[Scummvm-devel] Release plans, once again

[Marcus Comstedt]

Max Horn <max at quendi.de> writes:

>> Yes, this idea does indeed have merit.  The unloading will cause some
>> memory fragmentation, but it probably won't be a problem.
>>
> Indeed, but I see not much we can do about that... :-/

Theoretically I could relocate the remaining module again if I keep
the reloc data, but this would get messy if there are external
references to relocated code/data, such as virttables.  Better not go
there...


>> I'll have to write the dynamic loader myself anyway, so I can add
>> whatever features I like.
>
> Aye. The current "Plugin" code isn't set in stone, by the way, it's
> still unfinished in some ways... if you need changes or have
> questions, feel free to discuss this with me :-)

Well, the first change I seem to need is moving the "standard"
definitions of PLUGIN_PREFIX and PLUGIN_SUFFIX from Makefile.common to
Makefile so that I can override them.  No problem with that I suppose?

Secondly, I'll need to add another dependency to the dynamic plugin
link rule; right now it depends on $(EXECUTABLE), but I need it to
depend on a linker script generated _from_ $(EXECUTABLE).  Any
suggestions for a name for the makefile variable to hold this extra
dependency?


  // Marcus