On Mon, 3 Feb 2020, Markus Elfring wrote:

> Hello,
>
> I have tried the following small script variant out for
> the semantic patch language.
>
> @replacement@
> constant c;
> identifier text;
> expression x;
> statement is, es;
> @@
> char* text = x(...);
>
> if (
> +   strcmp(
>     text
> +   ,
> -   == std::string(
>     c
>     )
>    )
>    is
> else
>    es
>
>
> A corresponding transformation result from the software
> “Coccinelle 1.0.8-00029-ga549b9f0” looks promising (in principle).
>
> elfring@Sonne:~/Projekte/Coccinelle/Probe> spatch --c++ replace_std_string_check_by_strcmp2.cocci PulseEffects-source_output_effects-excerpt2.cpp
> …
> @@ -1,7 +1,7 @@
>  void on_message_element(const GstBus* gst_bus, GstMessage* message, SourceOutputEffects* soe) {
>    char* src_name = GST_OBJECT_NAME(message->src);
>
> -  if (src_name == std::string("equalizer_input_level")) {
> +  if (strcmp(src_name, "equalizer_input_level")) {
>      soe->equalizer_input_level.emit(soe->get_peak(message));
>  // Deleted part
>    } else if (src_name == std::string("webrtc_output_level")) {
>
>
> 1. But I wonder about an additional space character at the beginning
>    of the shown function in the generated patch.

Where is the space that yoyu are concerned about?

>
> 2. Will it become possible to achieve a similar change
>    if the specification “auto” would be used instead of
>    the data type “char*”?
>    https://github.com/wwmm/pulseeffects/blob/acb5161a6ab8d3b0c395ed2809d3318ccf4931bc/src/source_output_effects.cpp#L6

I guess it would be fine if you put

typedef auto;

among the metavariables?

julia