Re: [PATCH] etags: Add a small list of interpretors for Python

[Iustin Pop <iustin@google.com>]

On Wed, Sep 01, 2010 at 03:09:14PM +0200, Francesco Potortì wrote:
> >Currently, etags has no interpretors defined for Python, which prevents
> >it from recognising Python scripts that have no extension. This patch
> >adds a small list of interpretors for this language, which while not
> >perfect (e.g. it will fail on specific versions like python2.4), it is
> >better than the current situation where all scripts without an extension
> >will default to Fortran.
> 
> Sure.
> 
> >Note: I debated adding a more detailed list, but it would get outdated
> >quickly, so I left just these three versions. A regular expression for
> >the interpreter name would be much better here, would implementing that
> >make sense?
> 
> It is simpler than that, see below.

Sorry for the very late response.

> >+ static const char *Python_interpreters [] =
> >+   { "python", "python2", "python3", NULL };
> 
> You just add "python" here.
> 
> >!   { "python",Python_help,Python_functions,Python_suffixes,NULL,Python_interpreters},
> 
> This is allright.
> 
> Moreover, in get_language_from_interpreter you change this:
> 
>   for (lang = lang_names; lang->name != NULL; lang++)
>     if (lang->interpreters != NULL)
>       for (iname = lang->interpreters; *iname != NULL; iname++)
>         if (streq (*iname, interpreter))
>             return lang;
> 
> to something like this:
> 
>   for (lang = lang_names; lang->name != NULL; lang++)
>     if (lang->interpreters != NULL)
>       for (iname = lang->interpreters; *iname != NULL; iname++)
> -->     if (strneq (*iname, interpreter, strlen(*iname)))     <--
>             return lang;

I did this, but then:

> This changes the semantics of get_language_from_interpreter so that the
> list of interpreters is now a list of initial substrings that are
> accepted as interpreters names: I think that this new semantic is more
> useful than the previous one.
> 
> If this change is done, one should look through the etags builtin help,
> the man page and the info manual and update them with the new semantics,
> if necessary.

I looked but didn't find any documentation about this; probably due the
fact that the interpretors are not user-customizable?

If you can point me to which docs needs to be changed, I'll be glad to.
Neither doc/man/etags.1 nor doc/emacs/maintaining.texi list anything
about it.

Interdiff from the previous patch:

=== modified file 'lib-src/etags.c'
--- lib-src/etags.c     2010-12-23 13:52:41 +0000
+++ lib-src/etags.c     2010-12-23 14:05:35 +0000
@@ -757,7 +757,7 @@
 static const char *Python_suffixes [] =
   { "py", NULL };
 static const char *Python_interpreters [] =
-  { "python", "python2", "python3", NULL };
+  { "python", NULL };
 static const char Python_help [] =
 "In Python code, `def' or `class' at the beginning of a line\n\
 generate a tag.";
@@ -1477,7 +1477,7 @@
   for (lang = lang_names; lang->name != NULL; lang++)
     if (lang->interpreters != NULL)
       for (iname = lang->interpreters; *iname != NULL; iname++)
-       if (streq (*iname, interpreter))
+       if (strneq (*iname, interpreter, strlen(*iname)))
            return lang;
 
   return NULL;

-- 
thanks,
iustin