| Age | Commit message (Collapse) | Author |
|---|
|
|
|
|
|
|
|
|
|
newer than 2.3
|
|
|
|
value), which clearly nobody uses, as it has been broken .... for a while
|
|
those who expect all software to follow GNU guidelines
|
|
really needed anyway
|
|
builds
|
|
helper binaries such as depend, strgen and settingsgen
|
|
compiled for the build enviroment
|
|
versioning scheme since ICU 49
|
|
aren't needed anymore with newer GRFCodec/NFORenum
|
|
GRFCodec determine the best compression
|
|
|
|
|
|
reduce the chance that it happens significantly with betas/RCs/nightlies by doing the same as is done for stable releases: check the NewGRF version of server vs client.
Previously this check was not done for nightlies/betas/RCs due to missing versioning information in the source tarballs, but they have that for a while now. So just force the NewGRF version check for all versions, and remove the broken --revision configure option
|
|
when grfcodec/nforenum were not usable.
|
|
|
|
configure, including a version check so a proper error can be given when a too old grfcodec or nforenum is used
|
|
parts some compilers failed at, are integrated in other parts of the code now too
|
|
prevent some invalid GCC warnings
|
|
many useless warnings with GCC 4.7
|
|
|
|
without using it (Rubidium)
|
|
|
|
config.log
|
|
unusable
|
|
with LTO use number of processes passed to make -j
|
|
compiler can do more optimisations
|
|
official interface, so update the variable/definition naming slightly as it's not broken just different from the earlier iconv interface
|
|
|
|
|
|
you'll more likely get the right version/revision out-of-the-box
|
|
natural sort.
|
|
|
|
builds
|
|
application under MinGW.
|
|
work around a compiler bug in gcc-4.0.1 which breaks graphics display in x86_64 binaries (Rubidium)
|
|
|
|
savegame size by 10 to 30% with slightly more CPU usage. With maximum settings it reduces savegame size by 20 to 30%, but that takes 7 to 14 times longer. Map saving + downloading takes, on average, 5% less.
|
|
|
|
information and were not optimised at all, causing many compile-time warnings
|
|
other compilers would likely barf on GLIBC code
|
|
|
|
|
|
bit easier.
The (prebuilt) grf will still remain, so if you don't have NFORenum and GRFCodec it will still work.
|
|
|
|
when building with asserts disabled
|
