Each engine now only has to provide a single configure.engine file
adding the engine into the configure script, which then produces the
required other files automatically.
This is the third and final commit enabling fully pluggable engines.
Now providing an engine folder contains a configure.engine, engine.mk
and engine-plugin.h file, it will be picked up automatically by the
configure script.
This is the second part of allowing engines to be added dynamically.
Each folder in engines/ which must contain a file named "engine.mk"
containing the make definitions for that engine.
This is the first part of allowing engines to be added dynamically.
They are placed into a folder in engines/ which must contain a file
named "configure.engine" to add the engine, which is pulled into the
top level configure script automatically.
This allows to keep the engines to specfiy the files for translation close to
the engine sources itself.
Thanks to criezy for his suggestion on this approach.
Also changed the German HELP command to HILF at the translator
request (which means both HILF and HILFE work since the parser
only matches the start of the string).
This affects the Console / debugger classes of multiple engines.
An alternative solution would have been to remove the unused _vm
member vars. However, it seems likely that in the future, the _vm
member could be useful for methods added to the console. So instead,
we add a simple assert(_vm) to silence the clang warning.
The original interpreter only used english commands in the terminal
even when playing one of the localised version (and even though
everything else in the terminal was localised). This adds the possibility
to either use the English commands or the localised ones.
Localized commands have been added for French, German (thanks to
SimSaw, who also proposed that idea) and Italian (thanks to Maff).
This fixes bug #3590814.
The case for KP0-KP9 can probably be removed as it should be
handled by the case above. But I don't have a keyboard with
keypad to test. So I am leaving it for now.
It could re-delete previously deleted objects, causing
ex data (frame/text) corruption. This is the likely cause of bug #3591088.
Asm conversion error from eaf87bdfa7.
This is partly for readability and partly to avoid any possible
portability issues i.e. though they should be the same, size_t is
defined by system headers, whereas uint is defined within our build
system.
Also, replaced a array size calculation with our ARRAYSIZE macro.
Removed the buffer from being a global non-const variable. Also, the
code changes should allow the buffer size to be increased by just
changing the size of _keyBuffer if needed.
This is possible, since the game is freeware now. It is very
helpful when playing the game in order to avoid referring to the
manual. The extra help text is shown when the player uses the
'help' command inside the game terminal.
Also, the list of games where copy protection is skipped has been
reordered alphabetically in the README
Theoretically, it should be almost as big as _inputLine, so let's
make them the exact same size just for simplicity. This fixes crashes
for me when trying to use very long parameters to the monitor's READ
command, and could possibly fix a bug reported in the German version
where you supposedly have to read "kurzmitteilung".
Using Ryan's watch within the game, the watch time always started at
19.30 from point of use, which did not seem correct.
Checking with the original CD and Floppy interpreters under DOSBox
showed they used the current system time, so this was incorrect.
Bisection shows that this regression was introduced by commit
57e940f678 i.e. "DREAMWEB: Move all saved
variables to a GameVars struct" and was probably a side effect of a
minor call ordering change in the equivalent of setupInitialVars()
with respect to the getTime() call.
However, to ensure no further regressions, it was easier to fix this
by replacing the initial value setting by a getTime() call.
