On Mon, 5 Apr 2021, Markus Elfring wrote:

> >> I would like to avoid the repetition of parsing efforts as much as possible.
> >> Under which circumstances can replacement lists be taken better into account?
> >
> > Why does my suggestion involve a repetition of parsing effort?
>
> The selection of the applied programming interfaces has got significant influences
> on the run time behaviour.
>
> See also:
> https://github.com/coccinelle/coccinelle/issues/200#issuecomment-653775288
>
>
> > You want to use a regexp.
>
> This view depends on some factors.
> I would prefer to search for string literals (and their exclusion) by higher level means.
>
>
> > I'm asking you to put the regexp in a python function.
>
> How do you think about to improve the following software situation
> besides the application of regular expressions?
>
> @initialize:python@
> @@
> import re
>
> @display@
> identifier i =~ "^(?:[A-Z]+_){3,3}[A-Z]+";
> expression e : script:python() { re.match('"', e) };
> @@
> *#define i e
>
>
> elfring@Sonne:~/Projekte/PipeWire/lokal> spatch ~/Projekte/Coccinelle/janitor/show_define_usage7.cocci spa/include/spa/node/type-info.h
> …
>   File "<string>", line 5
>     coccinelle.result = (re . match ( " , e ))
>                                              ^
> SyntaxError: EOL while scanning string literal

This looks like a problem.  Thanks for the report.

julia