Re: [Tinycc-devel] Largish mob update from me

[grischka]

Michael Matz wrote:
> Unfortunately I had to add one other place where it checks nocode_wanted
> and that's in the ugly dealing with VT_CMP/VT_JMP values in vtop.  See
> testcases in the above, and my comment included here for citing/discussing:
>
> +       Don't do this when nocode_wanted.  vtop might come from
> +       !nocode_wanted regions (see 88_codeopt.c) and transforming
> +       it to a register without actually generating code is wrong
> +       as their value might still be used for real.  All values
> +       we push under nocode_wanted will eventually be popped
> +       again, so that the VT_CMP/VT_JMP value will be in vtop
> +       when code is unsuppressed again.  */
> +    if (vtop >= vstack && !nocode_wanted) {
>          v = vtop->r & VT_VALMASK;
>          if (v == VT_CMP || (v & ~1) == VT_JMP)
>              gv(RC_INT);
>
> I hope I haven't broken anything that you've fixed with your three 
> nocode_wanted patches ...

No, I guess it makes sense.  Except there is the same clause in
vswap() down below which we'll probably want to treat the same way.

Also I'd share the assumption that nocode-branches must be neutral
to vstack and it is up to the nocode-user to care about.

The gsym() in vpop() is ok IMO, it will fixup a jump chain on
vtop to the current "ind" pointer. That 'ind' doesn't change
under 'nocode' does not matter.

Btw, what's "pseudo leak" in a one-shot tool, in libtcc will write
to memory that already has been free'd by the underlying "TAL" manager.
Sorry if I didn't fix it it was just to watch my new memtest at work ;)

Vlad Vissoultchev's T(iny) AL(locater):
http://repo.or.cz/tinycc.git/commitdiff/cdc16d428f32294

If someone wants to try the difference, comment out tccpp.c:112:
    #define USE_TAL

--- grischka