[Cocci] Specifying that an expression is a string constant?

[Cocci] Specifying that an expression is a string constant?

[Timur Tabi]

The following rule fails with an error on the last line:

@rule3@
expression x, y, z;
@@
-DBG_PRINT_STRING_VALUE(x, y, z);
+NV_PRINTF2(x, y "0x%x\n", z);

Presumably, cocci doesn't know that "y" is a string constant and so
the resulting expression is valid.  I'm guessing that I need to
specify that "y" is a string constant, but even after reading
http://coccinelle.lip6.fr/docs/main_grammar.html, I can't figure out
what that could be.
_______________________________________________
Cocci mailing list
Cocci@systeme.lip6.fr
https://systeme.lip6.fr/mailman/listinfo/cocci

Re: [Cocci] Specifying that an expression is a string constant?

[Julia Lawall]

On Tue, 19 Feb 2019, Timur Tabi wrote:

> The following rule fails with an error on the last line:
>
> @rule3@
> expression x, y, z;
> @@
> -DBG_PRINT_STRING_VALUE(x, y, z);
> +NV_PRINTF2(x, y "0x%x\n", z);
>
> Presumably, cocci doesn't know that "y" is a string constant and so
> the resulting expression is valid.  I'm guessing that I need to
> specify that "y" is a string constant, but even after reading
> http://coccinelle.lip6.fr/docs/main_grammar.html, I can't figure out
> what that could be.

SmPL doesn't support this.  The only solution I can see is to use python
code to construct an identifier variable that actually has the form of y
"0x%x\n".  You can see examples of how to use python to initialize
identifier metavariables in demos/pythontococci.cocci

julia
_______________________________________________
Cocci mailing list
Cocci@systeme.lip6.fr
https://systeme.lip6.fr/mailman/listinfo/cocci

Re: [Cocci] Specifying that an expression is a string constant?

[Markus Elfring]

>> The following rule fails with an error on the last line:
>>
>> @rule3@
>> expression x, y, z;
>> @@
>> -DBG_PRINT_STRING_VALUE(x, y, z);
>> +NV_PRINTF2(x, y "0x%x\n", z);
>>
>> Presumably, cocci doesn't know that "y" is a string constant and so
>> the resulting expression is valid.  I'm guessing that I need to
>> specify that "y" is a string constant, but even after reading
>> http://coccinelle.lip6.fr/docs/main_grammar.html, I can't figure out
>> what that could be.
>
> SmPL doesn't support this.

Will any users (or contributors) become motivated enough to adjust involved
software limitations also for wider applications of the semantic patch language?

Regards,
Markus
_______________________________________________
Cocci mailing list
Cocci@systeme.lip6.fr
https://systeme.lip6.fr/mailman/listinfo/cocci