2000-04-29 * changed the register_* routines to use the cctokens and to only work with non-globals. This is now up to the point where we change the structure-header to use __inherits and friends. This is a major rewrite, which will be finished in version 0.3.0. * changed the debugging output routines. * Release 0.2.9 2000-04-29 * changed mark_cpplines bracketings from #{..} to #[...] which makes parsing some c declarations easier. * Release 0.2.1 2000-04-24 * Release 0.2 2000-04-24 * seperated out *many* things :-), the least to mention is the extra/ subdir so that the gtk specifics are to be found in there. All the special hashes are now initialized from these files which look not unlike a C header. I am sure you will find out how these are parsed - for the moment is no doc here. 2000-04-23 * tokenccode.pl invented, makes parsing a lot easier and safe. perhaps one should base a doc-generator on top of it. hmm. * the include-file checker does now also check with -f, so one should be able to run the things in seperate stages. * don't make the nonstatic-method be based on the static-methods, it does actually create bigger codepieces (the nm says there are "W" routines for the static-methods being in use). * remember to use "-fno-rtti -fno-exceptions" with your c++ compiler or you will end up with *much* bigger objectfiles. 2000-04-22 * and so I'll start to rearrange the code to make it much cleaner and more obvious of what it does. In the course I'll try to make cc'moding of GDK work. The cc-header will go out as the official version 0.2 - or better, it will only go out as pre first... * flaw: need to grok "char* what[]" (gtkclist.h), fixed. * detected flaws in gtkextra, need to contact its author... 2000-04-21 * gdkdefs still included although not there. commented out in perl-script. * perl 5.002 does not like "for my $i". seperate out. * makefile runs the perl-script directly but you could end up with "interpreter not found". feed it explicitly into perl now. * an older version of g++ had complains, it thought the in-structure declaration of "( )" beiong "(...)" and the out-of-structure definition of that same inline with "( )" being a "(void)". Fix it with a oneliner in the script. * added make freshcopy to ease things with testing. 2000-04-20 * tried the skript on gtkextra, and so I came about a bug... * the regex for register_function wasn't quite behaving if stomping at a tab. t'seems a tab isn't actually matching \s, damn. Well, perl has a \b, so let's use it. Yep, it works :-). * the gdk-tweak has gone wrong, damn. so you shouldn't apply it for the moment, sorry. It's sth. about some typedefs from gdkprivate.h being needed in gdktypes.h to declare the inline-prototypes inside the struct declaration, but gdkprivate.h does include gdktypes.h. 2000-04-19 * doc fixes, the gtk-- lives now at sourceforge and the include header is of course (by default) in * typo in pl-source: redefined value must match key of source lookup (t'seems I've been a lil' tired lately..) 2000-04-18 * bug: applying the script on glib.h will make 'struct _GDate' into _gdate so that it does not recognize g_date_* procedures. * not fixed: applying the script on gdk.h/gdktypes.h will emit the inline-definitions to early because there are two(!!) extern"C" sections (possibly due to concatenation of both these header files). * added a Makefile with tar: and install: targets 2000-04-16 * first version being put on sourceware.de, dubbed gtk-cc for now