Part of the verbosity of the previous version was due to me having
misinterpreted a SMPL parse error. Here is a more concise (although still
not optimal) version. Now it will also convert partially converted
structures.
Although, actually it might not do the right thing for eg:
struct ty { int a; int b; int c; int d; };
struct ty x = { 1, .b = 2, 3, 4 };
I am assuming that nothing needs to be done to .b = 2 and that c matches
up with 3 and d with 4. But what if it had been instead:
struct ty x = { 1, .c = 2, 3, 4 };
Does that means that b should match up with c?
Perhaps doing nothing for partiall C99 declarations was a better idea?
julia
---------------
@r@
field list[n] ds;
type T;
identifier i, virtual.ty;
@@
struct ty {
ds
T i;
...
};
@rr@
type T;
identifier i, virtual.ty;
@@
struct ty {
T i;
...
};
// --------------------------------------------------------------------
@@
initializer list[r.n] is;
expression a;
identifier nm;
identifier r.i, virtual.ty;
@@
struct ty nm = {
is,
+ .i =
a,
...
};
@@
expression a;
identifier nm;
identifier rr.i, virtual.ty;
@@
struct ty nm = {
+ .i =
a,
...
};
// --------------------------------------------------------------------
@@
initializer list[r.n] is;
expression a;
identifier nm;
identifier r.i, virtual.ty;
@@
struct ty nm[...] = { ..., {
is,
+ .i =
a,
...
}, ...};
@@
expression a;
identifier nm;
identifier rr.i, virtual.ty;
@@
struct ty nm[...] = { ..., {
+ .i =
a,
...
}, ...};
// --------------------------------------------------------------------
@@
initializer list[r.n] is;
expression a;
identifier nm;
identifier r.i, virtual.ty;
@@
struct ty nm[] = { ..., {
is,
+ .i =
a,
...
}, ...};
@@
expression a;
identifier nm;
identifier rr.i, virtual.ty;
@@
struct ty nm[] = { ..., {
+ .i =
a,
...
}, ...};
_______________________________________________
Cocci mailing list
[email protected]
http://lists.diku.dk/mailman/listinfo/cocci
(Web access from inside DIKUs LAN only)
