Re: and/or missbehaviour in embedded lua 5.1

[Andreas Stenius <kaos@...>]

Kein-Hong Man skrev:
> Andreas Stenius wrote:
> > Kein-Hong Man skrev:
> > > It should be like x86 except for int... hmmm... yeah, the debugging 
> > > nightmare would be if it's a problem with the C compiler. So it 
> > > could be any one of the following:
> > > (a) some place in the code generator doesn't like an int of 2 and it 
> > > borks
> > > (b) C compiler miscompiles the code generator due to the way it 
> > > interprets C, some lack of information, some kind of shortcut taken 
> > > or an internal bug
> > > (c) double type problem due to math library or conversion
> > > (d) ???
> > >
> > > Without a debugger, peppering lcode.c with printfs should be fine.
> > Said and done :)
> > After a few hours of peppering, I got this log file: 
> > http://kaos.explosive.se/files/Lua_5.1_060317.TXT
> > Not inlined, due to it's quite long..
>
> Thanks, I'll take a look. Curiosity kills the kitty...

Thanks alot!

Sorry for the extra noise, but I've made some omprovements to the output 
of expdesc, realized that it's one of the interesting parts to look 
closer at..
new log here: http://kaos.explosive.se/files/Lua_5.1_060317_2.TXT

Just toss me a line if you're interested in the results of anything 
besides "=true and 1".

I found this part the most interesting at a quick glance:

lcode.c:"luaK_posfix"(850): constants (0) for 0x1204f2:
op: 13, e1 (0x1aac): VTRUE, t: -1, e2 (0x1a4e): VKNUM, nval: 1

lcode.c:"luaK_dischargevars"(398): constants (0) for 0x1204f2:
e (0x1a4e): VKNUM, nval: 1

lcode.c:"luaK_concat"(267): constants (0) for 0x1204f2:
l1: 6840, l2: 0

lcode.c:"luaK_exp2anyreg"(522): constants (0) for 0x1204f2:
e (0x1aac): VKNUM, nval: 7.9591334224152e-74


Noting that the expdesc e1 in luaK_posfix has the same adress (is the same?)
as the one on luaK_exp2anyreg, but with a different expkind. 
Anyone know if this is intended?


// Andreas