GNOME Bugzilla – Bug 512155
gets confused by multiline typedef
Last modified: 2010-01-01 23:03:19 UTC
The following typedef: -typedef cairo_status_t -(*cairo_pdf_operators_use_font_subset_t) (unsigned int font_id, - unsigned int subset_id, - void *closure); makes gtk-doc think cairo_status_t is a function. Reformatting to: +typedef cairo_status_t (*cairo_pdf_operators_use_font_subset_t) (unsigned int font_id, + unsigned int subset_id, + void *closure); fixes it.
Yes, the beauty of a perl-regexp parser at work trying to parse c :( I'll see what I can do.
I'll added some variants to the tests and its actually only the first one that fails. I don't see an easy fix though :/ /** * bug_512155a_function_pointer_t: * * http://bugzilla.gnome.org/show_bug.cgi?id=512155 */ typedef int (*bug_512155a_function_pointer_t) (unsigned int arg1, unsigned int arg2, unsigned int arg3); /** * bug_512155b_function_pointer_t: * * http://bugzilla.gnome.org/show_bug.cgi?id=512155 */ typedef int (*bug_512155b_function_pointer_t) (unsigned int arg1, unsigned int arg2, unsigned int arg3); /** * bug_512155c_function_pointer_t: * * http://bugzilla.gnome.org/show_bug.cgi?id=512155 */ typedef int (*bug_512155c_function_pointer_t) (unsigned int arg1, unsigned int arg2, unsigned int arg3);
commit 8d4e46323b6ea82ed966c4b65f1344b651e20255 Author: Stefan Kost <ensonic@users.sf.net> Date: Sat Jan 2 01:00:03 2010 +0200 scan: improve the scanner, Fixes #512155 Handle multiline typedefs. Better debug logging. Handle function pointers better.