-Mon Jan 24 22:52:47 2000 George Lebl <jirka@5z.com>
-
- * Release 0.92.2
-
-Mon Jan 24 22:23:47 2000 George Lebl <jirka@5z.com>
-
- * src/main.c: add a protecting define before the object typedef
-
- * doc/gob.1.in: add docs for the new stuff
-
-Sun Jan 23 16:19:54 2000 George Lebl <jirka@5z.com>
-
- * src/{lexer.l,parse.y,tree.[ch],main.c}: make more types of header
- %{ %} blocks, "headertop" and "privateheader" and made short as
- well as long names possible.
-
- * src/main.c: add an --no-extern-c option to not put entern "C" into
- the output, which is now put in by default even if --for-cpp is
- on so that it is possible to call those from C by default
-
-Thu Jan 20 01:50:45 2000 George Lebl <jirka@5z.com>
-
- * src/main.c: put the _real_ functions in our normal private
- namespace of ___* to be more consistent.
-
-Mon Jan 10 00:16:43 2000 George Lebl <jirka@5z.com>
-
- * src/{main.[ch],util.[ch],checks.[ch]}: offload some functionality to
- extra files
-
- * src/main.c: cleanup the generation routine by spliting it up into
- more functions
-
- * src/{main.h,lexer.l,parse.y,out.c}: put some needed externs into the
- main.h file and include that rather then doing it in the .c files
-
-Thu Jan 06 18:33:44 2000 George Lebl <jirka@5z.com>
-
- * src/main.c: use ___ for the signal marshalling typedefs as well
-
-Thu Jan 06 17:32:46 2000 George Lebl <jirka@5z.com>
-
- * Release 0.92.1
-
-Thu Jan 06 17:13:22 2000 George Lebl <jirka@5z.com>
-
- * src/main.c: be consistent with the use of "namespacing", we
- prefix ALL internal gob stuff with "___". This should make it
- easier to avoid name conflicts with user code and glibc/gcc (Which
- seems to use __ prefix)
-
- * src/main.c: check for init and class_init names for non-constructor
- methods and give an error if found.
-
- * src/main.c: check argument types and flags for correctness, error
- on unknown type, but just a warning on unknown flag
-
- * doc/gob.1.in: add a section for name conflicts
-
-Wed Jan 05 02:23:21 2000 George Lebl <jirka@5z.com>
-
- * src/main.c: doh! the symbols for the aliases need to be declared as
- static.
-
-Wed Jan 05 02:13:59 2000 George Lebl <jirka@5z.com>
-
- * src/main.c: just realized that args... for macros is a GNU
- extention and so the PARENT_HANDLER no longer uses it, but prints
- the argument names with __ prefixed into the argument list.
-
-Wed Jan 05 01:58:32 2000 George Lebl <jirka@5z.com>
-
- * src/main.c: if using gnu c, also define macros using varargs for
- function aliases. This should allow the compiler to optimize
- further and if we have arguments for the defines there are
- no adverse side effects as there are for straight defines.
-
-Wed Jan 05 01:48:12 2000 George Lebl <jirka@5z.com>
-
- * src/main.c: fixup the aliases. They didn't actually work right.
- now we set it up as a const pointer to a function using typeof.
-
-Tue Jan 04 17:57:55 2000 George Lebl <jirka@5z.com>
-
- * src/lexer.l: add support for 'requires x.y.z' line. This is
- so that a .gob file can require at least version x.y.z of gob
-
- * src/test.gob: update
-
- * doc/gob.1.in: update for requires and run trhough ispell
-
-Tue Jan 04 17:10:54 2000 George Lebl <jirka@5z.com>
-
- * src/main.c: fix the non-gnu aliases
-
-Tue Jan 04 16:46:16 2000 George Lebl <jirka@5z.com>
-
- * src/{main.c,parser.y,tree.[ch]}: add optional C type specification
- to arguments for full type safety
-
- * doc/gob.1.in: updated
-
-Tue Jan 04 15:34:16 2000 George Lebl <jirka@5z.com>
-
- * src/main.c: --no-gnu option to completely disable any use
- of GNU C extentions. This is for those weird people that don't
- wish to use those extentions even with GNU C.
-
-Tue Jan 04 14:43:49 2000 George Lebl <jirka@5z.com>
-
- * src/main.c: use some GNU extentions (guarded with #ifdef __GNUC__)
- to make arguments fully typesafe, and to make it easier for
- GNU to optimize the aliases, as well as to avoid the ugly
- hack on GNU C.
-
- * src/main.c: don't put #line 0 for the .gob file for method prototypes
-
- * src/main.c: move the _priv structure freeing into finalize to enable
- people to work with the private stuff inside the destroy handler
-
- * src/main.c: use #line before every precondition check so that failed
- checks should point to the function prototype rather then into
- generated file.
-
- * src/main.c: remove the _END_ARG macro
-
- * src/main.c: add GOB_VERSION_MAJOR, GOB_VERSION_MINOR and
- GOB_VERSION_PATCHLEVEL defines into the generated C file
-
- * src/out.c: don't redundantly put #line for outfile if we're already
- in the outfile
-
- * configure.in: raise version number
-
-Tue Jan 04 02:22:33 2000 George Lebl <jirka@5z.com>
-
- * src/main.c: add argument wrapper macros to make the argument system
- a little more type safe.
-
-Fri Dec 31 17:39:22 1999 George Lebl <jirka@5z.com>
-
- * Release 0.92.0
-
-Thu Dec 30 20:20:36 1999 George Lebl <jirka@5z.com>
-
- * src/main.c: correctly handle the complex fundemental types,
- so that we can handle ALL the fundemental types for signals
-
-Thu Dec 30 19:29:56 1999 George Lebl <jirka@5z.com>
-
- * src/main.c: don't use function types for casts in marshallers,
- but default types we get from our GTK_TYPE_* -> C type translation,
- and use this fact to check for same function types and combine
- signal prototypes. This doesn't yet handle well complex types,
- but I have an idea how to make complex types work. Though
- I dunno if it's actually worth it.
-
-Thu Dec 30 17:18:27 1999 George Lebl <jirka@5z.com>
-
- * src/main.c: figure out some types from GTK types and cast properly
- when emitting a signal. This catches some inconsistencies in
- signal func argument lists and the GTK types of the arguments.
-
-Wed Dec 29 18:04:13 1999 George Lebl <jirka@5z.com>
-
- * src/parse.y: allow scope, public/private/protected, to come before
- the signal or virtual keyword for consistency's sake.
-
-Wed Dec 29 01:39:30 1999 George Lebl <jirka@5z.com>
-
- * src/{tree.[ch],parser.y}: cleanup scope stuff
-
- * src/{lexer.l,parser.y,tree.h,main.c}: add protected keyword and
- methods. protected methods are put into the -private files and
- are real exported functions just like public. protected data
- members act like pre-0.91.x private datamembers, that is they
- get stuck into the public header, but marked /* protected */.
-
- * src/main.c: change the method body printing function so that we
- can catch the no return warning properly and send us to the .gob
- file instead of the .c file.
-
- * src/{parse.y,main.c}: handle empty publics/privates/overrides
- and handle the body being set to ';' rather then '{}'
-
- * src/main.c: check for empty nonvoid regular methods, and error out
- on finding them as that would produce warnings which we can't catch
- and it is generally bad to do anyway
-
- * src/{parse.y,tree.[ch],main.c}: use char pointers for C buffers,
- and kill leading/trailing whitespace from them to make the C files
- nicer and easier to recognize empty functions.
-
- * src/main.c: add a --version switch
-
- * doc/gob.1.in: Update with the above changes
-
- * configure.in: raise version
-
-Tue Dec 28 20:49:56 1999 George Lebl <jirka@5z.com>
-
- * src/main.c: apply patch from ChiDeok Hwang <cdhwang@sr.hei.co.kr>
- to make the hack function static so that we don't pollute namespace
-
-Mon Dec 27 19:56:24 1999 George Lebl <jirka@5z.com>
-
- * Release 0.91.2
-
-Mon Dec 27 03:00:25 1999 George Lebl <jirka@5z.com>
-
- * src/main.c: fixup PARENT_HANDLER to work well with non-void
- functions, the macro becomes an expression if the function is
- not void and returns the return of the parent or the onerror
- expression if the function pointer was NULL.
-
-Mon Dec 27 02:25:32 1999 George Lebl <jirka@5z.com>
-
- * src/parse.y: allow completely empty classes
-
- * src/main.c: don't add the ugly hack function if there are no
- methods
-
-Mon Dec 27 01:26:26 1999 George Lebl <jirka@5z.com>
-
- * src/parse.y: fixup the virtual rule as I forgot to shift the
- argument numbers when adding to the expression
-
-Mon Dec 27 01:07:14 1999 George Lebl <jirka@5z.com>
-
- * src/parse.y: add optional public keyword to virtuals and signals,
- (which doesn't do anything, only consistency). I should perhaps
- rewrite and simplify this part of the parser
-
- * src/lexer.l: fix a bug with empty braces producing a segfault
- on public and private methods
-
- * doc/gob.1.in: add a section about Constructor methods (init
- class_init) and fix init_class -> class_init
-
- * configure.in: raise version to 0.91.2
-
-Tue Dec 14 00:06:00 1999 George Lebl <jirka@5z.com>
-
- * Release 0.91.1
-
-Mon Dec 13 23:25:06 1999 George Lebl <jirka@5z.com>
-
- * src/{main.c,out.c}: create a private header file and stick
- the private structure definition inside it. Add option
- --no-private-header for 0.91.0 behaviour and
- --always-private-header to force private header creation
-
- * doc/gob.1.in: update WRT above changes
-
- * examples/my-person.gob: include the private header
-
- * NEWS: update
-
-Mon Dec 13 18:30:09 1999 George Lebl <jirka@5z.com>
-
- * src/main.c: don't print spurious #line's into the private
- structure
-
- * src/{main.c,tree.[ch],lexer.l,parse.y}: implement support for
- array types for data members and function arguments
-
-Mon Dec 13 00:48:47 1999 George Lebl <jirka@5z.com>
-
- * Release 0.91.0
-
-Sun Dec 12 22:55:12 1999 George Lebl <jirka@5z.com>
-
- * doc/gob.1.in: added paragraph noting that private data members
- cannot be accessed above the class definition in the .gob file
-
-Sun Dec 12 22:08:31 1999 George Lebl <jirka@5z.com>
-
- * src/main.c: implement real private data members as promised in
- the documentation, this however breaks some compatiblity with
- things that already used the 'private' keyword for data members
-
- * doc/gob.1.in: document private data member stuff
-
- * NEWS: update
-
- * src/main.c: a couple of sanity checks of the code to prevent weird
- errors on some broken .gob code
-
- * configure.in: update version to 0.91.0
-
-Sun Dec 05 14:20:26 1999 George Lebl <jirka@5z.com>
-
- * src/{lexer.l,main.c}: if we find the class header #include
- statement in a %{ %} section above class definition, don't
- include it ourselves.
-
- * doc/gob.1.in: remove the note about the include always first from
- BUGS, and make a new section for the include file.
-
-Thu Nov 25 13:09:08 1999 George Lebl <jirka@5z.com>
-
- * Release 0.90.5
-
-Thu Nov 25 13:00:52 1999 George Lebl <jirka@5z.com>
-
- * src/main.c: don't error out on a symbol conflict if the types of
- the symbol node don't match (variable,method)
-
-Sat Nov 20 16:15:42 1999 George Lebl <jirka@5z.com>
-
- * src/main.c: add PARENT_HANDLER macro to make calling parent
- handlers in override functions easier
-
-Fri Nov 19 16:41:09 1999 George Lebl <jirka@5z.com>
-
- * src/main.c: add a type macro
-
-Thu Nov 18 22:56:09 1999 George Lebl <jirka@5z.com>
-
- * src/parse.y: add a warning check if the number of GTK types of
- a signal doesn't seem to be correct for the given number of
- function arguments
-
-Tue Nov 16 01:23:45 1999 George Lebl <jirka@5z.com>
-
- * src/main.c: support a "no-touch-headers" mode in which the
- headers are not touched unless they actually really changed,
- not on by default as it confuses automake
-
-Tue Nov 16 00:36:42 1999 George Lebl <jirka@5z.com>
-
- * src/{main.c,lexer.l,parser.y,out.c}: Implement a C++ mode, get
- rid of C++ errors/warnings when in C++ mode, and get rid of
- the 'this' pointer finally (it's been deprecated long enough)
-
-Mon Nov 15 23:45:37 1999 George Lebl <jirka@5z.com>
-
- * src/main.c: add prototype for the really bad hack function to
- avoid warning
-
-Mon Nov 15 23:42:06 1999 George Lebl <jirka@5z.com>
-
- * src/main.c: generate correct function pointers in the structure
- for signals and virtuals. Also use a much nicer system instead
- of the method name defines, we now just set a bunch of static
- pointers, this removes a whole lot of headaches actually
-
- * doc/gob.1.in: repair to reflect the above (I removed the define
- BUGS paragraph)
-
-Sun Nov 14 17:57:34 1999 George Lebl <jirka@5z.com>
-
- * doc/gob.1.in: repair some things that confused troff and add
- a C preprocessor warning to BUGS
-
-Sat Nov 13 17:22:49 1999 George Lebl <jirka@5z.com>
-
- * **/Makefile.am, configure.in: fix the conditional installation
- stuff, don't even go into the doc directory if we aren't installing
-
-Sat Nov 13 16:53:20 1999 George Lebl <jirka@5z.com>
-
- * doc/gob.1.in: describe BUGS better
-
-Sat Nov 13 16:20:41 1999 George Lebl <jirka@5z.com>
-
- * src/lexer.l: don't kill C++ style comments from C blocks
-
-Sat Nov 13 16:16:59 1999 George Lebl <jirka@5z.com>
-
- * src/{lexer.l,main.c,out.c}: fixed spurious newline printing, don't
- kill comments from C blocks so that line counts are right, accept
- newlines in strings, and don't take \'\"\' as a string delimiter
-
-Sat Nov 13 14:53:26 1999 George Lebl <jirka@5z.com>
-
- * src/main.c: don't add gtk_object_class to class_init if we only
- have non-GtkObject overrides to avoid warning
-
-Sat Nov 13 14:31:25 1999 George Lebl <jirka@5z.com>
-
- * configure.in,src/Makefile.am,doc/Makefile.am: check for
- ../NOINST_GOB and in case it's found don't install self
-
-Fri Nov 12 16:25:35 1999 George Lebl <jirka@5z.com>
-
- * configure.in,Makefile.am,src/main.c: don't use popt, our arguments
- are simple enough and popt doesn't seem to be widely used yet
-
-Sun Sep 05 22:08:40 1999 George Lebl <jirka@5z.com>
-
- * Release 0.90.2
-
-Sun Sep 05 22:08:33 1999 George Lebl <jirka@5z.com>
-
- * doc/gob.1.in: updated for private virtual and signal wrappers
-
-Sun Sep 05 21:50:19 1999 George Lebl <jirka@5z.com>
-
- * configure.in: raise version and actually set POPT_LIB correctly
-
- * src/Makefile.am: use POPT_LIB
-
-Sun Sep 05 21:45:54 1999 George Lebl <jirka@5z.com>
-
- * src/tree.h,src/main.c,src/parse.y: add support for private signals
- and virtuals
-
- * src/parse.y: fix bug with bogus vararg methods
-
-Sun Sep 04 17:44:37 1999 George Lebl <jirka@5z.com>
-
- * Released 0.90.1
-
-Sat Sep 04 17:18:22 1999 George Lebl <jirka@5z.com>
-
- * src/lexer.l,src/parse.y,src/main.c,src/tree.[ch]: support
- variable arguments for public and private methods
-
-Sat Sep 04 16:51:11 1999 George Lebl <jirka@5z.com>
-
- * lexer.l,parse.y: accept const in parameter lists
-
-Sun Aug 29 13:46:33 1999 George Lebl <jirka@5z.com>
-
- * Release 0.90.0
-
-Sat Aug 28 23:36:48 1999 George Lebl <jirka@5z.com>
-
- * src/main.[ch],parse.y: make error/warning reporting public
- and use it to report depreciated use of 'this'
-
-Sat Aug 28 22:41:52 1999 George Lebl <jirka@5z.com>
-
- * src/tree.[ch],parse.y,main.c: check for duplicate variables and
- signals and arguments as well.
-
-Sat Aug 28 22:00:19 1999 George Lebl <jirka@5z.com>
-
- * src/parse.y,src/main.c: change "this" to "self", but accept "this"
- as well
-
- * src/main.c: check for duplicate methods, and warn on non-public
- "new" method
-
- * configure.in: change to version 0.90.0
-
- * doc/gob.1.in: change this to self
-
-Tue Aug 24 20:37:26 1999 George Lebl <jirka@5z.com>
-
- * Release 0.0.4
-
-Thu Aug 19 03:18:03 1999 George Lebl <jirka@5z.com>
-
- * src/main.c: also count overrides to avoid an unused variable on
- class_init
-
-Thu Aug 19 03:08:49 1999 George Lebl <jirka@5z.com>
-
- * src/parse.y: in a check, we can check not only about numbers
- but for tokens as well
-
-Wed Aug 18 12:54:17 1999 George Lebl <jirka@5z.com>
-
- * Released 0.0.3
-
-Tue Aug 17 22:24:47 1999 George Lebl <jirka@5z.com>
-
- * src/lexer.l,src/parse.y: don't use reserved words for check type
- last first and null
-
- * src/{lexer.l,parse.y,tree.[ch],main.c}: add #line's to output
- files
-
- * src/out.[ch]: new functions for output so that we can easily add
- #line's to files
-
- * src/main.c: fix get/set arguments
-
- * doc/gob.1.in: updated for newer version and spellchecked, also
- gob.1 is generated by configure now to get versions correct
-
- * gob.spec.in: added a spec file
-
-Fri Jul 30 02:19:05 1999 George Lebl <jirka@5z.com>
-
- * src/parse.y: accept simple char types
-
- * src/main.c: make prototypes look better, and don't output arg stuff
- when it won't be used
-
- * Makefile,src/Makefile: add test make target
-
+Stub changelog, see the git history for changes.
projects / gob-dx.git / blobdiff