[Rpcemu] 0.8.6 64bit dynarec glitch
musus at verelanthe.co.uk
Sat Mar 20 15:44:31 PDT 2010
Hi list ;)
I've just built 0.8.6 (dynarec) on 64-bit Fedora 11 and I'm seeing some
strange behaviour. Initially I noticed it when trying to comment out a
RMEnsure line in an obey file, but that's not really important... What
/is/ important, however, is that a 32-bit build of 0.8.6 doesn't show
this behaviour at all, and nor does a 64-bit build without dynarec.
Anyway, what I'm seeing is that attempting to insert characters at the
start of a line longer than 30 characters in StrongEd causes the line to
be corrupted (eventually, deleting characters will crash StrongEd
For example... If I have the line:
displayed and attempt to insert any character at the start of it, it
Note that "0123" has become "1233". This pattern of corruption
continues for other insertions. Insertions at other points in the line
behave similarly (the corrupted character will be different depending on
the insertion point's (memory) word-alignment) as long as the insertion
point is more than ~30 characters from the end.
(For all I'm referring to a "line" here, the entire text is what
So, on the face of it, it looks like StrongEd has some code it uses to
move lumps of memory larger than ~30 bytes and the part of that that
deals with the 'dangling ends' of the block is failing in RPCemu.
FWIW, StrongEd version doesn't seem to make a difference; I initially
saw this with 4.60 so I upgraded to 4.68 (just in case ;) but that made
Can anybody reproduce this?
Erst wenn die Wolken schlafengehn | Personal: adny at verelanthe.co.uk
kann man uns am Himmel sehn | Techie: musus at verelanthe.co.uk
wir haben Angst und sind allein | WWW: verelanthe.co.uk/musus/
Gott weiss ich will kein Engel sein! | UT: adny.rsa at googlemail.com
-- Rammstein / Engel
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Size: 198 bytes
Desc: This is a digitally signed message part.
More information about the Rpcemu