PLANET option and makes in-tree Makefile builds
run without further work. Out-of-tree builds need a symlink to planet from
the build directory.
This makes it possible to ship the bundle stand-alone. Also
note that CMake will automatically pack the game data for
release builds, but sym-link the game data for debug builds.
Note this means you will only see the parts of planet/ that
are mentioned in OC_C4GROUPS in CMakeList.txt! This is equivalent
to the behaviour of the shipped build, so I don't see this as
a problem.
Günther Brammer