It seems the recent MinGW gcc versions use the "TMP" env variable to determine
the temporary file path. Since we use that variable to hold temporary values it
will make g++ crash since it cannot access the path.
I "fixed" that problem by renaming the TMP variable we use to TMPR.
Tested with mingw-g++ 4.5.2.
WinCE is using -O3 by default, which includes -finline-functions, which in turn
breaks SCI. By disabling that opimization we get SCI working fine again.
Only applies when using loadable modules.
This requires a second compile run or black voodoo linker tricks.
Not implemented at this time. Disable it, so we get DS build failure
reports from buildbot (which are useful with all the DISABLE_EVERYTHING
defines).
The resulting binary might be too big to be usable.
mt32emu is using floats everywhere, and we don't have hardware-assisted
floating point computations on this armeabi. most (if any) devices can't
handle this
timidity doesn't make sense, because there's no daemon to connect to
Add an object with an empty .ctors and .dtors section to the build
of each plugin. This ensures that there will be an output section
.ctors and one .dtors in the output file (they could be empty though).
If these output sections were to be removed, the symbols defined inside
by the linker script could disappear as well, and we don't want that.
svn-id: r55149
This changes removes the tmpnam warning when building the Sword 2.5 engine.
This fix sets the top level define which is used by engines/sword25/util/lua/luaconf.h to configure the LUA interpreter build, but only for linux* targets.
Other targets may need a similar fix.
Note : LUA_USE_POSIX is used in preference to LUA_USE_LINUX as this causes another warning associated with the dlopen usage.
svn-id: r55130
Following lordhoto's suggestion, I implemented a simple allocator that grabs the size of the biggest available plugin in memory. This is an elegant solution to the fragmentation problem, with the caveat that memory is wasted. As such, it's not suited for the DS, so I added a #define to disable it there.
svn-id: r55009
This merge was extremely difficult to carry out. It wasn't entirely SVN's fault -- there were several merges to the branch that were done by hand. Please check for any issues and regressions. Also note that the DS makefile was not copied over since the "one at a time" plugin mode currently has too much fragmentation ie. it doesn't work.
svn-id: r54051
