=== ChangeLog discontinued ===
With the move to git, we stop maintaining a separate ChangeLog and
rely on proper commit messages instead. Web view of changes:
<http://git.gnome.org/browse/perl-Glib/>.
2009-03-31 Torsten Schoenfeld <kaffeetisch@gmx.de>
* t/a.t: The g_log bug got fixed upstream, so don't skip its test
when glib >= 2.20.1.
2009-03-29 Torsten Schoenfeld <kaffeetisch@gmx.de>
* t/a.t: Include 's390' in the list of platforms for which the
problematic g_log test needs to be skipped.
2009-03-29 Torsten Schoenfeld <kaffeetisch@gmx.de>
* copyright.pod
* Glib.pm
* NEWS
* README: Stable release 1.221.
2009-03-29 Torsten Schoenfeld <kaffeetisch@gmx.de>
* t/a.t: Using multiple log levels breaks g_log on some platforms,
so don't do it on those platforms. Tracked down by Wim Lewis, and
reported here: <http://bugzilla.gnome.org/show_bug.cgi?id=577137>.
2009-03-24 Torsten Schoenfeld <kaffeetisch@gmx.de>
* Glib.exports: Add gperl_option_context_get_type and
gperl_option_group_get_type to the export list so that dependent
modules link correctly on MSWin32. Patch by Martin
Schlemmer. (GNOME #576372)
2009-03-24 Torsten Schoenfeld <kaffeetisch@gmx.de>
* t/h.t: Correct a skip count. Patch by Stefan B. (RT #44428)
2009-03-19 Torsten Schoenfeld <kaffeetisch@gmx.de>
* Makefile.PL: Space-separate the libs and cflags strings from
gobject and gthread to avoid unwanted concatenation. Patch by
Daniel Macks. (RT #44055)
2009-03-19 Torsten Schoenfeld <kaffeetisch@gmx.de>
* MakeHelper.pm (postamble_docs_full): Use qq// instead of qq{} to
quote the copyright string in order to avoid problems with dmake.
Patch by Martin Schlemmer. (GNOME #573099)
2009-03-17 Torsten Schoenfeld <kaffeetisch@gmx.de>
* Glib.pm
* Makefile.PL
* NEWS
* README: Stable release 1.220.
2009-03-08 Torsten Schoenfeld <kaffeetisch@gmx.de>
* Glib.pm
* NEWS
* README: Unstable release 1.214.
2009-03-08 Torsten Schoenfeld <kaffeetisch@gmx.de>
* GClosure.xs (gperl_callback_invoke): Fix a giant leak by
properly unsetting all GValues used for the callback arguments.
Patch by Kevin Ryde.
2009-03-08 Torsten Schoenfeld <kaffeetisch@gmx.de>
* GenPod.pm (xsub_to_pod): Don't wrap the argument types of xsubs
with a superfluous pair of '=over' and '=back'.
2009-03-08 Torsten Schoenfeld <kaffeetisch@gmx.de>
* GenPod.pm (podify_enums_and_flags): Don't generate 'ENUMS AND
FLAGS' entries for Glib::Enum and Glib::Flags since they don't
have values anyway.
2009-03-01 Torsten Schoenfeld <kaffeetisch@gmx.de>
* t/7.t
* GSignal.xs (gperl_signal_connect): Don't leak the GClosure
object when an invalid signal name is passed in. Patch by
Kevin Ryde.
2009-03-01 Torsten Schoenfeld <kaffeetisch@gmx.de>
* GType.xs (gperl_fundamental_type_from_obj): Check the SV for
definedness before passing it to SvRV. This fixes a test failure
on MSWin32 in t/c.t.
2009-03-01 Torsten Schoenfeld <kaffeetisch@gmx.de>
* t/filename.t: Add more tests for Glib::filename_to_uri. Patch
by Kevin Ryde.
2009-02-22 Torsten Schoenfeld <kaffeetisch@gmx.de>
* GBoxed.xs
* GType.xs: Fix errors reported by podchecker.
2009-02-21 Torsten Schoenfeld <kaffeetisch@gmx.de>
* Makefile.PL: Tell the CPAN indexer to ignore the package 'MY'
used by Glib::MakeHelper.
2009-02-21 Torsten Schoenfeld <kaffeetisch@gmx.de>
* GenPod.pm (xsdoc2pod): Instead of relying on one final '=cut',
close every opened POD paragraph separately. This fixes POD
errors that occurred when '=for position COPYRIGHT' was used.
Original patch by Ryan Niebur.
2009-02-18 Torsten Schoenfeld <kaffeetisch@gmx.de>
* GBookmarkFile.xs: Don't document Glib::BookmarkFile::set_added,
set_modified, and set_visited as functions; they're methods.
* t/h.t: Test Glib::BookmarkFile::set_added, set_modified, and
set_visited and their associated getters.
2009-02-18 Torsten Schoenfeld <kaffeetisch@gmx.de>
* Glib.pm: Use an '=encoding utf8' directive for the AUTHORS POD
section so that non-ASCII characters are handled correctly.
2009-02-13 Torsten Schoenfeld <kaffeetisch@gmx.de>
* Glib.pm
* NEWS
* README: Unstable release 1.213.
2009-02-11 Torsten Schoenfeld <kaffeetisch@gmx.de>
* GError.xs: Document that Glib::Error->location returns raw
bytes. Patch by Kevin Ryde.
2009-02-11 Torsten Schoenfeld <kaffeetisch@gmx.de>
* Glib.pm: In the documentation for filename_from_uri, hint at
path separator issue. Patch by Kevin Ryde.
2009-02-11 Torsten Schoenfeld <kaffeetisch@gmx.de>
* GOption.xs (copy_string): Fix C99-ism. Patch by Kevin Ryde.
2009-02-06 Torsten Schoenfeld <kaffeetisch@gmx.de>
* t/filename.t: Test Glib->filename_to_uri with hostname == undef.
2009-02-05 Torsten Schoenfeld <kaffeetisch@gmx.de>
* MakeHelper.pm: Deprecate
Glib::MakeHelper->get_configure_requires_yaml in favor of EU::MM's
new META_MERGE and META_ADD features.
2009-02-05 Torsten Schoenfeld <kaffeetisch@gmx.de>
* Makefile.PL: Unconditionally use META_MERGE; so you now need
ExtUtils::MakeMaker >= 6.46 to roll a Glib tarball with a proper
META.yml. Also, tell the CPAN indexer (via META.yml) to not index
xsapi.pod.head and xsapi.pod.foot.
2009-02-05 Torsten Schoenfeld <kaffeetisch@gmx.de>
* GClosure.xs (gperl_closure_marshal): Synchronize the way signals
with no return value are handled with how it's done in
gperl_signal_class_closure_marshal. This means that Perl handlers
for these kinds of signals are now always called in void context,
as they should be.
2009-02-05 Torsten Schoenfeld <kaffeetisch@gmx.de>
* MANIFEST
* GType.xs (gperl_signal_class_closure_marshal)
* t/signal_marshal.t: Correctly handle signals with no return type
when invoking signal class closures. Patch by Kevin Ryde.
2009-02-05 Torsten Schoenfeld <kaffeetisch@gmx.de>
* GType.xs (gperl_register_fundamental_full): Handle
wrapper_class_by_type being NULL gracefully. Patch by Kevin Ryde.
2009-02-05 Torsten Schoenfeld <kaffeetisch@gmx.de>
* Glib.xs: In Glib::filename_from_uri and filename_to_uri, use the
gchar converters for the hostname. Patch by Kevin Ryde.
2009-02-05 Torsten Schoenfeld <kaffeetisch@gmx.de>
* GClosure.xs (gperl_callback_invoke): Use the <"%s", pointer>
pattern when passing an error string pointer to croak().
2009-02-05 Torsten Schoenfeld <kaffeetisch@gmx.de>
* GType.xs
* t/c.t: Make the various Glib::Flags methods more robust with
respect to receiving undefined input. Patch by Kevin Ryde.
2009-02-05 Torsten Schoenfeld <kaffeetisch@gmx.de>
* GBoxed.xs (gperl_register_boxed)
* GObject.xs (gperl_register_object)
* GType.xs (gperl_register_fundamental): When inserting into the
package → type hash tables, use g_hash_table_replace instead of
_insert so that the new string pointer is used as the key. Patch
by Kevin Ryde.
2009-02-05 Torsten Schoenfeld <kaffeetisch@gmx.de>
* Glib.pm
* NEWS
* README: Unstable release 1.212.
2009-02-03 Torsten Schoenfeld <kaffeetisch@gmx.de>
* t/a.t
* GLog.xs: Make sure that messages with % chars in them make it
through Glib::log() and friends safely.
2009-02-02 Torsten Schoenfeld <kaffeetisch@gmx.de>
* ParseXSDoc.pm (xsdocparse): Always use Data::Dumper on MSWin32
for serialization. Storable seems to consistently fail there,
stating 'Magic number checking on storable file failed'.
2009-02-02 Torsten Schoenfeld <kaffeetisch@gmx.de>
* Glib.xs
* GSignal.xs: Hush some compiler warnings.
2009-01-20 Torsten Schoenfeld <kaffeetisch@gmx.de>
* Glib.pm: Add documentation for Glib->filename_to_uri and
filename_from_uri. Patch by Kevin Ryde.
2009-01-18 Torsten Schoenfeld <kaffeetisch@gmx.de>
* GParamSpec.xs
* t/e.t: Add Glib::ParamSpec->value_cmp. Patch by Kevin Ryde.
2009-01-18 Torsten Schoenfeld <kaffeetisch@gmx.de>
* t/e.t: Use UTF-8 instead of ISO-8859-1 to encode umlauts.
2009-01-18 Torsten Schoenfeld <kaffeetisch@gmx.de>
* GParamSpec.xs
* t/e.t: Add Glib::ParamSpec->value_validate. Patch by Kevin
Ryde.
2009-01-18 Torsten Schoenfeld <kaffeetisch@gmx.de>
* t/options.t: Don't use UTF-8 umlauts unconditionally as this
causes trouble on non-UTF-8 locales.
2009-01-18 Torsten Schoenfeld <kaffeetisch@gmx.de>
* MakeHelper.pm: Add
Glib::MakeHelper->postamble_precompiled_headers to generate a make
target for precompiling headers.
* Makefile.PL: Use the above to create a 'precompiled-headers'
target that precompiles gperl.h. This speeds up the compilation
of Glib by 16% on my machine.
2008-12-30 Torsten Schoenfeld <kaffeetisch@gmx.de>
* GType.xs
* t/c.t: Allow calling Glib::Flags::as_arrayref directly, as an
alternative to the @{} syntax. Patch by Kevin Ryde.
2008-11-24 Torsten Schoenfeld <kaffeetisch@gmx.de>
* Glib.pm
* NEWS
* README: Unstable release 1.211.
2008-11-22 Torsten Schoenfeld <kaffeetisch@gmx.de>
* Glib.xs
* GOption.xs (added)
* gperl.h
* Makefile.PL
* MANIFEST
* t/options.t (added)
* typemap: Add Glib::OptionContext and Glib::OptionGroup, wrapping
glib's command line option parser.
2008-11-22 Torsten Schoenfeld <kaffeetisch@gmx.de>
* gperl-private.h: Don't include perl.h. This header is always
included after gperl.h which includes everything already. And
including perl.h alone doesn't work and isn't enough anyway.
2008-11-15 Torsten Schoenfeld <kaffeetisch@gmx.de>
* GBoxed.xs (gperl_register_boxed): Use the copy of the passed-in
package name for the key of info_by_package to avoid problems when
the package name is dynamically allocated.
* GBoxed.xs (gperl_register_boxed)
* GObject.xs (gperl_register_object)
* GType.xs (gperl_register_fundamental): Always insert into the
lookup tables in the correct order to avoid prematurely freeing
the package name.
2008-11-12 Torsten Schoenfeld <kaffeetisch@gmx.de>
* GObject.xs: Add newlines at the end of debugging messages that
lack them to keep perl from adding line number information.
2008-11-03 Torsten Schoenfeld <kaffeetisch@gmx.de>
* GType.xs (gperl_type_from_package): Also try
gperl_param_spec_type_from_package on the package.
2008-11-03 Torsten Schoenfeld <kaffeetisch@gmx.de>
* Glib.pm
* Makefile.PL
* NEWS
* README: Unstable release 1.210.
2008-11-03 Torsten Schoenfeld <kaffeetisch@gmx.de>
* GType.xs (add_interfaces): Fail loudly and with a clear error
message when we encounter an unknown interface when registering a
new type.
2008-11-03 Torsten Schoenfeld <kaffeetisch@gmx.de>
* t/lazy-loader.t -> t/lazy_loader.t: Rename this test to blend in
with the others which use underscores to separate words.
2008-11-03 Torsten Schoenfeld <kaffeetisch@gmx.de>
* Glib.exports
* gperl.h: Properly export gperl_param_spec_type_from_package and
gperl_register_param_spec, they are meant to be public.
2008-11-02 Torsten Schoenfeld <kaffeetisch@gmx.de>
* GType.xs: Use gperl_register_fundamental_alias to register the
"Glib::Uint" backwards-compatibility alias.
2008-11-02 Torsten Schoenfeld <kaffeetisch@gmx.de>
* GBoxed.xs
* GObject.xs
* GType.xs
* Glib.exports
* gperl.h: Provide gperl_register_boxed_alias,
gperl_register_fundamental_alias, and gperl_register_object_alias
to register aliases for other registered types. An alias is a
package name that will resolve to the specified type, while the
type will still resolve to its originally registered package when
going the other way.
2008-11-01 Torsten Schoenfeld <kaffeetisch@gmx.de>
* t/9.t
* GMainLoop.xs: Wrap g_child_watch_add as Glib::Child::watch_add.
Patch by Kevin Ryde.
2008-11-01 Torsten Schoenfeld <kaffeetisch@gmx.de>
* GMainLoop.xs
* Glib.pm: Add constants Glib::SOURCE_CONTINUE and SOURCE_REMOVE
for use in source-type callbacks. Patch by Kevin Ryde.
2008-10-19 Torsten Schoenfeld <kaffeetisch@gmx.de>
* Makefile.PL: ExtUtils::MakeMaker 6.46 removed support for
EXTRA_META in favor of the new META_MERGE and META_ADD. Adapt.
2008-09-20 Torsten Schoenfeld <kaffeetisch@gmx.de>
* Glib.pm
* Makefile.PL
* NEWS
* README: Stable release 1.200.
2008-09-20 Torsten Schoenfeld <kaffeetisch@gmx.de>
* TODO: Remove the entry about hushing about documentation
generation. That's done.
2008-09-07 Torsten Schoenfeld <kaffeetisch@gmx.de>
* Glib.pm
* NEWS
* README: Unstable release 1.193.
2008-09-07 Torsten Schoenfeld <kaffeetisch@gmx.de>
* GClosure.xs (gperl_callback_invoke): Instead of putting a mortal
copy of the user data on the stack, increment its ref count and
mortalize it. This ensure that modifications to $_[-1] in the
callback actually change the user data. Patch by Kevin Ryde.
* MANIFEST
* GObject.xs
* t/lazy-loader.t: Make sure the lazy loader can handle being
invoked on packages which aren't registered with the Glib type
system. This can happen when a non-registered package is setup to
inherit from a registered package, and when then some method is
invoked on the non-registered package. This gets rid of the
spurious "asked to lazy-load Foo, but that package is not
registered" warnings.
2008-09-06 Torsten Schoenfeld <kaffeetisch@gmx.de>
* GParamSpec.xs:
* GValue.xs: Handle NULL GParamSpecs in newSVGParamSpec by
returning undef. Consequently, remove the NULL handling in
_gperl_sv_from_value_internal. This partly reverts the commit
from 2008-08-18 and brings newSVGParamSpec in line with the other
SV* constructors.
* GObject.xs
* t/5.t: Fix the stack handling in Glib::Object::get to be robust
against stack movement due to reallocation. This can happen if a
subclass' GET_PROPERTY triggers a stack resize. Patch by Kevin
Ryde.
2008-08-31 Torsten Schoenfeld <kaffeetisch@gmx.de>
* Glib.pm
* NEWS
* README: Unstable release 1.192.
2008-08-31 Torsten Schoenfeld <kaffeetisch@gmx.de>
* MakeHelper.pm (postamble_docs_full): Fix the POD index
generation. For five months now, broken index pages (like
Glib::index) were being created due to shell command quoting
issues.
2008-08-23 Torsten Schoenfeld <kaffeetisch@gmx.de>
* GType.xs (gperl_convert_flag_one, gperl_convert_flags): Improve
the wording of the error messages for invalid flags a bit.
* t/4.t: Prettify a bit by using ok(), pass(), and fail()
functions.
2008-08-18 Torsten Schoenfeld <kaffeetisch@gmx.de>
* Glib.pm
* NEWS
* README: Unstable release 1.191.
2008-08-18 Torsten Schoenfeld <kaffeetisch@gmx.de>
* t/e.t
* GValue.xs (_gperl_sv_from_value_internal): Handle NULL
GParamSpecs gracefully by returning undef.
* GParamSpec.xs (newSVGParamSpec): Croak on NULL GParamSpecs.
2008-08-17 Torsten Schoenfeld <kaffeetisch@gmx.de>
* GParamSpec.xs: Add hierarchy POD sections to the various
Glib::ParamSpec subclasses.
2008-08-16 Torsten Schoenfeld <kaffeetisch@gmx.de>
* GenPod.pm: Document how to hide arguments with "=for arg". Cope
with _noinc_ornull type variants.
2008-08-03 Torsten Schoenfeld <kaffeetisch@gmx.de>
* GObject.xs (g_object_new): Check that we got a correct number of
arguments.
* t/2.t: Fix a typo uncovered by the above change.
2008-07-13 Torsten Schoenfeld <kaffeetisch@gmx.de>
* GType.xs
* t/c.t: In the overloaded flags operators, don't choke on undef
in the swap argument. This happens when a normal 'x' operator is
used for handling 'x='. Patch by Kevin Ryde.
2008-06-22 Torsten Schoenfeld <kaffeetisch@gmx.de>
* Glib.pm
* Makefile.PL
* NEWS
* README: Unstable release 1.190.
2008-06-13 Torsten Schoenfeld <kaffeetisch@gmx.de>
* GObject.xs (gperl_get_object_check): Croak on objects that don't
carry magic.
2008-06-01 Torsten Schoenfeld <kaffeetisch@gmx.de>
* GType.xs: Fix typo in error message output by Glib::Flags::new.
* GUtils.xs
* t/1.t: Provide Glib::strerror and Glib::strsignal. Patch by
Kevin Ryde.
2008-05-31 Torsten Schoenfeld <kaffeetisch@gmx.de>
* GSignal.xs: In the xsub for g_signal_add_emission_hook, make
sure that the type class exists before we try to fetch information
about one of its signals.
2008-05-23 Torsten Schoenfeld <kaffeetisch@gmx.de>
* GType.xs
* t/c.t: Add Glib::Flags::new, a constructor for flags objects
that can be used with the overloaded operators. Patch by Kevin
Ryde and muppet.
2008-05-22 Torsten Schoenfeld <kaffeetisch@gmx.de>
* GType.xs
* Glib.pm
* t/c.t: Overload '!=' and 'ne' for flags values for consistency.
2008-05-20 Torsten Schoenfeld <kaffeetisch@gmx.de>
* GClosure.xs (gperl_callback_invoke): Put a mortal copy of the
user data on the stack to avoid prematurely destroying it in
certain cases. Patch by Kevin Ryde.
* t/c.t: Test that empty flags values ([], undef) work.
* GType.xs (gperl_convert_flag_one): Don't call
gperl_type_flags_get_values needlessly.
2008-05-04 Torsten Schoenfeld <kaffeetisch@gmx.de>
* GBoxed.xs
* GType.xs: Use const char* to store the return value of
sv_reftype.
* Subclass.pm: Improve the documentation of GET_PROPERTY and
SET_PROPERTY. (Patch by Kevin Ryde)
2008-04-19 muppet <scott@asofyet.org>
* xs/GType.xs
* t/c.t: Register Glib::Enum and Glib::Flags. Remove special case
logic for these from Glib::Type::register().
2008-03-30 Torsten Schoenfeld <kaffeetisch@gmx.de>
* Glib.pm
* Makefile.PL
* NEWS
* README: Stable release 1.182, requiring ExtUtils::Depends >=
0.300.
2008-03-30 Torsten Schoenfeld <kaffeetisch@gmx.de>
* GType.xs: For the 64 bit integer converters: use
g_ascii_strtoll, g_ascii_strtoull, and _atoi64 when available.
Only use _strtoi64 and _strtoui64 if _MSC_VER >= 1300.
* GenPod.pm: Add setters set_copyright, set_year, set_authors, and
set_main_mod for the corresponding package variables.
* MakeHelper.pm (postamble_docs_full): Use more portable ways to
quote commands and strings, partly by using the new functions
mentioned above.
(postamble_rpms): Don't do anything unless $ENV{GPERL_BUILD_RPMS}
is set.
(quiet_rule): Use two lines for the prettified commands so that
long commands don't get even longer.
(do_pod_files): Prettify information message.
* t/1.t
* t/filename.t
* t/g.t: Fix failures on MSWin32.
* ParseXSDoc.pm (clean_out_empty_pods): Stop being chatty.
2008-03-22 Torsten Schoenfeld <kaffeetisch@gmx.de>
* Makefile.PL: Trap exceptions thrown by EU::PkgConfig->find.
When one occurs, exit with status 0 to indicate to CPAN testers
that their system isn't supported.
2008-03-14 Torsten Schoenfeld <kaffeetisch@gmx.de>
* Glib.pm
* NEWS
* README: Stable release 1.181.
2008-03-14 Torsten Schoenfeld <kaffeetisch@gmx.de>
* Makefile.PL: Remove the no_index specification from EXTRA_META
because the YAML implementations currently in use don't seem to
support this syntax yet.
* Makefile.PL (MY::postamble): Handle the EU::Depends object $glib
being undefined gracefully.
* Makefile.PL: Don't warn about this release being unstable
because it isn't.
* copyright.pod: Update for 2008.
2008-03-10 Torsten Schoenfeld <kaffeetisch@gmx.de>
* Glib.pm
* NEWS
* README: Stable release 1.180.
2008-02-24 Torsten Schoenfeld <kaffeetisch@gmx.de>
* Makefile.PL: Tell the compiler to always look for headers in '.'
first so that ours are preferred in case of name
clashes. (Suggestion by Daniel Macks in RT #32883)
* MakeHelper.pm: Add a hyphen to the NAME section of POD
indices. (Suggested by Quentin Sculo)
2008-02-24 Torsten Schoenfeld <kaffeetisch@gmx.de>
* Glib.pm
* NEWS
* README: Unstable release 1.174.
2008-02-24 Torsten Schoenfeld <kaffeetisch@gmx.de>
* MakeHelper.pm: Rework the BLIB_DONE stuff to not rely on
Makefile conditionals. We should now build fine with any `make´
again. Also fix some incorrect and add missing dependencies for
various documentation files.
2008-02-12 Torsten Schoenfeld <kaffeetisch@gmx.de>
* Glib.pm
* NEWS
* README: Unstable release 1.173.
2008-01-17 Torsten Schoenfeld <kaffeetisch@gmx.de>
* GType.xs
* t/c.t: Make Glib::Type::list_values return the value of each
enum/flags entry in addition to the name and nickname. (Patch by
Kevin Ryde)
* t/9.t
* t/filename.t: Hopefully fix test failures on Windows.
2008-01-16 Torsten Schoenfeld <kaffeetisch@gmx.de>
* t/4.t: Test Glib::Type->list_ancestors and
Glib::Type->package_from_cname.
2008-01-09 Torsten Schoenfeld <kaffeetisch@gmx.de>
* Glib.pm
* NEWS
* README: Unstable release 1.172.
2008-01-09 Torsten Schoenfeld <kaffeetisch@gmx.de>
* ParseXSDoc.pm (parse_file): When parsing an #else pre-processor
directive, make sure we have actually seen and parsed a
corresponding #if before trying to negate its entry. This can
happen when our parsing heurestics fail. The result was an
exception: "Modification of non-creatable array value attempted."
2008-01-09 muppet <scott@asofyet.org>
* GType.xs: Defer the instantiation of a newly-registered
object class's class until after we've parsed the list of
signals, properties, and interfaces to add to it. This
allows perl-derived GObjects to override GInterfaces that
are implemented by parent classes. Previously, the early
instantiation confused GLib and caused the call to
g_type_class_add_interface() to complain that the given type
already conforms to the interface, because the type system
considered the initialization of the class to be complete.
Note that we do not currently have a way of chaining up to
a parent's GInterface implementation, as can be done in C
via g_type_interface_peek_parent(). Black magic may be
required for that.
2008-01-07 muppet <scott@asofyet.org>
* GObject.xs: Load GInterface types immediately, instead of
leaving them for lazy loading. Otherwise, it is possible to
get into situations in which a GInterface has not triggered
its lazy loading by the time it is needed.
2008-01-07 Torsten Schoenfeld <kaffeetisch@gmx.de>
* CodeGen.pm
* GBoxed.xs
* GError.xs
* GObject.xs
* GSignal.xs
* GType.xs
* GValue.xs
* Glib.exports
* Glib.xs
* gperl.h
* typemap: Rename gperl_sv_defined to gperl_sv_is_defined.
* GBoxed.xs
* GError.xs
* GType.xs
* gperl.h: Add new macros: gperl_sv_is_array_ref,
gperl_sv_is_code_ref, and gperl_sv_is_hash_ref.
* GObject.xs (class_info_finish_loading): When handling a class'
@ISA, instead of simple av_fetch calls and a final av_clear, use
repeated av_shift calls. Invoking av_clear on an @ISA array seems
to break the caching magic associated with it when running under
perl 5.10.0. This in turn breaks our lazy-loading scheme.
2008-01-07 muppet <scott@asofyet.org>
* GSignal.xs, t/signal_emission_hooks.t: Bind and test
g_signal_get_invocation_hint().
2007-12-30 Torsten Schoenfeld <kaffeetisch@gmx.de>
* gperl_marshal.h: Mention in the POD that
dGPERL_CALLBACK_MARSHAL_SP should always come last in a list of
declarations.
2007-12-22 Torsten Schoenfeld <kaffeetisch@gmx.de>
* gperl.h (SvGObject)
* t/2.t: Use gperl_get_object_check instead of gperl_get_object so
that inproper usage of Glib::Object methods results in an error
message and not in a segfault.
* CodeGen.pm (gen_boxed_stuff): Correctly cast the return value of
the Sv$classname macros for boxed types so as to avoid handing out
void* pointers.
2007-12-16 Torsten Schoenfeld <kaffeetisch@gmx.de>
* Glib.pm
* NEWS
* README: Unstable release 1.171.
2007-12-16 Torsten Schoenfeld <kaffeetisch@gmx.de>
* GBookmarkFile.xs: Hush compiler warning.
* CodeGen.pm
* gperl.h
* typemap: Replace the last occurences of SvOK with
gperl_sv_defined.
2007-11-20 Torsten Schoenfeld <kaffeetisch@gmx.de>
* Glib.exports
* Glib.xs
* gperl.h: Add gperl_sv_defined, which checks an SV for
definedness just like Perl's defined() does.
* GBoxed.xs
* GError.xs
* GObject.xs
* GSignal.xs
* GType.xs
* GValue.xs
* Glib.xs: Use gperl_sv_defined instead of SvOK to check SVs for
definedness. This keeps Glib from erroneously thinking your
scalars are undefined, for example when using tied hashes.
* MANIFEST
* t/tied_definedness.t: Test that values from tied hashes get
through the definedness checks.
2007-11-18 Torsten Schoenfeld <kaffeetisch@gmx.de>
* GError.xs (gperl_gerror_from_sv): Fix the check for undefined
domain keys in GError hashes.
* GType.xs (add_interfaces): Fix the check for undefined interface
names.
2007-11-15 Torsten Schoenfeld <kaffeetisch@gmx.de>
* Glib.exports: Add SvGKeyFile. (Reported by spicy jack)
2007-10-28 Torsten Schoenfeld <kaffeetisch@gmx.de>
* Glib.pm
* Makefile.PL
* NEWS
* README: Unstable release 1.170.
2007-10-28 Torsten Schoenfeld <kaffeetisch@gmx.de>
* GType.xs: Define and use PORTABLE_LL_FORMAT and
PORTABLE_ULL_FORMAT.
2007-10-18 Torsten Schoenfeld <kaffeetisch@gmx.de>
* AUTHORS
* MakeHelper.pm (postamble_docs_full): When constructing the
add_types call, make sure to properly quote the paths. (Patch by
Andres König)
2007-10-14 Torsten Schoenfeld <kaffeetisch@gmx.de>
* Makefile.PL: Use BEGIN { require 5.008; } instead of use 5.008;
so automatic CPAN testers know which perl version we need. Also
get rid of the MY package by prefixing the relevant subs with
`MY::´.
2007-10-08 Torsten Schoenfeld <kaffeetisch@gmx.de>
* GObject.xs (Glib::Object::CLONE): Check that perl_gobjects !=
NULL before working with it, to avoid an assertion.
2007-10-07 Torsten Schoenfeld <kaffeetisch@gmx.de>
* MakeHelper.pm
* t/make_helper.t: Add
Glib::MakeHelper::get_configure_requires_yaml to generate YAML
suitable for use with EU::MM's EXTRA_META.
* Makefile.PL: Use the above to tell EU::MM to generate a correct
configure_requires entry. Also output a no_index entry to tell
the CPAN indexer what to ignore.
2007-09-24 kaffeetisch
* CodeGen.pm
* GenPod.pm
* MakeHelper.pm
* Makefile.PL: Prettify build output by summarizing what is done
instead of printing executed command lines, and by removing
unnecessarily verbose output. Use `make NOECHO=" "´ if you want
to see the command lines.
* MakeHelper.pm: Implement custom Makefile rule generators in MY
that print stuff like [ XS File.xs ] instead of the overly long
command lines. Again, use `make NOECHO=" "´ if you want to see
the command lines.
2007-09-16 kaffeetisch
* Glib.pm
* Makefile.PL
* NEWS
* README: Stable release 1.160.
2007-09-15 kaffeetisch
* GKeyFile.xs
* GMainLoop.xs
* GUtils.xs
* gperl.h
* t/1.t
* t/9.t
* t/g.t: Change version checks to refer to glib 2.14 instead of
2.13.
2007-08-13 kaffeetisch
* Glib.pm
* NEWS
* README: Unstable release 1.153.
2007-08-13 kaffeetisch
* GMainLoop.xs
* GObject.xs
* GType.xs
* GenPod.pm: Output a warning if a type name isn't recognized by
the doc parser. Add 'package' and 'list' to the basic types and
fix the remaining type name issues the new warning uncovered.
* GenPod.pm
* ParseXSDoc.pm: Parse preprocessor conditionals in the XS code to
add Since: tags to methods' POD.
* GenPod.pm
* ParseXSDoc.pm: Make it possible to document specific packages in
different POD files by providing a new =for object variant.
* t/1.t
* t/9.t
* t/filename.t: Fix win32 test failures.
2007-08-01 kaffeetisch
* gperl.h
* typemap: Add char_own_ornull and gchar_own_ornull typemaps.
2007-07-08 kaffeetisch
* Glib.pm
* NEWS
* README: Unstable release 1.152.
2007-07-07 kaffeetisch
* GObject.xs: Register G_TYPE_INTERFACES as Glib::Interfaces to
avoid having it appear under Glib::Object::_Unregistered in
hierarchies in the generated POD.
* GMainLoop.xs
* t/9.t: Wrap g_main_context_is_owner and g_timeout_add_seconds.
2007-06-24 kaffeetisch
* Glib.pm, NEWS, README: Unstable release 1.151.
2007-06-23 kaffeetisch
* GUtils.xs
* gperl.h
* typemap
* t/1.t: Wrap g_get_user_special_dir and provide converters for
the new enum GUserDirectory.
* t/g.t: In floating point tests, don't check for equality.
Instead, check that numbers don't differ too much from their
expected values.
2007-06-10 kaffeetisch
* GKeyFile.xs, t/g.t: Wrap g_key_file_load_from_dirs.
2007-02-24 kaffeetisch
* MakeHelper.pm: Use BSD make syntax for the BLIB_DONE stuff on
BSD systems, unless the environment variable FORCE_GMAKE is set.
2007-01-08 kaffeetisch
* Glib.pm, Makefile.PL, NEWS, README, copyright.pod: Unstable
release 1.150.
2007-01-07 kaffeetisch
* GenPod.pm: Only mark those objects as deprecated that have a
non-empty list of replacements; this avoids marking each and every
object as deprecated. Move up the deprecation message in the POD
to make it more prominent. Also, hush the "podifying Foo"
messages.
2007-01-06 Emmanuele Bassi <ebassi@gnome.org>
* ParseXSDoc.pm:
* GenPod.pm: Add a "deprecated_by" key to the apidoc parser;
you should use this to mark an object deprecated in favour of
another. Also add a __deprecated__ marker for functions and
methods.
2006-12-30 kaffeetisch
* Glib.xs: Call g_threads_init before g_type_init. Required (and
probably enforced in the near future) by recent glib versions.
2006-12-02 kaffeetisch
* NEWS: Merge from stable-1-14.
* Makefile.PL: Check that glib >= 2.12.0, not 2.11.0, before
including GBookmarkFile.xs.
2006-11-23 kaffeetisch
* GType.xs, gperl.h: Revert the G_OS_WIN32 change and just use
WIN32 instead. G_OS_WIN32 doesn't seem to be defined on most
win32 boxes.
* AUTHORS, GType.xs: Use _strtoi64 and and _strtoui64 on win32 if
compiling with MSVC; patch from Serguei Trouchelle.
2006-11-05 kaffeetisch
* GType.xs, gperl.h: Use G_OS_WIN32 instead of WIN32 or _WIN32.
2006-10-21 kaffeetisch
* GBookmarkFile.xs: Remove a custom "=for signature" directive
that incorrectly specified a boolean return value for remove_item.
2006-10-17 kaffeetisch
* Glib.xs, MANIFEST, ppport.h: Get rid of ppport.h, we don't seem
to need it.
2006-10-03 kaffeetisch
* CodeGen.pm, GenPod.pm, ParseXSDoc.pm: Fix POD errors. (Mainly
=cut's not followed by a blank line.)
2006-09-13 kaffeetisch
* GBoxed.xs (gperl_new_boxed_copy): Check for NULL before passing
the boxed pointer to g_boxed_copy.
* GType.xs: Add macros, PORTABLE_STRTOLL and PORTABLE_STRTOULL, to
centralize the preprocessor madness for 64bit integer support. On
win32, which apparently lacks strtoll or similar, use strtol and
strtoul for now to make it at least compile.
* t/7.t: Apply a patch from Michael G Schwern that introduces
helper functions which make this complicated test easier to read.
* t/filename.t: Remove two debug prints that were causing perl
warnings.
2006-09-09 kaffeetisch
* t/4.t, t/5.t, t/6.t, t/7.t, t/9.t: In hand-rolled test output,
use '-' instead of '#' as the comment designator. In TAP (Test
Anything Protocol), the protocol we use, '#' is for directives and
'-' is for comments.
2006-09-04 kaffeetisch
* Glib.pm, Makefile.PL, NEWS, README: Stable release 1.140.
2006-09-04 kaffeetisch
* t/9.t: Don't test Glib::main_depth on glib <= 2.4.0, where it
first appeared.
* t/g.t: Fix skip count.
2006-08-29 kaffeetisch
* README: Add a bug reports section.
2006-08-07 kaffeetisch
* Glib.pm, NEWS, README: Unstable release 1.132.
2006-08-07 kaffeetisch
* GKeyFile.xs, Glib.xs: Remove outdated FIXME comments.
* GMainLoop.xs (async_watcher_install): Add FIXME comment about a
leak.
* GObject.xs: Remove FIXME comment about gperl_object_check_type's
semantics. We can't do anything about anymore.
* GSignal.xs (foreach_closure_matched): Fix to use guint instead
of int for the return type.
* Glib.pm: Don't use indirect object syntax.
* TODO: Update.
* typemap: Add custom typemaps for guchar* to avoid compiler
warnings.
* t/g.t: Change version checks to refer to stable releases only.
2006-07-23 kaffeetisch
* Glib.pm, Makefile.PL, NEWS, README: Unstable release 1.131.
2006-07-23 kaffeetisch
* GError.xs, GKeyFile.xs, Glib.xs, gperl-gtypes.h, gperl-gtypes.c,
gperl.h: Change version checks to check for 2.12 instead of 2.11.
2006-07-19 Emmanuele Bassi <ebassi@gmail.com>
* GObject.xs: Bind g_object_class_find_property()...
* t/4.t: ... and test it.
2006-07-14 kaffeetisch
* Glib.pm, NEWS, README, copyright.pod: Unstable release 1.130.
2006-07-14 kaffeetisch
* MakeHelper.pm: Turn a few 'use Config's into 'require Config's.
Also remove some loud MSVC/nmake bashing. :-)
2006-06-20 Emmanuele Bassi <ebassi@gmail.com>
A GBookmarkFile.xs
A t/h.t
* GBookmarkFile.xs: Bind GBookmarkFile, a parser for desktop
bookmarks, that comes with GLib 2.11/2.12.
* t/h.t: Test suite for Glib::BookmarkFile.
* GError.xs:
* Glib.xs:
* MANIFEST:
* Makefile.PL:
* doctypes:
* gperl-gtypes.[hc]:
* gperl.h:
* typemap: Build glue for Glib::BookmarkFile.
* Makefile.PL:
* Glib.pm: Update $VERSION to 1.130 and warn the user.
2006/06/19 kaffeetisch
* GKeyFile.xs, t/g.t: Bind and test g_key_file_set_double,
g_key_file_get_double, g_key_file_set_double_list, and
g_key_file_get_double_list. Fix a few API doc glitches.
2006/06/19 kaffeetisch
* GObject.xs: Register GInitiallyUnowned as
Glib::InitiallyUnowned.
2006/05/21 kaffeetisch
* GObject.xs (class_info_finish_loading): Free the list returned
by g_type_interfaces when we're finished with it.
2006/03/12 kaffeetisch
* Glib.pm, Makefile.PL, NEWS, README: Stable release 1.120.
2006/03/06 kaffeetisch
* Glib.pm, NEWS, README: Unstable release 1.118.
2006/03/04 kaffeetisch
* gperl-private.h, MANIFEST: Add a new header file for private
functions that are used in more than one xs file. For now these
are _gperl_sv_from_value_internal, _gperl_fetch_wrapper_key, and
some new threading-safety stuff described below.
* Glib.xs, gperl-private.h: Add _gperl_set_master_interp and
_gperl_get_master_interp to get and set a global master
interpreter that can be used to setup perl's thread-local storage.
For convenience, the new macro GPERL_SET_CONTEXT encapsulates this
setup step. Inspired by Dobrica Pavlinusic's approach in the Fuse
module.
* Glib.xs: Call _gperl_set_master_interp in BOOT to set the global
master interpreter to the value of PERL_GET_INTERP.
* GObject.xs (gobject_destroy_wrapper): Use GPERL_SET_CONTEXT to
prevent a segfault that occured when this function was called from
a different thread. Apparently, PL_in_clean_objs tries to access
invalid memory if you don't setup perl's context correctly.
* GLog.xs (gperl_log_handler): Use GPERL_SET_CONTEXT to replace
the hack that put the value of PERL_GET_CONTEXT into the user
data. The result should be the same: correctly setup perl's
context before calling warn() to avoid segfaults when called from
a different thread.
* GType.xs, GObject.xs: Include gperl-private.h to get rid of the
"extern" declarations.
2006/02/27 kaffeetisch
* Glib.pm, Makefile.PL, NEWS, README: Unstable release 1.117.
2006/02/14 kaffeetisch
* GType.xs: Add a missing semicolon that caused compilation
failures on perl 5.8.8.
2006/02/13 kaffeetisch
* Glib.pm, Makefile.PL, NEWS, README: Unstable release 1.116.
2006/02/13 kaffeetisch
* MANIFEST, README.api-changes: Remove this outdated file.
2006/02/11 muppetman
* GObject.xs: SvLEN() is the length of the SV's buffer; SvCUR() is
the length of the SV's string. Use SvCUR() instead of SvLEN()-1 in
_gperl_fetch_wrapper_key(), or we get garbage key names with perl
5.8.8, resulting in failures in t/f.t related to properties with no
getter or setter.
* ParseXSDoc.pm: Data::Dumper uses a lot of memory, and causes the
docgen for Gtk2 to use up to 10MB. Switch the serialization code
over to Storable, using store_fd(), instead. Runs a bit faster, too.
Also allow passing by reference to parse_xsub(), to avoid copying
potentially large xsub bodies.
2006/01/30 kaffeetisch
* Glib.pm, NEWS, README: Unstable release 1.115.
2006/01/28 muppetman
* MakeHelper.pm: Slightly reduce the potentially ludicrous line
length of the xsdocparse command.
2006/01/22 muppetman
* GObject.xs: Honor the "don't warn about unregistered subclasses of
this type" flags. Use those flags to avoid creating _Unregistered
names for what are typically private implementation classes. This
leaves the method a little misnamed, but oh well. Updated the pod
accordingly. Also remove unused code.
2006/01/22 kaffeetisch
* AUTHORS, Makefile.PL: Apply a patch from Wim Lewis that makes
sure we don't use threading stuff if perl isn't configured to
handle threads.
2006/01/21 muppetman
* GObject.xs: Hide CLONE from the docs; it's not something user
code should call.
2006/01/18 kaffeetisch
* GObject.xs: Modify the MAKE_UNDEAD and REVIVE_UNDEAD macros so
no casts are needed when using them.
2006/01/18 kaffeetisch
* Glib.pm: Fix a typo.
2006/01/16 kaffeetisch
* Glib.pm, NEWS, README: Unstable release 1.114.
2006/01/16 kaffeetisch
* GObject.xs: Apply patch from Rafael Garcia-Suarez that fixes the
undead object macros on 64-bit platforms.
2006/01/16 kaffeetisch
* GObject.xs (class_info_finish_loading): Hush a compiler warning.
2006/01/16 kaffeetisch
* GLog.xs (gperl_log_handler): Add a comment describing a rare
threading-related crash.
2006/01/15 muppetman
* Glib.pm, GObject.xs, GType.xs: Change the way GObject type
mappings are initialized.
Previously, gperl_register_object() would not only map a GType to a
perl package name, but would set up the @ISA entry for that type by
introspection; if the parent type was not yet registered, the
mapping would go into a "pending isa" list, and the code would walk
that list on each registration. This left a nice loophole that was
triggered by the addition of GInitiallyUnowned into libgobject -- if
a type's parent is unknown to the bindings, that type simply doesn't
get its ancestry set up correctly. This ended up breaking all
nontrivial gtk2-perl programs. Shame on me for thinking up such a
fragile scheme.
Changed the code in various subtle ways. Added the idea of "lazy
loading", inspired by Gtk-Perl. gperl_register_object() still
registers the association of a GType and a package name, but adds to
@ISA only "Glib::Object::_LazyLoader". ISA setup is performed by
the internal function class_info_finish_loading(), which replaces
the _LazyLoader name in @ISA with the class's parent and implemented
interfaces (preserving manually appended and prepended entries).
This is triggered one of two ways: either by calling
gperl_object_package_from_type(), which happens when you bless an
object; or via an AUTOLOAD, isa, or can invocation in _LazyLoader,
which happens when you try to invoke methods on the class. When a
class is set up, all of its parents are checked as well, so this
happens fairly quickly. This should be completely transparent to
all existing code, as the new _LazyLoader stuff in private, and the
external interfaces have not changed. The one spot that could break
is calling UNIVERSAL::(can|isa) on a Glib::Object-derived package
name before that class has been set up.
In addition, instead of giving up on unknown classes, we now create
fake names for them in a "safe" namespace, along the lines of
Glib::Object::_Unregistered::$c_type_name, and treat them normally.
For foreign types, this still effectively means they can be treated
as the first known parent, but the value of ref() will be different
(but you shouldn't have been relying on that, anyway).
2006/01/13 muppetman
* GObject.xs: Ignoring the undead flag in _gperl_fetch_wrapper_key()
resulted in crashes when manipulating properties of objects with
undead wrappers.
* Glib.pm, Glib.xs, t/filename.t: Bind, test, and document
g_filename_display_name and g_filename_display_basename.
2006/01/01 kaffeetisch
* Glib.pm, NEWS, README: Unstable release 1.113.
2005/12/19 kaffeetisch
* GObject.xs (gperl_register_object): Move around the loop that
puts interface packages into @ISA so it gets run on every call to
gperl_register_object. With the old code, you ran into problems
when a GInterface was registered after the last GObject had been
registered.
2005/12/12 kaffeetisch
* Glib.pm, NEWS, README: Unstable release 1.112.
2005/12/12 kaffeetisch
* GType.xs: Hush a compiler warning.
* MANIFEST.SKIP: Use the more correct G.+\.c$ instead of G*\.c$.
2005/11/30 ebassi
* GKeyFile.xs: fix the argument types in order to allow 'undef' values.
* t/g.t: new tests for the functions allowing undefined values.
2005/11/18 kaffeetisch
* GType.xs Register the fundamentals for gint64 and guint64 as
Glib::Int64 and Glib::UInt64 respectively.
2005/11/14 kaffeetisch
* Glib.pm, NEWS, README: Unstable release 1.111.
2005/11/13 muppetman
* GValue.xs, GObject.xs: The GObject property interface passes around
pointers to boxed objects, but does not transfer ownership of them.
Thus, the old code was letting perl hold on to dead pointers. We
basically need to be able to specify whether to copy boxed objects
when converting from GValue to SV. Since gperl_sv_from_value() is
public and API frozen, add a new private function,
_gperl_sv_from_value_internal(), and call this from both
gperl_sv_from_value() and Glib::Object::get(). Fixes
http://bugzilla.gnome.org/show_bug.cgi?id=319204 and several other
hard-to-diagnose intermittent bugs.
2005/11/12 kaffeetisch
* xsapi.pod.head: Fix a few typos.
2005/10/18 kaffeetisch
* Glib.pm, MANIFEST, MANIFEST.SKIP, Makefile.PL, NEWS, README:
Unstable release 1.110.
2005/10/17 kaffeetisch
* GParamSpec.xs, GType.xs, GValue.xs, gperl.h, typemap, t/64bit.t:
Use custom converters for gint64 and guint64 that transform 64 bit
integers to and from strings if perl's IV type isn't big enough.
* AUTHORS, Glib.pm: Add some documentation about the above;
include a write up by A. Pagaltzis about the various big integer
modules for Perl.
* GSignal.xs (newSVGSignalQuery)
(gperl_signal_emission_hook): Fix some signedness warnings.
* GParamSpec.xs (gperl_param_spec_type_from_package): Fix a
variable-data-in-struct-construction warning.
2005/10/15 kaffeetisch
* GObject.xs (gperl_register_object): Free the interfaces list
after use.
2005/10/14 muppetman
* GObject.xs, AUTHORS: Patch from Steven Walter fixes nasty reference
counting bug that could leave Perl wrappers pointing at dead C objects
in some situations. Because we have no C containers in glib-object, we
can't really add tests for this.
2005/10/05 kaffeetisch
* Glib.pm, NEWS: Merge from stable-1-10.
* t/c.t: Change the flags test yet again. Make it explicit this
time by using is_deeply() and \@{} so it hopefully works with any
version of Test::More.
2005/09/30 kaffeetisch
* GObject.xs (gperl_register_object): Automatically add each known
interface an object implements to the corresponding @ISA.
(gperl_new_object): If we encounter an object which has no known
type in its ancestry except for G_TYPE_OBJECT, register it with a
unique package name and take a look at the interfaces it
implements -- if there is one we know, add it to the new package's
@ISA.
2005/09/05 kaffeetisch
* MANIFEST, META.yml: Remove META.yml.
* Glib.pm, Makefile.PL, NEWS, README: Stable release 1.100.
2005/07/27 muppetman
* CodeGen.pm: Forgot to delete the leftover gtk junk before committing.
2005/07/27 kaffeetisch
* Glib.pm, META.yml, NEWS, README: Unstable release 1.093.
2005/07/26 muppetman
* CodeGen.pm, Makefile.PL, MANIFEST: Take out the Gtk2-specific
portions of Gtk2::CodeGen, make it extensible, and call it
Glib::CodeGen.
2005/07/12 kaffeetisch
* Glib.pm, META.yml, NEWS, README: Unstable release 1.092.
2005/07/05 kaffeetisch
* GUtils.xs, t/1.t: Bind and test g_markup_escape_text.
2005/07/02 00:32 pcg
* GMainLoop.xs (add_watch): on native win32, convert the perl fileno
to a winsock fileno using win32_get_osfhandle and call
g_io_channel_win32_new_socket directly.
* Subclass.pm: fix my contact address while we are at it.
2005/06/28 kaffeetisch
* GenPod.pm (podify_ancestors): Use one less space for the
indention of ancestors to make them line up correctly.
2005/06/22 kaffeetisch
* t/c.t: Use is() instead of is_deeply() to test flags values
since the latter was changed to always use the string version of
overloaded objects in recent versions of Test::More, whereas the
former seems to work correctly. Tested with Test::More 0.45 and
0.60.
2005/06/20 kaffeetisch
* xs/GLog.xs (gperl_log_handler): Set perl's context to the one
gperl_handle_logs_for() was called in. This is to make sure
Perl_form accesses a valid memory pool, which prevents segfaults
when the log handler is called from a different thread.
* GType.xs: Fix a typo in gperl_package_from_type's POD.
2005/06/06 kaffeetisch
* Glib.pm, MANIFEST, META.yml, NEWS, README: Unstable release
1.091.
2005/05/30 kaffeetisch
* GKeyFile.xs: Fix the g_key_file_get_keys XSUB to actually return
anything, and make it omit NULLs. Fix an incorrect offset in the
g_key_file_set_locale_string_list XSUB. Replace "const gchar
const **" with "const gchar * const *" where appropriate.
* GObject.xs (Glib::Object::set_threadsafe): Mark `threadsafe' as
unused in the !GPERL_THREAD_SAFE branch.
* typemap: Add a GSignalFlags typemap. The corresponding
converters already exist and are exported.
* t/2.t: Test Glib::Object::freeze_notify and thaw_notify.
* t/7.t: Test Glib::Object::signal_stop_emission_by_name,
signal_handler_block, signal_handler_unblock and
signal_handler_is_connected.
* t/9.t: Test Glib::main_depth, Glib::MainLoop::is_running,
Glib::MainLooP::get_context, Glib::MainContext::new,
Glib::MainContext::default and Glib::MainContext::pending.
* t/c.t: Test the union operator of flags.
* t/g.t: Test Glib::KeyFile::set_locale_string_list,
get_locale_string_list, set_string_list, set_locale_string,
set_comment, set_boolean, set_value, get_keys, remove_comment,
remove_key, remove_group, to_data and set_list_separator.
2005/05/22 kaffeetisch
* Glib.xs: Initialize filename and hostname in filename_from_uri.
* t/filename.t: Add tests for the filename conversion facilities.
* t/a.t: Correct existing and add new tests for Glib::Log. Test
set_fatal_mask and set_always_fatal.
2005/04/27 ebassi
* doctypes: fix typo for GKeyFileFlags.
2005/04/27 ebassi
* Glib.xs, GError.xs, doctypes, gperl-gtypes.c, gperl-gtypes.h,
gperl.h, Makefile.PL, MANIFEST, typemap, GKeyFile.xs, t/g.t: bind,
document and test GKeyFile, a parser for .ini-like files.
2005/04/25 kaffeetisch
* Glib.pm, META.yml, Makefile.PL, NEWS, README: Unstable release
1.090.
2005/04/19 rwmcfa1
* GenPod.pm: fixed long running auto list of enums/flags problems.
enums were only getting added when there was info assocaited with them,
flags could always make it through. there was also a bug which allowed
enums and flags in parent properties to be listed in children.
2005/04/17 kaffeetisch
* Glib.pm, META.yml, NEWS, README: Merge from stable-1-08.
2005/04/04 muppetman
* GSignal.xs, t/signal_emission_hooks.t, MANIFEST:
Bind, document, and test g_signal_add_emission_hook and
g_signal_remove_emission_hook, with a little refactoring in
GSignal.xs.
2005/03/24 kaffeetisch
* GType.xs, GValue.xs, gperl.h: Add API that allows binding
developers to specify conversion functions when registering
fundamental types that are to be used by gperl_value_from_sv() and
gperl_sv_from_value(). gperl_register_fundamental_full() is the
new function that takes a GPerlValueWrapperClass struct which
contains a GPerlValueWrapFunc and a GPerlValueUnwrapFunc.
gperl_fundamental_wrapper_class_from_type() can be used to
retrieve the GPerlValueWrapperClass that corresponds to a given
GType.
* GBoxed.xs, GType.xs: Document that gperl_register_boxed() and
gperl_register_fundamental_full() assume that the wrapper class is
statically allocated and that it will always be alive.
2005/03/07 kaffeetisch
* Glib.pm, Makefile.PL, META.yml, NEWS, README: Stable release
1.080.
2005/03/06 kaffeetisch
* GBoxed.xs: Fix a typo.
2005/02/28 kaffeetisch
* Glib.pm, META.yml, Makefile.PL, NEWS, README: Unstable release
1.074.
2005/02/15 15:48 (-0500) rwmcfa1
* GMainLoop.xs: likely type-o on Glib::MainLoop->new causing ref count
to start out at 2 and thus main loops never die, should of been unref.
2005/02/07 23:40 (-0500) muppetman
* Glib.pm, NEWS, README: Unstable release 1.073
2005/02/07 23:30 (-0500) muppetman
* t/9.t: Disable the async signal test on win32.
2005/02/07 23:20 (-0500) muppetman
* GSignal.xs, gperl.h, Glib.exports: Bind and document g_signal_query
as Glib::Object::signal_query. Add newSVGSignalQuery to marshal a
GSignalQuery to perl.
* GType.xs: Move guts of list_signals to newSVGSignalQuery, and call
that function instead.
M MANIFEST
A t/signal_query.t: New test for signal_query and list_signals.
2005/02/07 23:20 (-0500) muppetman
* GLog.xs: Call abort() rather than croak() on G_LOG_LEVEL_ERROR
messages. croak()ing didn't do anything useful, and abort()ing
allows you to stop gdb on an error and get a backtrace.
2005/02/04 22:36 (-0500) muppetman
* MakeHelper.pm: (select_files_by_version) do not escape . in glob
patterns; it is unnecessary on unix and win32 thinks the \ is a path
separator. This one-character change fixes an intermittent win32
build problem.
2005/01/29 21:22 (-0500) muppetman
* ParseXSDoc.pm: better handling for xsubs with continuation lines.
2005/01/29 13:45 (-0500) muppetman
* copyright.pod: update for the new year.
2005/01/29 01:38 (-0500) muppetman
* GBoxed.xs, GenPod.pm: implement support for the boxed type GStrv
as a native perl anonymous array of strings
2005/01/25 23:34 (-0500) rwmcfa1
* t/c.t: remove [] around test 12. caused failures on newer
Test::More's
2005/01/09 23:50 (-0400) muppetman
* Glib.pm, META.yml, NEWS, README: Unstable release 1.072
2005/01/09 23:08 (-0400) muppetman
* Glib.xs: minor documentation updates
2005/01/09 16:29 (-0400) muppetman
* Glib.exports: add gperl_format_variable_for_output, which needs to
be exported but got missed somehow for 1.060.
2005/01/08 23:45 (-0400) muppetman
* GMainLoop.xs, t/9.t: Ensure that asynchronous signals, deferred by
perl's safe signal handling, are delivered on time when a main loop
is running. Thanks to Jan Hudec.
2005/01/04 15:15 (-0400) rwmcfa1
* t/1.t: we need to make it clear to perl that g_get_user_data_dir and
friends are functions or we'll get syntax errors and test failures when
no symbols with those names are found (even though they'll be skipped.)
2005/01/02 16:31 (+0100) kaffeetisch
* GParamSpec.xs, TODO: Remove documentation that doesn't apply
anymore.
* GUtils.xs, t/1.t: Bind and test g_get_user_data_dir,
g_get_user_config_dir, g_get_user_cache_dir,
g_get_system_data_dirs, g_get_system_config_dirs, and
g_get_language_names.
2004/11/29 21:28 (+0100) kaffeetisch
* Makefile.PL: Correct the unstable warning message to say 1.06x
instead of 1.04x.
* Glib.pm, META.yml, NEWS, README: Unstable release 1.071.
2004/11/01 21:22 (+0100) kaffeetisch
* Glib.pm, META.yml, Makefile.PL, NEWS, README: Unstable release
1.070.
2004/10/24 11:55 (-0400) rwmcfa1
* MANIFEST.SKIP: updates
* perl-Glib.spec.in: new scheme that addresses x86_64 problems
found by Carl Nygard
2004/10/20 19:32 (+0200) kaffeetisch
* GParamSpec.xs, e.t: Add and test support for GParamSpecUnichar.
2004/10/17 12:27 (+0200) kaffeetisch
* doctypes: Add a description for gunichar.
2004/10/14 15:27 (-0400) muppetman
* Makefile.PL: use PREREQ_PM in both WriteMakefile calls, so META.yml
gets created properly.
2004/09/27 09:35 (-0400) muppetman
* README, NEWS, Glib.pm: stable release 1.061
2004/09/19 23:31 (+0200) kaffeetisch
* GBoxed.xs: Use gperl_format_variable_for_output when croaking in
the unwrap function.
* GType.xs, Glib.xs: Fix two signedness issues.
* MakeHelper.pm: Print an ellipsis after "Including ApiDoc pod" to
indicate that it might take some time to comlete.
2004/09/11 15:23 (-0400) muppetman
* README, NEWS, Glib.pm: stable release 1.060
* Makefile.PL: remove the "unstable" warning.
2004/08/29 14:19 (-0400) muppetman
* README, NEWS, Glib.pm: unstable development release 1.055
2004/08/27 17:44 (-0400) rwmcfa1
* GenPod.pm: kaffee was seeing include failures which resulted in the
carp message comming out, but it was using $_ where it should of been
$pod which caused a second level of errors which were fatal.
2004/08/27 13:42 (-0400) rwmcfa1
* GenPod.pm: preprocess_pod wasn't being called on each pod section
as it used to be and still should be. couldn't get to sfcvs webui to
find out when it quit being, but it's been a while.
2004/08/15 17:33 (-0400) muppetman
* README, NEWS, Glib.pm: unstable development release 1.054
2004/08/13 22:14 (-0400) muppetman
* GenPod.pm: fix typo in doc
2004/08/06 21:10 (-0400) rwmcfa1
* GenPod.pm: get the types from signals prototypes and properties so
their enum values will be listed at the bottom of the pod. (this is in
addition to the xsubs and explicit stuff already in place)
2004/08/01 21:10 (-0400) muppetman
* README, NEWS, Glib.pm: unstable development release 1.053
2004/08/01 17:58 (+0200) kaffeetisch
* MakeHelper.pm: Move select_files_by_version and
read_source_list_file from Gtk2's Makefile.PL here so that other
modules can use them too. Also add Ross' cool "evenification"
hack to select_files_by_version.
* gperl.h, typemap: Add a char_own typemap.
2004/07/18 21:00 (-0400) muppetman
* README, NEWS, Glib.pm: unstable development release 1.052
2004/07/18 21:00 (-0400) muppetman
* GObject.xs: one last bug; always prefer to create the key with an
underscore in its name. this preserves the semantics of the old
property stuff from Glib::Object::Subclass::SET_PROPERTY(), although
it means we'll always try twice on keys with no dashes.
* Subclass.pm: remove the CHECK block (and the associated bugs) by
changing how Subclass works. we now rely on the new default property
handlers, and thus have no further need for GET_PROPERTY() and
SET_PROPERTY(); by prepending Glib::Object::Subclass to the new
class's @ISA, we can inherit new() normally. tested extensively in
my sandbox and a couple of people reported that all was well.
* t/f.t: yes, it's bad form to change the tests to fit the code, but
this test was verifying old behavior that we can't produce any longer,
and which was actually quite annoying in practice. verify the new
semantics.
2004/07/12 20:42 (+0200) kaffeetisch
* typemap: Make the OUTPUT variant of T_GPERL_GENERIC_WRAPPER do
the same as the INPUT one -- that is, strip leading "const"'s and
trailing asterisks from the type.
2004/06/28 00:47 (-0400) muppetman
* Glib.pm, README, NEWS: unstable development release 1.051
2004/06/28 00:40 (-0400) muppetman
* GType.xs: document the new property stuff.
2004/06/28 00:27 (-0400) muppetman
* GValue.xs: handle undef gracefully in gperl_value_from_sv(); let
the GValue default.
2004/06/28 00:17 (-0400) muppetman
* GObject.xs, GType.xs, t/f.t: add new semantics for Glib::Object
properties. when no GET_PROPERTY or SET_PROPERTY is defined, fall
back to using the key in the wrapper hash with the same name as the
property; also allow creation of properties with explicit getter and
setter functions, which override [SG]ET_PROPERTY and the fallback.
discussion of the ideas here is archived in the mailing list:
http://mail.gnome.org/archives/gtk-perl-list/2004-June/msg00091.html
2004/06/21 22:15 (-0400) muppetman
* Glib.xs: fine-tune gperl_format_variable_for_output(); don't
truncate references, and disambiguate undef.
2004/06/08 21:55 (-0400) muppetman
* AUTHORS, GLog.xs, GObject.xs, GParamSpec.xs, GType.xs, Glib.xs:
patch from Albert Chin keeps us honest in ANSI/C89 compatibility,
to allow compilation on IRIX. C++ comments are not allowed in C
files (some were on dead code, so i just removed them in applying
the patch); and C89 doesn't allow for non-static member
initialization (the FOO_TYPE_BAR macros usually evalulate to
function calls, so we have to do the param_types arrays the hard
and ugly way.)
2004/06/02 13:41 (-0400) muppetman
Unstable development release 1.050
* NEWS: updated.
2004/06/02 13:41 (-0400) muppetman
* GParamSpec.xs: cast to hush compiler warning.
2004/06/01 11:44 (-0400) muppetman
* GParamSpec.xs: remove stray debugging print
2004/06/01 11:18 (-0400) muppetman
Greatly expanded GParamSpec support.
* GParamSpec.xs: use attached-magic wrapper hashes and a GType
registry for GParamSpecs. Previous releases used blessed scalars
for Glib::ParamSpec and a list of simple hashes from
Glib::Object::list_properties; but they really should've been the
same object. this code puts the keys from the old simple hashes into
the magical wrapper hash now used for all Glib::ParamSpecs, to retain
backwards compatibility. we also add accessors for all of the
interesting members of the various GParamSpec child structures.
this allows you to get the extended metadata from paramspecs, such
as default or minimum values, etc. Glib::ParamSpec->param is now
supported, making it possible to use Glib::Object's notify signal.
lots of new manpages for the various param spec types, consolidated
where possible.
* gperl.h, Glib.exports: add new external function,
gperl_param_spec_package_from_type().
* GObject.xs: use the new and improved newSVGParamSpec() instead of
creating a plain hash.
* GType.xs: search through the paramspec type registry in
gperl_package_from_type(). register more fundamental types on boot,
with a nasty hack to correct the old mispelling of Glib::UInt (was
"Glib::Uint") with some semblance of backwards compat; use the new
correct name, but silently allow the old incorrect name.
* Glib.pm: update docs to reflect the type registration differences.
* MANIFEST, t/e.t: new test for all this new stuff
2004/05/31 21:05 (-0400) rwmcfa1
* GenPod.pm: add the type (flags|enum) back to the heading
2004/05/31 23:10 (+0200) kaffeetisch
* GObject.xs (CLONE): Remove unused variable.
* GType.xs: Document the interfaces keyword of register_object.
2004/05/31 15:11 (-0400) rwmcfa1
* GenPod.pm: reworked flags/enums code, added autodetection based on
return and param types. so all pod files should now docuement the
flags/enums used within them as returns and/or parameters.
2004/05/16 17:44 (-0400) rwmcfa1
* GObject.xs: code to allow gobject tracking which and threadsafeness.
needs lots of testing, documentation, and needs a way to tell if thread
support was compiled into perl. right now it's enabled (if compiled in)
by calling Glib::Object->set_threadsafe.
2004/05/16 11:25 (-0400) muppetman
* GObject.xs: disable stash caching. this fixes a few bizarre
things that go wrong when using perl threads; different threads
have separate copies of the stashes, but the caching kept only
one copy, which had a tendency to be incorrect in at least one
thread. to avoid an ABI break, we just look up the stash in
gperl_object_stash_from_type() every time, and save a pointer
on every registered object class.
2004/05/16 01:44 (-0400) muppetman
* GObject.xs: add missing Glib::Object::notify().
2004/05/11 19:35 (+0200) kaffeetisch
* GClosure.xs: Fix a debug message.
* GValue.xs: Adjust the apidocs for gperl_sv_from_value to match
the implementation.
* gperl.h
* Glib.xs: Implement and export new helper function
gperl_format_variable_for_output.
* GObject.xs: Use the new helper in gperl_get_object_check to
output a stringified version of the variable.
2004/05/04 18:11 (-0400) muppetman
* t/1.t, t/2.t, t/3.t, t/4.t, t/5.t, t/6.t, t/7.t, t/8.t, t/9.t,
t/a.t, t/b.t, t/c.t, t/d.t: add some commentary to explain what's
going on in each test. there's a lot of duplicated testing, so we
probably need to rework them at some point.
2004/04/21 12:36 (-0400) muppetman
* GSignal.xs: remove invalid alias do_stuff_by_func().
2004/04/14 19:50 (+0200) kaffetisch
Merge from the stable-1-04 branch:
* NEWS: Update.
* typemap, GBoxed.xs: use INT2PTR to hush warnings about improper
casting on cygwin.
* GError.xs: Mark unused variables.
* Glib.xs: Use explicit casts in the version checking part of the
BOOT code.
* GenPod.pm: document the __hide__, __gerror__, and __function__
modifiers for the apidoc directive. Add $MAIN_MOD to the head of the
see alsos list if it's set, so that the generated pages link to the
correct extension. Clean up the semantics of the copyright
generation stuff, and make the docs a little clearer.
* MakeHelper.pm: set $Glib::GenPod::MAIN_MOD with the module name in
Glib::MakeHelper::postamble_docs_full, either from the depends object
or the NAME key.
* GenPod.pm
* Glib.pm
* MakeHelper.pm
* ParseXSDoc.pm: Fix a spelling error here, adjust some grammar there.
* t/8.t
* GClosure.xs: don't clobber $_ when warning of unhandled exceptions.
apparently the SAVE_DEFSV trick from a while back doesn't work.
* Subclass.pm: beef up the docs.
2004/04/05 23:33 pcg
* Glib.xs: gperl_sv_from_filename didn't check the return status
of g_filename_to_utf8 properly, causing segfaults on invalid
filenames.
2004/04/04 00:52 (-0500) muppetman
* Glib.xs: actually do work in filename_from_unicode(). thanks to
Jan Hudec for noticing this was broken.
2004/03/21 00:00 (-0500) muppetman
merge from stable:
* GMainLoop.xs, GObject.xs: bump version check versions
* GUtils.xs, t/1.t: add Glib::GET_VERSION_INFO
* GenPod.pm: sort the index
* MakeHelper.pm: fix broken quoting
2004/03/17 17:09 (+0100) kaffeetisch
* gperl.h (SvGObject_ornull): Use SvOK instead of SvTRUE.
* gperl.h
* typemap: Add typemaps for "const char_ornull *" and
"char_ornull *".
2004/03/13 22:05 (-0500) muppetman
* MakeHelper.pm: portability fixes; ask Config what make program we're
using and generate the proper conditional syntax, lest we generate bogus
makefiles. we now work with Gnu Make and NMake.
2004/03/12 16:00 (-0500) muppetman
* Glib.pm, Makefile.PL, README: 1.04 will be released from
the stable-1-04 branch; bump HEAD's version to 1.050.
2004/03/11 23:15 (-0500) muppetman
* Glib.exports: sort the list, to make it easier to spot missing
symbols. add missing symbols.
* GBoxed.xs: fix dubious logic in parameter test.
* Glib.pm: give propers to kaffee
2004/03/11 21:51 (-0500) muppetman
* README: minor fixups.
2004/03/10 11:30 (-0500) muppetman
D Changes
M MANIFEST
remove the grossly-out-of-date and unused Changes file.
2004/03/08 23:52 (-0500) muppetman
FIXME sweep:
* GMainLoop.xs, GType.xs: cleanup and commentary
* GObject.xs, t/5.t: make the 'bad property' message from set() match
the one from new(). clean up properly before croaking if new()
encounters a bad property name. verify this behavior with tests.
* GParamSpec.xs: fix ALIAS on param_spec/boxed/object to remove
invalid symbol; document the fact that param_spec is unimplemented.
* GValue.xs: fix a nasty bug in gperl_value_from_sv that would've
resulted in a segfault when trying to read the C pointer out of an
SV containing a GParamSpec; implement the other direction in
gperl_sv_from_value. there's currently no way to trigger these in
Glib or Gtk2, but it would be conceivable to run across GParamSpec
properties in other GObject-based bindings, so we don't want to barf
on them.
* GenPod.pm: rework and simplify internal function podify_see_alsos()
* t/9.t, t/a.t: use version guards on tests that suffer from GLib
bugs that are now fixed.
* TODO: updated
2004/03/07 23:54 (-0500) muppetman
* GenPod.pm: don't add L<> around items that already have it in
podify_see_alsos
2004/03/07 03:33 (-0500) rwmcfa1
* GenPod.pm, ParseXSDoc.pm: added support for the __gerror__ keyword on
=for apidoc lines. to be used at a latter date.
2004/03/05 00:28 (-0500) muppetman
* Makefile.PL: add version checks to the import attempts for our
black-sheep ExtUtils modules; more recent versions actually support
the version checks.
* MakeHelper.pm: use almost excessive validation to detect whether
ExtUtils::Depends is of sufficient vintage, and bail out if not,
using the same trick as Makefile.PL to coax CPAN into updating.
2004/03/04 09:52 (-0500) muppetman
* MakeHelper.pm: fix a bug in the MY::const_cccmd override which
rendered makefiles broken and useless when there were no C or XS
files to compile.
* Makefile.PL: install the MakeHelper manpage.
2004/03/02 20:59 (-0500) rwmcfa1
* Glib.pm: alter the behavior of read only/write only tied properites.
setting a read only croaks, reading a write only returns
'[write-only]'. shouldn't affect anything/anyone, just make things
work a little better with Dumper and stuff like that.
* GObject.xs: doc changes ^
2004/03/02 00:42 (-0500) muppetman
* README, Glib.pm, NEWS: API-frozen unstable release 1.038
* GMainLoop.xs: change version guard on g_main_depth() to 2.3.5,
which was released today; minor cleanup and doc.
2004/03/01 12:17 (-0500) rwmcfa1
* GMainLoop.xs: g_main_depth bound as Glib::main_depth (ver 2.4)
2004/02/27 22:12 (-0500) muppetman
* GClosure.xs, GLog.xs, GMainLoop.xs, GType.xs: short descriptions
for the NAME sections
2004/02/27 14:38 (-0500) rwmcfa1
* GenPod.pm: mystery solved, Exporter needed quotes around it in the use
base call. as to why it died the way it was with "Unknown error" i don't
yet know, looking into it. use stict is now uncommented, and this issue
is closed so far as Glib is concerned.
2004/02/27 13:58 (-0500) rwmcfa1
* GenPod.pm: use warnings now used. Also turned on strict to get rid of
all of it's complaints, but even after all of them are gone it dies
with Unknown error, so use strict is commented out. lots and lots and
lots of warnings fixed, some of which were flat out bugs. one of which
was user visible, SEE ALSO's was wrong.
2004/02/27 02:00 (-0500) muppetman
* Glib.pm, README, NEWS: unstable release 1.037
2004/02/26 10:16 (-0500) muppetman
* t/1.t: correct the skip count. bad muppet.
2004/02/26 21:45 (-0500) muppetman
* README: fix the spelling of license; update the sandbox instructions.
* GUtils.xs, t/1.t: add more pod than code. enable, document, and test
various utility functions (get|set_application_name, get_user_name,
get_real_name, get_tmp_dir, get_home_dir).
* Glib.xs: correct the semantics of the boot-time version check
2004/02/26 22:59 (+0100) kaffeetisch
* GError.xs: Mark 'file' unused to hush compiler warnings.
* Glib.xs: Remove debug prints.
* ParseXSDoc.pm: Check if $firstline is defined before accessing
it to hush some warnings.
2004/02/25 19:28 (-0500) rwmcfa1
* ParseXSDoc.pm, GenPod.pm, GBoxed.xs, GError.xs, GMainLoop.xs,
GParamSpec.xs, GType.xs: =pod position=<pos> -> =for position <pos>
some pod parsers don't like text after =pod.
2004/02/25 15:21 (-0500) rwmcfa1
* GObject.xs: use new positioning pod mech. use #if 0 trick to get
tie_properties to act like a real method
* MakeHelper.pm: extra =back removed
* GBoxed.xs, GError.xs, GMainLoop.xs, GParamSpec.xs, GType.xs: use new
positioning pod mech.
* GenPod.pm, ParseXSDoc.pm: new positioning pod mech put in place
2004/02/23 17:05 (-0500) muppetman
* GError.xs: allow '' as well as undef when an SV into a GError,
as $@ is usually '' to say 'no error'.
2004/02/23 15:16 (-0500) rwmcfa1
* t/1.t: versions could be 0, which would fail before, check for
defined.
2004/02/23 14:38 (-0500) muppetman
* ParseXSDoc.pm: add hidden pods to the list of things to ignore when
checking if a package is empty; this keeps from clobbering the pod in
Glib.pm with Glib.pod, among other things.
2004/02/23 01:37 (-0500) muppetman
* GClosure.xs: don't clobber $_
* GenPod.pm: if an xsub is marked as a function, generate a
function's signature for it.
* GUtils.xs: mark the functions as functions.
* GSignal.xs: =arg -> =for arg
* GError.xs, gperl.h: add the ability to register new error domains
and throw Glib::Errors from Perl. New utility function,
Glib::Error::matches, makes it easy to test errors. New xs utility
function for parsing a perl data structure into a GError.
* t/d.t: bunches of new tests for the new Glib::Error features.
2004/02/22 21:44 (-0500) rwmcfa1
* GSignal.xs: =arg -> =for arg, apidoc error/type-o fixed.
* GUtils.xs, t/1.t: implemented CHECK_VERSION and friends with the same
scheme that went into Gtk2 eariler today.
* ParseXSDoc.pm: changed hide_hidden -> preprocess_pod. it now looks for
the __function__ keyword. changed the way these keywords are recorded,
they now go into xsub keys.
* GenPod.pm: use the above by looking for the function key in xsubs
and proceeding accordingly.
2004/02/22 15:26 (-0500) rwmcfa1
* GenPod.pm: removed warning about xsubs with no args, as they're
perfectly legal, and there are now some.
2004/02/22 12:48 (-0500) rwmcfa1
* GenPod.pm: method sorting implmented with suggestions from Marc and
Peter.
2004/02/20 01:18 (-0500) muppetman
* Glib.pm, README, NEWS: unstable release 1.036
2004/02/19 12:50 (+0100) kaffeetisch
* Glib.xs (gperl_alloc_temp): Return NULL and complain loudly if
the number of bytes is not greater than 0.
2004/02/19 01:18 (-0500) muppetman
* Glib.pm, GError.xs, TODO: documentation updates.
2004/02/18 01:35 (-0500) muppetman
A GError.xs
A gperl-gtypes.h
A gperl-gtypes.c
A t/d.t
M MANIFEST
M Makefile.PL
M Glib.exports
M Glib.xs
M Glib.pm
M gperl.h
change gperl_croak_gerror() to turn GErrors into exception objects.
the objects overload the stringify operator to be indistinguishable
from normal exceptions, resulting in no API change for old code.
gperl_croak_gerror()'s prefix argument is now useless, and has
been renamed to ignore, but not removed (to retain ABI and source
API compatibility). add gperl_register_error_domain(), and use it
for built-ins. doing this properly requires GEnum GTypes for the
error code enums, which are not provided by GLib, so we provide
just the relevant ones in private files which can be recreated
in the future as needed.
2004/02/18 01:31 (-0500) muppetman
* ParseXSDoc.pm: quell some doc messages unless NOISYDOC is set in
the environment.
2004/02/17 15:32 (-0500) muppetman
* GBoxed.xs, GObject.xs, GParamSpec.xs, GSignal.xs, GType.xs,
GUtils.xs, GValue.xs, Glib.pm, Glib.xs, MakeHelper.pm, Makefile.PL,
Subclass.pm, gperl.h, typemap: update copyright notices; add 2004 to
the stuff that has been modified this year.
* GenPod.pm: add time_t to %basic_types.
2004/02/12 19:00 (-0500) muppetman
* Glib.pm, Makefile.PL, README, NEWS: unstable release 1.035
2004/02/10 01:37 (-0500) muppetman
Some infrastructure changes allow us to enhance the quality of
generated POD.
* GenPod.pm: add a way to add entries to %basic_types at runtime.
* MakeHelper.pm: add postamble_docs_full(), and use it to implement
postamble_docs(). the new one takes parameters by name, and does a
lot more cool stuff than the old one. this allows us to make it
easier to set $Glib::GenPod::COPYRIGHT, provides a way to get
filenames through to Glib::GenPod::add_types(), and provides a more
extensible API for future expansion.
* Makefile.PL: use Glib::MakeHelper->postamble_docs_full() instead
of postamble_docs(). copyright information comes from new file
copyright.pod, and we use and install the new file doctypes, which
adds type mappings for stuff that isn't registered with the Glib
type system.
A doctypes
A copyright.pod
M MANIFEST
add copyright.pod and doctypes
* README: bump version requirement for ExtUtils::Depends.
2004/02/09 08:09 (-0500) rwmcfa1
* GObject.xs: needed version guards around the interface_properties
stuff b/c it's new to 2.3.x.
2004/02/08 15:16 (-0500) rwmcfa1
* GObject.xs, GenPod.pm: list_properties extended to be able to handle
interfaces. since interfaces don't inherite from GObject
Glib::Object::list_properties ($package) is required for it to work on
an interface. GenPod was updated to call list_properties this way.
2004/02/08 00:05 (-0500) rwmcfa1
* GParamSpec.xs, t/b.t: added GParamSpec->scalar, and tested it.
2004/02/08 20:46 (-0500) muppetman
* Glib.pm, README, NEWS: unstable release 1.034
* GenPod.pm: add bool and CV to %basic_types.
2004/02/07 02:22 (-0500) muppetman
* GType.xs: add the ability to add GInterfaces to new GObject types
with register_object. new key interfaces => \@list, where @list is a
list of package names. effectively does foreach (@list) {
$_->_ADD_INTERFACE ($newtype); }, where the _ADD_INTERFACE method is
supposed to do the actual work of adding and initializing the interface
implementation.
2004/02/05 21:10 (-0500) muppetman
* GType.xs: rework _INSTALL_OVERRIDES handling one more time. rather
than have a mess of inherited methods which chain up,
Glib::Type::register_object looks for _INSTALL_OVERRIDES in all
classes of the new type's ancestry, calling it if found. this allows
proper overriding behavior and lets us add vfunc implementations
without requiring changes to all client code.
Glib::Type::register_object now *always* instantiates the class upon
registration, to avoid problems with it maybe or maybe not existing
later. we also leak the reference to avoid the class needing to be
recreated, since we can't properly recreate it later, and perl
doesn't let classed die in normal circumstances, anyway.
* Glib.pm: add G_PARAM_READWRITE to the list of exportable constants.
* Glib.xs: better message?
2004/02/05 13:38 (-0500) rwmcfa1
* GObject.xs: fixed apidoc on ... param of Glib::Object->new
2004/02/04 23:44 (-0500) muppetman
* GObject.xs, GValue.xs: perl's IV is supposed to be as big as a
pointer, so we can't use glib's GINT_TO_POINTER family of casts,
because they'll truncate the IV on a 64-bit platform. use the
INT2PTR family of macros from the Perl api. this is important for
G_TYPE_POINTER GValues and g_object_[sg]et_data, for which we use
IVs to represent the pointer values.
* t/2.t: add some tests to ensure that we're handling int <=> pointer
conversions correctly, and that we can store pointers in object data.
as a side-effect, the test verifies that our unique hash wrappers work
correctly, as well.
2004/02/04 19:09 (-0500) rwmcfa1
* Glib.xs: implemented linked version checking in BOOT section
2004/02/04 20:19 (+0100) kaffeetisch
* gperl.h, Glib.xs: Add the new helper "object" GPerlArgv that
encapsulates everything related to argv parsing and thus
simplifies the various init xsubs.
2004/02/04 02:33 (-0500) muppetman
* GBoxed.xs, GObject.xs, GType.xs: GType is not a guint, it's a
gulong; therefore, GUINT_TO_POINTER/GPOINTER_TO_UINT are not valid
for GTypes. use direct casts, instead. this fixes nasty problems
on alpha. thanks to Mark Brockschmidt for finding the bug and
loaning me machine time to solve it.
2004/02/04 01:21 (-0500) muppetman
* GType.xs: call method _INSTALL_OVERRIDES on newly registered
Glib::Object subclasses, to allow the new types an opportunity to
install vfunc overrides. This method will typically be provided
by the class which introduces the vfuncs, and will install
implementations which marshal to inheritable perl methods.
Do nothing if no such method can be found; not all classes have
pure vfuncs.
2004/02/03 00:57 (-0500) muppetman
* GType.xs: remove debugging print
* GenPod.pm: add 'unsigned' to list of built-in mappings
* t/9.t, t/a.t: add mipsel, mips, and alpha to the list of 64-bit
platforms on which to skip tests which exercise glib bugs.
2004/01/30 22:36 (-0500) muppetman
* Subclass.pm: pass all unknown import parameters through to
Glib::Type->register_object. bump version.
2004/01/29 20:03 (-0500) rwmcfa1
* Glib.pm, README: unstable release 1.033
* NEWS: updated
1.032 isn't going public, was a test failure with no DISPLAY in Gtk2
2004/01/29 18:44 (-0500) rwmcfa1
* Glib.pm, README: unstable release 1.032
* NEWS: updated
2004/01/27 11:44 (-0500) rwmcfa1
* ParseXSDoc.pm, GenPod.pm: added a new =for type, see_also which
allows for appending things to see onto the end of the apidoc
autogenerated SEE ALSO sections, much desired/asked for feature. code
added to ParseXSDoc to find them and GenPod to use them.
2004/01/27 10:44 (-0500) rwmcfa1
* MakeHelper.pm: provide a mechinism for over-ridding the global
variables in GenPod.
* GenPod.pm: doc how to go about over-ridding the global vars.
2004/01/25 03:44 (-0500) muppetman
* Glib.pm, README: unstable release 1.031
* NEWS, MANIFEST.SKIP: updated
2004/01/25 01:00 (-0500) muppetman
A NEWS
M MANIFEST
add a file to summarize changes
* GObject.xs: add freeze_notify and thaw_notify.
* GenPod.pm: add gshort and gushort to %basic_types, and FIXME
notice on %basic_types to remind me that it needs to be extensible.
remove the gchar=>integer mapping, since it conflicts with
gchar=>string (the * is implied, you see)
* Glib.pm: add exportable constants under a 'constants' tag.
2004/01/24 21:44 (-0500) rwmcfa1
* MakeHelper.pm: added dist-srpms target, removed doc refering to
runtime_reqs as it has been replaced by the pkg-config trick
* Makefile.PL: removed runtime_reqs stuff, replaced by the pkg-config
trick
* perl-Glib.spec.in: use pkg-config for Requires version
2004/01/22 09:44 (-0500) rwmcfa1
* MakeHelper.pm: and i thought i was done. don't know what i was
thinking with the last one, laying in bed sleepless last night the
problem came to me. this is another attempt at a solution. this one
will work, but depends on makefile conditionals which may not be as
portable as i would like. also a fix to prevent the mkdir change from
a while ago from happening everytime make is done.
2004/01/21 20:56 (-0500) rwmcfa1
* MakeHelper.pm: yet another attempt at completely clearing up the
issues with the -jx make option and not having the doc rebuild at every
single make invocation. uses blib_done file, which depends on all of
the @xs_files. hopefully this will do it, i'm about out of ideas.
bascially MakeMaker sucks if you're tyring to do stuff this complicated
with it. with plain old make this wouldn't be an issue at all, but
with the hooks we're given this is not going to be easy.
2004/01/21 15:14 (-0500) muppetman
* ParseXSDoc.pm: allow space padding on either side of the arguments
in the xsub declaration. also make the error messages more
digestable (by removing the __WARN__ handler, was for debug only).
2004/01/21 14:57 (-0500) rwmcfa1
* MakeHelper.pm: we need to create the directory in which index.pod will
be placed under certian circumstances. this doesn't affect any of
Gtk2-Perl, but some modules (with certian structures) will cause
problems that creating the directory if need be will address.
2004/01/16 18:38 (-0500) rwmcfa1
* typemap: added types for gshort and gushort, patch by kaffee
2004/01/16 11:09 (-0500) rwmcfa1
* MakeHelper.pm: BLIB_MOD_EXIST -> BLIB_DONE. and set equal to something
that 'seems' to ensure that blib has completed. fixes -j make build
problems.
2004/01/16 00:23 (-0500) muppetman
* README: unstable release 1.030
* README, Makefile.PL: add emphasis of the unstable nature of this
series.
2004/01/16 00:02 (-0500) muppetman
A GUtils.xs
* MANIFEST, Makefile.PL, Glib.xs: add GUtils.xs, with version
information functions.
* ParseXSDoc.pm: don't complain about empty argument lists -- they are
legal on pure functions.
* AUTHORS: updated.
2004/01/15 22:08 (-0500) muppetman
* Glib.xs, Glib.pm: restore the prototypes on filename_to_unicode and
filename_from_unicode, to allow proper support for the function and
static method calling conventions. thanks to cxreg in irc for
reminding me that it would work.
also add initial implementations of filename_to_uri and
filename_from_uri, also callable as pure functions and class methods.
as currently implemented, there are issues with utf8 handling, but i
don't quite know how to resolve them.
make Glib be a full Exporter, so we can add these pure
functions to @EXPORT_OK and %EXPORT_TAGS.
update the docs accordingly.
2004/01/15 14:31 (-0500) muppetman
* GenPod.pm: name mapping for GPerlFilename_const
* gperl.h, typemap: add typedef and input typemap for
GPerlFilename_ornull
* GType.xs, gperl_marshal.h, t/7.t, t/8.t: preserve $@ across
closure invocations.
* GSignal.xs: add words of warning to documentation
2004/01/14 10:59 (-0500) rwmcfa1
* GObject.xs: added pod documentation for tie_properties
2004/01/14 00:21 (-0500) muppetman
* GBoxed.xs, GType.xs, GValue.xs: more spots where we should use
GINT_TO_POINTER and friends
* t/9.t, t/a.t: skip tests that exercise glib bugs on x86_64; we
need version API in the bindings to do this correctly.
2004/01/13 22:41 (-0500) muppetman
* AUTHORS, Glib.xs, GObject.xs, GType.xs, typemap: applied
patch (with extra fuzz) from Jacek Konieczny <jajcus at bnet dot pl>
to clean up warnings on 64-bit platforms.
2004/01/13 22:20 (-0500) muppetman
* ParseXSDoc.pm: support continuation lines in xsubs. this borks line
number handling, and is far more permissive than xsubpp, but xsubpp
will be the thing that causes people's code to break, and to do the
line number handling correctly would require far too much code tearup.
also, honor __hide__ correctly on =for apidoc comments that supply
a symbol name (by removing the end-of-line anchor from the regex).
2004/01/09 14:33 (-0500) rwmcfa1
* AUTHORS: mailing list addr correction.
2004/01/08 21:22 (-0500) rwmcfa1
* Glib.xs: bit the bullet and made filename_to_unicode and
filename_from_bullet work with either :: or -> syntaxes.
* Glib.pm: doc'd ^ as using the -> syntax.
there was great discussion on the list about this issue (in a more
general sense) bascially early on the decision/mandate was made that
the -> syntax was to be used everywhere. right or wrong it's to late
change now that stable stuff has hit the world. the above functions
work either way, so we're safe.
2003/12/31 02:03 (-0500) muppetman
* GValue.xs: paranoia -- check for sv==NULL before passing to SvOK.
2003/12/30 11:49 (-0500) rwmcfa1
* GenPod.pm, MakeHelper.pm, Subclass.pm: added version numbers
2003/12/30 10:12 (-0500) rwmcfa1
* MakeHelper.pm: moved the const_cccmd to the bottom of the file so
that the first package isn't MY (caused issues with CPAN and generally
wasn't a good idea.) Also converted the comment about it into a pod
NOTICE so that people will be aware that it's being added.
2003/12/29 16:32 (-0500) rwmcfa1
* MakeHelper.pm: added DATE to the list of replacements, in a format
that rpm is happy with
* perl-Glib.spec.in: use the new DATE replacement in conjunction with
VERSION to create the changlog on the fly, which is better.
2003/12/23 23:40 (-0500) muppetman
* typemap: ouch, there was a typo in the T_GCHAR_ORNULL output
typemap, resulting in unknown variable errors.
2003/12/23 00:52 (+0100) kaffeetisch
* typemap: Add typemaps for gsize and gssize.
2003/12/18 13:54 (-0500) rwmcfa1
* MakeHelper.pm: rpms now build in HOME/rpms dir, one change to fix
them all.
2003/12/16 22:00 (-0500) rwmcfa1
* GType.xs: warnings fixes, minor (non-problem) bugs
2003/12/16 13:16 (-0500) rwmcfa1
* MakeHelper.pm: the V deps have now been put under a var
BLIB_MOD_EXISTS which is set to pm_to_blib now. (1 point of change if
need be in the future)
2003/12/15 20:37 (-0500) rwmcfa1
* MakeHelper.pm: added pure_all as a dep on a few things.
2003/12/10 15:18 (-0500) rwmcfa1
* t/c.t: now that the register_enum|flags stuff is fixed, test it out
more thoroughly.
2003/12/10 12:45 (-0500) muppetman
* Subclass.pm: updated doc
* GType.xs: fix memory corruption problem in register_enum and
register_flags by duping and leaking the value names. there appears
to be no way to clean up that memory, unfortunately.
also moved those two xsubs up in the file to appear immediately
after register_object in the docs, and added examples to their docs.
move the existing register to register_object, since that's all it
does. created a new function named register, which acts as a
traffic-cop, calling the proper helper function based on the type
from which you are attempting to derive. this would've been easier
to do in Perl, but then getting the doc into the right place would
require a rather large tearup of Glib::ParseXSDoc.
* TODO: a few of these are completed, now.
2003/12/10 12:02 (-0500) rwmcfa1
* t/c.t: initial import, tests out the new register_enums|flags code
2003/12/09 13:26 (-0500) rwmcfa1
* GType.xs: another pass at getting things going. created
sanitize_package_name, used it where app.
2003/12/08 22:44 (-0500) rwmcfa1
* GType.xs: implemented first pass at register_enum and register_flags
2003/12/08 11:08 (-0500) muppetman
* GenPod.pm: add default translation for stdio FILE => "file handle"
2003/12/08 09:44 (-0500) rwmcfa1
* t/6.t: use strict and warnings, b/c newer Test::More do anyway,
required a few minor syntax changes to work.
2003/12/04 13:07 (-0500) rwmcfa1
* MakeHelper.pm: @ARGV is now searched for disable-apidoc which,
surprise, disables the generation of the api doc pods thus speeding up
the build process.
2003/12/04 00:12 (-0500) muppetman
* Glib.pm: credit where credit is due
* Glib.pm: bump version to 1.030 to keep cvs head ahead of the
stable series.
2003/12/02 23:49 (-0500) rwmcfa1
* : Merged from rel-1-02-branch
* GenPod.pm: implement a sorter function that pushes GObject based pkgs
to the front, instead of alpha. this is so that all classes will be
instantiaed (sp?) by the time we do the interfaces.
2003/12/02 18:30 (-0500) muppetman
* t/b.t: inexplicitly missing a test
2003/12/02 18:30 (-0500) muppetman
* GType.xs: base_init support for perl-derived GObjects. during
class initialization, invoke the method INIT_BASE in the object's
package if it exists. see http://lists.gnome.org/archives/gtk-\
perl-list/2003-November/msg00194.html (and indeed the rest of that
thread) for discussion.
* t/6.t, MANIFEST: test the base_init stuff by ensuring multiple
derivation works as expected.
2003/12/02 18:26 (-0500) muppetman
* GParamSpec.xs: stray code in a switch statement prevented the
successful operation of Glib::ParamSpec->object().
2003/12/01 16:21 (-0500) rwmcfa1
* GenPod.pm: put an extra \n after each of the pods to keep pod stuff
happy
2003/11/29 12:45 (-0500) muppetman
* Glib.pm, README: bump version to 1.013, make release from
rel-1-02-branch.
2003/11/29 12:23 (-0500) rwmcfa1
* GenPod.pm: modified copyright text as list decided.
2003/11/28 22:00 (-0500) muppetman
* GType.xs: remove unused var and code.
* Glib.xs: apparently this bit of doc has been b0rken for a very
long time.
2003/11/28 14:47 (-0500) rwmcfa1
* ParseXSDoc.pm, GenPod.pm: moved inclusion processing into GenPod
rather that ParseXSDoc, it will need to be run later than parse.
* MakeHelper.pm: DOC_PL_DEPENDS is now POD_DEPENDS, related to ^
2003/11/25 12:02 (-0500) muppetman
* GType.xs: clean up some odd leaks; use proper contexts in the
signal accumulator and custom class closure.
2003/11/25 11:39 (-0500) rwmcfa1
* GType.xs, devel.pod: link correction
* GenPod.pm: added parents return from list ancestors. use it for
podify_see_alsos, which is new. and added get_copyright
* Glib.pm: removed reference to old Glib::PkgConfig
* MakeHelper.pm: added new DOC_PL_DEPENDS var, should prove useful
for includes
* ParseXSDoc.pm: added include system =for include file,
=for include !cmd
2003/11/24 23:35 (-0500) rwmcfa1
* Makefile.PL, MakeHelper.pm: clean up and resolution of issues about
xsapi.pod.
2003/11/21 02:12 (-0500) muppetman
* Glib.pm, README: bump version to 1.012 and release.
2003/11/21 01:00 (-0500) muppetman
* GBoxed.xs: implement a generic Glib::Boxed::copy, and some doc
* GBoxed.xs, GObject.xs, GType.xs, GValue.xs, typemap: replace
SvTRUE() with SvOK() as a more efficient test for definedness.
SvTRUE() is rather heavy, in that it evaluates to an awful lot
of code, where what we generally need is just to see if the scalar
is the Perl equivalent of C's NULL.
* GParamSpec.xs: now that there's a gunichar typemap, implement
things that were waiting on it. not sure why you'd want a unichar
param in perl, but the point is to make it possible. :-)
2003/11/19 14:08 (-0500) muppetman
cleanup for win32:
* GType.xs: hush some warnings from msvc
* MakeHelper.pm: if we're using cl, use its weird command switch
to put the object files in the right place... /Fo<file>
* Glib.exports: added some missing symbols
2003/11/18 17:21 (-0500) muppetman
* Glib.xs: use GPerlFilename_const in the right place to hush the
compiler. also use the length returned from g_filename_(to|from)_utf8
to avoid having to call strlen() in gperl_sv_from_filename and
gperl_filename_from_sv.
2003/11/18 01:09 (-0500) muppetman
* GType.xs: interfaces can have signals as well; adjust logic to
allow them through as well. this still misses some signals because
on 2.0.x we don't have a way to list interfaces' prerequisite types
to ensure that they are loaded. the beginnings of code to use
g_type_interface_prerequisites() in 2.2.x is commented out, waiting
for me to finish it.
* ParseXSDoc: handle length() arguments in xsubs. they basically
just get stripped.
* GenPod.pm: 'gchar_length' means 'string'.
* typemap: a corresponding OUTPUT typemap for gunichar.
2003/11/17 22:02 (+0100) kaffeetisch
* typemap: Implement a gunichar INPUT typemap.
2003/11/16 20:32 (-0500) muppetman
* gperl_marshal.h: add similar macros for GPerlCallback, to make sure
that the callback is executed by the correct interpreter.
* GClosure.xs: use the new GPerlCallback helper macros in
gperl_marshal.h
2003/11/16 02:44 (-0500) muppetman
* GClosure.xs, GType.xs: clean up the stack properly after call_sv
to stop internal bleeding.
* TODO: updated
2003/11/14 04:31 pcg
* Glib.pm: rework the doc for filename conversion. horribly buggy
it still was.
2003/11/13 21:32 (-0500) muppetman
* Glib.pm, README: bump version to 1.011 for first 1.02 beta
2003/11/13 13:23 (-0500) muppetman
* GenPod.pm: add a sensible type name mapping for GPerlFilename;
will show up in docs as 'localized file name'.
Fix podify_methods to take only the package name, and remove the
second my $package in the same scope.
Minor re-work to logic of when to skip and when to count methods;
intent was to support hidden xsubs, but those are caught at the parser
level now.
Changed the message that gets written when all methods are defined
but missing.
* Glib.pm: fix some typos
* Glib.xs: hide filename_(to|from)_unicode from the docgen stuff,
to keep from overwriting the manpage generated for Glib.pm. these
functions are already documented by hand in Glib.pm, anyway.
* ParseXSDoc.pm: allow the token \b__hide__\b on the =for apidoc
line to hide an xsub.
* GType.xs: fix a little bug in list_values causes segfaults if you
pass in type that's neither a flags nor enum type.
2003/11/12 20:45 pcg
* gperl.h, typemap: Add GPerlFilename_const.
* Glib.xs, Glib.pm: implement and document filename_to_unicode and
filename_from_unicode.
2003/11/12 02:48 pcg
* gperl.h, Glib.xs, typemap: Implement GPerlFilename type and
gperl_filename_from_sv & gperl_sv_from_filename utility functions.
2003/11/11 22:25 (-0500) muppetman
* GClosure.xs, gperl_marshal.h: yeah, that wasn't the right way to
handle that. i forgot the update the non-PERL_IMPLICIT_CONTEXT side
of the #ifdef when adding the second arg. backed out the change to
GClosure.xs, fixed gperl_marshal.h.
2003/11/11 10:40 (-0500) rwmcfa1
* GClosure.xs: if PERL_IMPLICIT_CONTEXT is not defined then the call
to the macro GPERL_CLOSURE_MARSHAL_INIT has the wrong number of
prarameters. there's probably a better way to handle this, buf for
now this will allow things to compile.
2003/11/11 00:36 (-0500) muppetman
* MANIFEST, debian/*: remove the debian packaging files, since having
them in the upstream dist makes it difficult for the maintainers
2003/11/10 13:12 (-0500) muppetman
* gperl_marshal.h: new file
* GClosure.xs, MANIFEST, Makefile.PL: use and install new file.
Since client code may install custom marshallers, it is a good idea
to provide some macros to tidy up the error-prone boilerplate that
needs to go into each of those marshallers. gperl_marshal.h is
installed to the same place as gperl.h, but not #included by default,
and should be #included whereever you implement a custom marshaller.
it is loaded with commentary and an example marshaller skeleton.
* GSignal.xs: update commentary on gperl_signal_set_marshaller_for()
to pointer binding authors to gperl_marshal.h for guidance.
2003/11/10 01:54 (-0500) muppetman
* GenPod.pm: be a little smarter about enums and flags package names.
2003/11/09 22:59 (-0500) rwmcfa1
* GenPod.pm: improved the logic for placement of the no methods exist
message, moved it to podify_methods.
2003/11/09 22:30 (-0500) rwmcfa1
* ParseXSDoc.pm: allow the (Object::To::Doc) construct in =for object
directives.
* GenPod.pm: support the new object directive allowing files to be
named independantly from the objects documented in them.
2003/11/10 00:59 pcg
* Glib.pm, GType.pm: implement == and eq for flags, the latter
because some functions use eq as a generic comparison function.
Also enable fallback for flags values, should work just fine.
2003/11/09 18:54 (-0500) muppetman
writing documentation while mechagodzilla and titanosaur destroy
tokyo... again. http://us.imdb.com/title/tt0073373/
* GClosure.xs, GLog.xs, GMainLoop.xs, GObject.xs, GParamSpec.xs,
GSignal.xs GType.xs: added lots of apidoc pod. All the packages in
GMainLoop.xs go into Glib::MainLoop, and i reordered some of the
xsubs in GObject.xs but changed no code.
* GenPod.pm: pod updates, cleanup; use two =over paragraphs inside
xsub docs to make the argument lists indent readably in the generated
manpages.
* Glib.pm: big updates to the SEE ALSO section.
* Makefile.PL: even yet still more cleanup
* ParseXSDoc.pm: pod cleanup
* TODO: updated
2003/11/09 16:29 (-0500) rwmcfa1
* ParseXSDoc.pm: added the method clean_out_empty_pods to prevent empty
entries from the $data member of doc.pl. this prevents completely empty
pod files from being created.
* GenPod.pm: if no methods are bound to a package then print a message
saying it is likely that it doesn't exist in the version the module was
compiled against. the message might could stand some improvement.
2003/11/08 12:10 pcg
The "more evil things?" release.
* GType.xs: Allow |, & and ^ as operators on flags, too.
* GType.xs: GFlags are now represented as blessed IV's, which indeed
gave a _substantial_ speed improvement at an unimportant place. But
it also saves memory and code, and looks nice, so I just had to do it.
* GLog.xs: simplify newSVGLogLevelFlags, as excluding single values
should no longer be necessary.
2003/11/08 11:30 pcg
* GType.xs: In addition to the change yesterday, make sure
that gperl_convert_back_flags only ever returns distinct,
non-overlapping flag values. This requires an inherent ordering
in the flag values, but if this isn't the case we can always sort
the values first on bit-count (doh), or use a heuristic such as
prefering single bit values.
2003/11/08 01:03 (-0500) muppetman
* GClosure.xs: simplify, man... both dSP and the PERL_UNUSED_VAR()
should've been outside the conditional. this may need to be
macro-ified to prevent bugs in custom marshallers.
* Glib.xs, GType.xs, gperl.h: move private function streq_enum
to Glib.xs as new public function gperl_str_eq(), string compare that
considers dash and underscore equivalent. add gperl_str_hash(), a
hashing function that considers dash and understore equivalent.
* Glib.exports: updated with new functions
* GSignal.xs: use custom hash and compare functions for the signal
marshallers hash.
* gperl.h: add prototypes for gperl_fundamental_type_from_package()
and gperl_fundamental_package_from_type().
* GObject.xs: remove unused var (which was shadowing an already-defined
var in the parent scope).
2003/11/07 20:13 pcg
* GType.xs: test SvPOK instead of SVt_PV, as many scalars
can be valid strings.
* GType.xs: apply muppet's patch to bless flags, modified to
- implement overloaded operations on flags.
- change the flag test from "set & mask" to "set & mask == mask",
which hopefully fixes the GdkModifier problem.
- since the testsuites of both glib and gtk work fine, as
well as my apps, I checked it in (low-risk-patch).
* Glib.pm: add overload glue + documentation.
* In a second, future step, we might want to replace GFlags by
bless(RV(IV)) while keeping the user-api intact. This might
break some obscure parts in Gtk2, so... this is a later step
that can be done without user-visible-changes.
2003/11/07 12:57 (-0500) rwmcfa1
* GenPod.pm: instead of usings methods trick, use ->can, improves
things quite a bit.
2003/11/06 21:29 (-0500) rwmcfa1
* GenPod.pm: using the old methods trick i've come up with a way to
only have xsubs pod for package methods that are bound. so ifdef'd out
code (stuff that's not available in the compiled against version) won't
be put into the pods. one more doc issue down.
2003/11/06 16:53 (-0500) muppetman
* MakeHelper.pm: make the pod index have actual links.
2003/11/06 10:30 (-0500) muppetman
* ParseXSDoc.pm: handle IN_OUTLIST arguments, which previously got
a type of "(!!)". note: still does not handle IN_OUT or OUT args,
as i don't have any of those to test.
2003/11/03 14:03 muppetman
* GenPod.pm: use the actual instance arg name rather than cutting
up the type name. you'd expect the arg name to be used, anyway.
2003/11/03 12:04 muppetman
* GType.xs: (list_interfaces) watch out for unregistered interface
classes. complain, but don't blow up. also, free the returned array
of GTypes. thanks to Bjarne Steinsbø.
* AUTHORS: updated contributors
2003/11/02 11:54 muppetman
* GClosure.xs: the case of the missing ref; refcount bug in
gperl_closure_marshal kept complex objects from finalizing properly.
kudos and thanks to marc for finding and debugging this one.
2003/11/02 16:55 muppetman
* ParseXSDoc.pm: match void\s* instead of just void when looking for
methods with void return. use scalar(localtime) instead of the
identical strftime format.
2003/11/02 16:17 muppetman
* Glib.pm: undo my silly "atleast" version-check import stuff, and
do something compatible with the version check supplied by Exporter.
we don't use Exporter because, well, we don't export anything.
2003/11/02 02:15 muppetman
* ParseXSDoc.pm: split_aliases wasn't properly deep-copying args
arrays. since that's not exactly something that you'd want to do
unless you are going to generate docs from the parser results,
moved the splitting of aliases from parse_file() to xsdocparse()
as canonicalize_xsubs().
2003/11/01 01:15 muppetman
* typemap: input (Perl-to-C) typemap for T_GPERL_GENERIC_WRAPPER
now uses a more robust regex -- strip leading const\s+ and ensure
that we strip the trailing \s*\* from the very end. requested
by kaffee. the output typemap is unchanged.
2003/10/30 13:47 rwmcfa1
* Glib.pm: added tie_properties stuff
* t/b.t: initial import, test tie_properties stuff
2003/10/30 11:29 muppetman
* GenPod.pm: make sure that all signatures get the same pod directive.
don't add sigils to return type names.
2003/10/30 09:30 rwmcfa1
* Glib/MakeHelper.pm: improved/safer rules for building the documentation
* Glib/Makefile.PL: we have some 'special' dependancies for Glib that we need to make sure happen before xsdoc2pod can proceed
2003/10/29 02:00 muppetman
* Glib.pm: add a version check to Glib's import.
* GenPod.pm, MakeHelper.pm: rework package to filename handling to
support packages with :: in the name. use FULLEXT (makefile var that
is the full extension path under INST_LIB) and File::Spec."
* ParseXSDoc.pm: don't clobber package with object; store object
separately and use it when retrieving pkgdata, but not in place of
the package name. this allows xsubs to be placed into a different
package for doc, but not lose their actual package for the symname.
2003/10/28 23:18 muppetman
* MakeHelper.pm: don't be an Exporter. take a shift-off class arg
on all subs. this allows them to be overridden, and removes the
requirement for each client package to import the symbols.
added pod for just about everything.
do_pod_files now returns the generated pod_files hash rather than
fiddling with one in main::. it still uses the local @gend_pods,
though.
postamble_clean() takes an optional list of extra clean files.
postamble_docs() takes a list of xs files instead of expecting
to find @main:xs_files.
* Makefile.PL: use the new semantics of the MakeHelper functions.
2003/10/28 16:47 rwmcfa1
* Glib/GClosure.xs, Glib/GLog.xs: put things in the appropriate doc
object
* Glib/GenPod.pm: eval the list_values call. make sure dir creation
succeeds, other cleanups
* Glib/MANIFEST: missing files added
* Glib/Makefile.PL: updates to new system, MakeHelper. use strict and
use warnings safe now
* Glib/ParseXSDoc.pm: replace the warn handler with what was there
before rather than deleting
* Glib/MakeHelper.pm: initial import
2003/10/27 29:08 muppetman
* Makefile.PL: bad quoting kept the build from working...
2003/10/27 19:29 muppetman
* GenPod.pm: move =for enum parsing to the parser.
have xsdoc2pod optionally write a list of the packages it has created,
use 'unsigned' instead of 'integer' for unsigned data types.
fix podify_signals to print nothing if there are no signals.
don't call Glib::Type->package_from_cname on types with :'s in
them -- they can't be C type names anyway (shuts up about half of
the warnings you get doing Gtk2). watch for =arg name (__hide__),
and don't display the arg having that keyword.
* Makefile.PL, xsdocparse.pl(removed), ParseXSDoc.pm(new): remove the
xsdocparse.pl script; the code in is Glib::ParseXSDoc now. having
xsapi.pod in %pm_files creates a circular dependency; have to take it
out (still in man3pods, and we'll put it into blib by hand).
2003/10/27 02:42 muppetman
* Makefile.PL: manpages have the :: in the filename
* GParamSpec.xs, GLog.xs, GMainLoop.xs, GSignal.xs: added some apidoc;
register enums and flags as necessary; implement g_param_spec_enum and
g_param_spec_flags
* GenPod.pm: pod, cleanup, lots of things. watch for
=for (enums|flags) in package pod and podify their values.
change around how we print things to make them a little more
visually distinct.
2003/10/26 21:31 rwmcfa1
* GenPod.pm: put the pod files in blib using the fully correct paths
and file names
* Makefile.PL: merge bug, well that and not looking closely.
2003/10/26 17:10 rwmcfa1
* GObject.xs: test out a blurb
* GenPod.pm: xsdoc2pod script incorporated, will make it easier for
other modules to use this stuff
* Makefile.PL: go ahead and include GenPod.pm in the scheme of things.
use GenPod in it's current incarnation to do the api docs
2003/10/26 11:36 muppetman
* GSignal.xs: some apidoc comments to play around with
* GenPod.pm: sigils don't go in %basic_types, they're added by
convert_return_type as needed. count the number of properties we
find that match this type, to return an empty string if this type
has none of its own. added some developers' pod for the various
helper functions. allow 'for' on signature and arg lines. show
default argument values in call signatures.
* Makefile.PL: install GenPod.pm. its manpage will be nonsense
for now.
* xsdoc2pod.pl: typo
* xsdocparse.pl: split & duplicate aliased xsubs automagically.
match up apidoc pods and xsubs.
2003/10/25 21:44 rwmcfa1
* GType.xs: can't call newSVpv if package name is not valid, shouldn't
happen so croak on that error. some improved error checking to make
sure that we have a valid interface and that our class is
instantiatable, may need to double check.
* apidoc.pl: get rid of the groups stuff since it's not used
* xsdocparse.pl: get rid of the groups stuff since it's not used.
* GenPod.pm: initial import
added blurb retrieval
* xsdoc2pod.pl: initial import
2003/10/24 17:18 muppetman
* xsdocparse.pl: new file. program extracts pod and xsub signatures
from XS files and dumps a big honkin' data structure to stdout.
differentiates between the C section and the xsub section, and looks
for a new set of pod directives, e.g. =for apidoc and =for object.
this is destined to be the heart of the shiny new automatically
generated api reference documentation. more on this later.
* apidoc.pl, Makefile.PL: use xsdocparse.pl instead of podselect
to extract pod used to build xsapi.pod.
* GType.xs: new function Glib::Type->package_from_cname() turns
a c type name into registered Perl package name, if found.
also added some =for apidoc pod to test out xsdocparse.pl.
2003/10/23 16:04 rwmcfa1
* GType.xs: Glib::Type->list_ancestors, list_interfaces, and
list_signals added. with that V commit this starts to bring together
some petty powerful magic.
2003/10/23 15:51 muppetman
* GType.xs: Glib::Type->list_values ($enum_or_flags_package).
also takes a C type name, e.g., "GtkWidgetFlags", in case the type
is registered with glib but not with the bindings.
* GObject.xs: fix list_properties to allow package names instead
of just objects.
2003/10/18 02:36 muppetman
* Glib.pm: bump version for devel series
* typemap, gperl.h: new typedef and typemaps for the much-needed
gchar_ornull.
* GLog.xs: use the new gchar_ornull typemap to clean things up a bit.
* GType.xs: commentary
2003/10/12 13:42 rwmcfa1
* (most).xs, gperl.h: UNUSED is no more, PERL_UNUSED_VAR used where
needed, relies on new ParseXS to prevent the other warnings that were
its fault, we now only quite the ones that are ours
2003/10/09 22:23 muppetman
* Glib.pm, META.yml, README, debian/changelog: 1.00
* GType.xs: typo in the documentation.
2003/10/03 15:54 muppetman
* Glib.pm, META.yml, README, debian/changelog: 1.00rc4
2003/10/02 03:34 rwmcfa1
* Glib.xs: core dumps occur in gperl_croak_gerror when NULL is passed
for err, assertion now checks for it. only binding authors would run
across this one, but know it will be more ovbious what they've done
wrong, speaking from expirence...
2003/10/02 03:13 muppetman
* gperl.h: bad nesting of parens in a debugging macro
* GClosure.xs: bug in error-handling code in gperl_callback_invoke
made it a little hard to track down the fact that a callback was
created with the wrong parameter type.
2003/10/01 11:20 rwmcfa1
* Glib/GBoxed.xs, Glib/GLog.xs, Glib/GObject.xs, Glib/GSignal.xs,
Glib/GType.xs, Glib/gperl.h, Gtk2/xs/GtkAccelGroup.xs,
Gtk2/xs/GtkIconFactory.xs, Gtk2/xs/GtkListStore.xs,
Gtk2/xs/GtkObject.xs, Gtk2/xs/GtkToolbar.xs, Gtk2/xs/GtkTreeView.xs,
Gtk2/xs/GtkWidget.xs: // comments -> /*
2003/09/26 04:10 muppetman
* Glib.pm, META.yml, README, debian/changelog: 1.00rc3
* GClosure.xs: assert that we have a callback before doing anything
to it. Clean up a bad comment char.
2003/09/21 20:19 rwmcfa1
* Makefile.PL: fixed bug in specfile generation
2003/09/21 15:03 rwmcfa1
* (lots of files): license updates/additions
2003/09/19 00:21 muppetman
* Glib.pm, META.yml, README, debian/changelog, debian/control: updated
for 1.00rc2 release
* Makefile.PL: there was no version 0.1 of ExtUtils::PkgConfig...
2003/09/17 10:44 rwmcfa1
* Makefile.PL: ExtUtils::PkgConfig can now deal with version
requirements using pkg-config's interface, make use of it.
* t/6.t: removed, this tested Glib::PkgConfig
2003/09/16 23:50 rwmcfa1
* Makefile.PL, perl-*spec.in: somewhat automated versioning system
implemented for depenancy modules
* PkgConfig.pm: moved to new ExtUtils::PkgConfig module
* MANIFEST: PkgConfig.pm removed from manifest
2003/09/16 15:00 muppetman
* GClosure.xs, gperl.h: changed type of "tag" in signature of
gperl_remove_exception_handler() to avoid signedness warnings.
the tag's supposed to be a guint anyway.
* GObject.xs, GMainLoop.xs, GParamSpec.xs, GSignal.xs, GType.xs,
GValue.xs, Glib.xs: cleanup: casting to hush pedantic compiler
warnings; can't have UNUSED in PREINIT:, put it in CLEANUP:;
quell no-return warnings on functions that croak.
2003/09/15 22:26 rwmcfa1
* Makefile.PL: spec file dependancies improved
2003/09/15 17:43 muppetman
* gperl.h, Glib.exports: add gperl_signal_set_marshaller_for and
gperl_closure_new_with_marshaller
* GClosure.xs: add gperl_closure_new_with_marshaller, allowing you
to specify a custom marshaller to use for a GPerlClosure. you can't
just call g_closure_set_marshal again, because glib asserts that it
hasn't been set yet. this is used by gperl_signal_connect, as part
of the changes described below. includes POD that warns you that
you really don't want to use this function. gperl_closure_new now
just calls this.
* GSignal.xs: add thread safety to the closures list. added
gperl_signal_set_marshaller_for, to allow client code to specify a
special marshaller to use for all invocations of a particular named
signal on a particular class; this is intended to allow workarounds
for unhelpfully defined signal parameter types.
* Glib.pm: bumped version to rc2, since we export new C-level APIs.
2003/09/11 23:33 muppetman
* Glib.pm, MANIFEST, META.yml, README, debian/changelog: updated
for 1.00rc1 release
2003/09/11 23:10 rwmcfa1
* gperl.h: added UNUSED macro
* ChangeLog, GClosure.xs, GLog.xs, GMainLoop.xs, GObject.xs,
GParamSpec.xs, GType.xs: make use of UNUSED macro to quite warnings
about class and/or ix being unused, other slight/small warnings fixed
when -Wall
2003/09/11 22:30 muppetman
* Subclass.pm: documentation about creating properties, creating
signals, and overriding class closures.
* gperl.h: more than one file needs gperl_type_class
* GType.xs: unused variable
2003/09/11 21:50 muppetman
* GSignal.xs, Glib.exports, gperl.h: implement
g_signal_chain_from_overridden and a wrapper for GSignalFlags;
make signal_emit actually catch return values and give them back
to perl; add output-only wrapper for GSignalInvocationHint.
* GType.xs, t/7.t, t/8.t: make class closures optional for new
signals; allow the creation of signals that have return values;
allow perl to supply custom accumulators for valued signals;
make overriding class closures actually work.
2003/09/11 13:27 muppetman
* t/7.t: use strict and warnings in this one, too.
2003/09/11 10:35 rwmcfa1
* t/[1234568a].t: use strict and warnings, fixes to make doing so work.
2003/09/06 19:23 rwmcfa1
* Makefile.PL: dist-rpms build target added
* perl-Glib.spec.in: initial import
2003/09/05 01:50 muppetman
* README, debian/changelog: update for 0.97 release
2003/09/05 00:51 muppetman
* Glib.pm: more pod (too much? should it be in a separate doc?)
* t/a.t, Glib.exports, Glib.xs, Makefile.PL, gperl.h, MANIFEST:
add support for routing g_log messages through perl
2003/09/01 19:52 muppetman
* Glib.pm: don't set dl_load_flags on darwin, it generates a warning.
also bump version so that cvs versions of Gtk2 which depend on
T_GPERL_GENERIC_WRAPPER will force upgrades.
* typemap: use a new generic typemap for all the types which follow
the SvMyType/newSVMyType naming convention.
2003/08/28 23:30 muppetman
* gperl.h, GClosure.xs, Glib.xs: merged changes from the exceptions
branch. gtk2-perl now attempts to do reasonable handling of
exceptions in callbacks. still needs pod for the perl-level stuff,
but i'm not sure where to put it.
* Glib.exports: new exportable symbols
* GParamSpec.xs: hush uninitialized value warning (drop in the bucket)
* Glib.pm, MANIFEST, README, debian/changelog: update for 0.96 release
2003/08/28 17:45 muppetman
* GSignal.xs: clean up comments, bind signal_handler_is_connected
2003/08/22 01:11 muppetman
* Glib.pm, README, META.yml, debian/changelog, MANIFEST:
prep for 0.95 release
2003/08/19 21:11 rwmcfa1
* Makefile.PL: added realclean removal of build dir to postamble
section.
2003/08/18 23:45 muppetman
* t/7.t: test out signal stuff on a subclassed object.
2003/08/18 17:28 muppetman
* GSignal.xs: quiet down remember_closure and forget_closure, now
that i'm sure they work.
2003/08/16 20:53 muppetman
* GSignal.xs: implemented wrappers for g_signal_handlers_block_by_func,
unblock_by_func, and disconnect_by_func. code for the _matched
versions of these functions (the generic ones) is commented out,
because there is no typemap for GSignalMatchType.
2003/08/15 09:35 muppetman
* debian/*, AUTHORS: patch from James Curbo adding maintainence
files for debian packages.
* MANIFEST, README, Glib.pm: updated for 0.94 release
2003/08/14 16:54 muppetman
* apidoc.pl, Makefile.PL: use a simple perl script instead of shell
commands in the makefile to create xsapi.pod; not all platforms have
the shell tools, and this will let us sneak in more sophisticated
stuff later.
* PkgConfig.pm: portability fixes
2003/08/14 11:50 muppetman
* devel.pod, Makefile.PL, Glib.pm: new manpage, Glib::devel, describes
the philosphy of how the bindings are designed and how they work; an
overview-style supplement to Glib::xsapi, which details the APIs
themselves.
2003/08/13 23:44 muppetman
* GObject.xs: fatalistic commentary
* GIOChannel.xs, Glib.xs, Makefile.PL: new file GIOChannel.xs
* GClosure.xs, GMainLoop.xs, GSignal.xs, gperl.h: closure
correctness and simplification fixes
* GValue.xs: reordering to assure that all the same types are
handled in both switches; croak on unhandled types rather than
just warn --- an unhandled type is a bug that we need to know
about; it may require a handler registration mechanism.
* TODO: updates
2003/08/13 18:15 muppetman
* xsapi.pod.foot, xsapi.pod.head: new files
* GBoxed.xs, GClosure.xs, GObject.xs, GSignal.xs, GType.xs, GValue.xs,
Glib.xs, Makefile.PL, gperl.h: moved all the gtk-doc commentary in
gperl.h into embedded pod in the various XS files; now we can extract
all the pod from the XS files and convert that into an installable
API reference document, Glib::xsapi.
2003/08/12 12:06 rwmcfa1
* PkgConfig.pm: can now take multiple pkg name parameters, see t/6.t
for examples of usage.
* t/6.t: test out PkgConfig.pm some.
2003/08/09 23:57 pcg
* gperl.h: export gperl_register_fundamental.
* Glib.exports: add it here, too (actually, this was an anonymous
check-in by muppetman... tsk, tsk..)
2003/08/09 02:09 muppetman
* GBoxed.xs, t/2.t: removed a FIXME
2003/08/01 12:41 muppetman
* MANIFEST, README, Glib.pm: updated for 0.92 release
2003/07/31 19:40 muppetman
* Glib.exports, Makefile.PL, gperl.h: a few windows-specific things
2003/07/30 09:21 rwmcfa1
* GBoxed.xs: bug found by matthias blasing that when getting a legal
property off off a gobject that had no value assigned would croak with
a NULL to boxed error. croak changed to warn and wrapped in a NOISY.
returns undef from the null test
2003/07/29 14:52 rwmcfa1
* GObject.xs: #ifdef'd out the warning when returning undef from
gperl_new_object, tree's can rightfully ellicit this, if you're in a
state of heavy devel on wrappers you might want to re-enable it
2003/07/29 00:58 pcg
* Subclass.pm: fix typoe.
2003/07/25 10:21 muppetman
* Glib.pm, README: bump version to 0.91
2003/07/23 23:27 pcg
* PkgConfig.pm: allow better win32 compatibility (I made a native,
win32 build of Glib today).
2003/07/18 12:05 muppetman
* Glib.pm, README, META.yml
2003/07/17 11:44 muppetman
* GBoxed.xs, GObject.xs, GTypes.xs, Glib.xs, Makefile.PL: add the
necessary locking to make the perl bindings thread-safe. thanks go
to Brett Kosinski for contributing the initial patch and bugging me
about it, which he needed for work on GStreamer bindings.
2003/07/10 09:50 muppetman
* Glib.pm: $VERSION is actually treated as a string, write it as one
to avoid bootstrap problems when using a locale that writes numbers
differently. (thanks to thierry for finding that.)
2003/07/06 16:48 muppetman
* GSignal.xs: add g_signal_stop_emission_by_name
2003/07/05 06:37 pcg
* typemap: add gchar_own*, for functions that return allocated
strings and gchar_length* for functions expecting a length() input.
* gperl.h: add gchar_own and gchar_len typedefs.
2003/07/04 13:35 muppetman
* Makefile.PL: patch from Thierry Vignaud fixed my copy and paste
bug which leads to errors when ExtUtils::Depends isn't present.
* ChangeLog: bump to 0.26 for release
2003/06/28 03:16 pcg
* GType.xs: remove superfluous gperl_object_new call.
* GObject.xs: important bugfix: the calls with own==FALSE in
GType.xs caused premature death of the perl wrapper. The fix
incurs a slight performance hit, so only do it when own==TRUE, as
this should be the only case where it is ever required.
* t/5.t: test for this case.
2003/06/27 17:00 muppetman
* GObject.xs: get_data and set_data work only with UVs now,
because it's not safe to treat them as anything else, and
since we can use hash keys they aren't needed for anything else.
2003/06/27 12:00 muppetman
merged realobjects-branch back onto HEAD. this was a week's worth
of work, mostly by pcg with some help from me. here's a summary of
changelog items from that branch:
* README.api-changes: clearinghouse for API gripes, please amend
* GParamSpec.xs, Glib.xs, typemap, gperl.h, Makefile.PL: support
for pspecs, so we can add properties to gobjects.
* Subclass.pm, t/[345].t: extra module to ease implementation of
subclasses, and some tests for it
* Glib.pm: pod updates
* GObject.xs: big change in how perl wrappers are implemented.
the new implementation will only ever create one perl wrapper (a
real hash, with magic containing the gobject's address) for any
gobject. The combined perl+gobject will stay alive as long as one
of the partners is alive. The only real changes are in
gperl_new_object and in the DESTROY method.
* GType.xs: rename INSTANCE_INIT to INIT_INSTANCE and make it a
function call as opposed to a method call. Also call FINALIZE_INSTANCE
on object finalization time. implemented creation of properties
for an object in Glib::Type::register, and implemented calling
of SET_PROPERTY and GET_PROPERTY on the resultant object.
other changes during the merge:
* Makefile.PL: install MAN3PODS with the correct names.
* PkgConfig.pm: add some documentation
2003/06/25 09:25 muppetman
* GValue.xs, GClosure.xs: add a MODULE directive to appease
ExtUtils::ParseXS
2003/06/20 11:29 muppetman
* README, Makefile.PL: bump version number for 0.24 release
2003/06/19 00:44 pcg
* Depends.pm: moved to ExtUtils-Depends module.
* Makefile.PL: add ExtUtils-Depends as prerequisite, please
install the ExtUtils::Depends module seperately from cvs.
2003/06/16 14:21 muppetman
* GBoxed.xs, gperl.h: converted GBoxed to use vtable-based wrappers;
this allows client code to install new wrapper classes that make
gperl_sv_from_value and gperl_value_from_sv call the right code
to convert boxed types in custom ways. GPerlBoxedPackageFunc has
disappeared and the call signature for gperl_register_boxed has
changed (well, the meaning of that one extra parameter -- if you
already passed NULL, there's nothing to worry about).
* GMainLoop.xs, typemap: added support for some GMainContext and
GMainLoop methods.
2003/06/13 14:12 muppetman
* GObject.xs: rename Glib::Object->_new to Glib::Object->new -- it
should be that because people expect it. other cleanup. NOTE: we
need to have object wrappers be able to override things like "take
ownership", "wrap", "unwrap", and "destroy"... as it is, if you
create a wrapper for a GtkObject from Glib::Object->new, the wrapper
will be created incorrectly (gtk_object_sink will not be called).
to do all of this correctly will also require wrapper caching,
which i haven't made work reliably... but i'm working on it now.
2003/06/12 15:18 muppetman
* GClosure.xs: protect yourself from maniacal code that passes in
something to get a return value from closures/callbacks with no
return value. this makes gperl_value_from_sv stop warning about
bad types on the activate signal for GtkEntry, for example.
2003/06/11 09:25 muppetman
* GValue.xs: undef is also false; use SvTRUE instead of SvIV when
evaluating an SV as a gboolean to allow undef to pass quietly.
* GType.xs: add Glib::Double
2003/06/17 18:21 muppetman
* gperl.h, xs/GObject.xs: move sink functions into Glib,
to avoid some problems in which GtkObject created via
Glib::Object->new don't have gtk_object_sink called on them.
this problem is actually more generic than just GtkObject (other
libraries do similar things).
2003/06/09 13:49 muppetman
* gperl.h, typemap: corrected broken handling for GObject output
typemap; now uses SvGObject/newSVGObject-style macros and the
_ornull/_noinc variants that Gtk2 uses.
2003/06/06 23:55 muppetman
* GObject.xs, GType.xs: copy scalars that will be stored, don't just
mess with the reference count, for along that way lies madness.
* MANIFEST: added some key missing files (AUTHORS, LICENSE, ChangeLog)
* Glib.pm: bumped version number to 0.22 for public release
2003/06/01 16:16 muppetman
applied patch from Marc Lehmann (pcg at goof dot com)
* GObject.xs: allow $object->get and $object->set to work on
multiple properties.
* Glib/GClosure.xs: don't try to PERL_SET_CONTEXT on non-threaded perls
2003/05/30 23:55 muppetman (while watching zep at the albert hall)
* GType.xs: initial values, try to avoid warnings and bugs
* AUTHORS, GClosure.xs, gperl.h, typemap: applied more patches from
Brett Kosinski; added typemaps for gint64 and guint64, ensure that
closures and callbacks are executed by the interpreters that created
them.
2003/05/29 15:51 muppetman
* GType.xs, gperl.h: applied patch (with liberal modifications) from
Brett Kosinski (brettk at frodo dot dyn dot gno dot org) to add
gperl_try_convert_flag, like the existing gperl_try_convert_enum.
rearranged things to avoid copied code, and use the _try_ versions
as the actual workhorses. added a little more commentary, too.
2003/05/26 04:28 muppetman
* Glib.pm: bumped version for development release
* GBoxed.xs: ensure that the scalar from which we try to extract
a boxed pointer is actually a reference.
2003/05/22 10:30 muppetman
* AUTHORS, LICENSE every other file: in every file in the project
added a header pointing to the new AUTHORS and LICENSE files.
changed the license clause in the docs from "same as perl" to GPL.
2003/05/18 14:00 muppetman
* ChangeLog: since breaking the monolithic build into pieces,
ChangeLog entries for Glib will be in here. i've included the
contents of the toplevel log up to this point for history.
* MANIFEST: updated for proper make dist
* Glib.pm: changed version number
* README: updated
2003/05/17 09:06 rwmcfa1
* Glade/*: first pass at GladeXML added
* Makefile: added, see comments within
* Makefile.PL: is no more
* lots and lots of files: G -> Glib and other related/required changes
2003/05/16 14:55 muppetman
* Gtk2/xs/GtkSpinButton.xs: removed get_value_as_float because it is
deprecated (and had the wrong return type anyway).
2003/05/15 11:45 muppetman
* G/GSignal.xs: hush unused parameter warning, more efficient
this way, anyway
* Gtk2/xs/GtkContainer.xs: implemented foreach
* Gtk2/xs/GtkIconFactory.xs: minor bugfix
* G/GType.xs, G/GValue.xs, G/gperl.h, G/typemap,
Gnome2/xs/GnomeProgram.xs, Gtk2/xs/GtkCombo.xs, Gtk2/xs/GtkDialog.xs,
Gtk2/xs/GtkFrame.xs, Gtk2/xs/GtkItemFactory.xs, Gtk2/xs/GtkStock.xs,
Gtk2/xs/GtkTextBuffer.xs, Gtk2/xs/GtkToolbar.xs,
Gtk2/xs/GtkTooltips.xs, Gtk2/xs/GtkTreeViewColumn.xs: use newSVGChar
and SvGChar instead of newSVpv and SvPV_nolen for gchar*, to ensure
valid utf8 handling. this definitely needs testing.
2003/05/06 12:56 rwmcfa1
* Gtk2/t/1.GtkWindow.t, Gtk2/t/2.GtkButton.t: prevent windows from
fighting over focus and thus stall tests (happends with wmaker)
* Gtk2/t/16.GtkMenu-etc.t: we don't really know how to use
tearoff_state so for the time being we won't
2003/05/05 23:11 muppetman
* G/G.pm, Gtk2/Gtk2.pm: pod updates
2003/05/05 16:35 muppetman
* Gtk2/xs/GtkTooltips.xs: work around a (bug|feature) in the Gtk+
C library by storing a GtkTooltips reference in the GtkWidget's user
data. doesn't hurt normal behavior, and prevents some hard-to-explain
pitfall errors.
2003/05/03 11:17 joered
* Gtk2/: Gtk2.pm, Makefile.PL, pm/Helper.pm: removed deprecated
timeout/idle/input methods from Gtk2.pm; added Gtk2/pm/Helper.pm
with a convenience implementation of add_watch/remove_watch
2003/05/02 18:11 muppetman
* Gtk2/gtk-demo/apple-red.png, Gtk2/gtk-demo/background.jpg,
Gtk2/gtk-demo/gnome-applets.png, Gtk2/gtk-demo/gnome-calendar.png,
Gtk2/gtk-demo/gnome-foot.png, Gtk2/gtk-demo/gnome-gimp.png,
Gtk2/gtk-demo/gnome-gmush.png, Gtk2/gtk-demo/gnome-gsame.png,
Gtk2/gtk-demo/gnu-keys.png: images needed by Gtk2/gtk-demo/pixbufs.pl,
directly from the gtk+-2.2.1 source distribution.
* G/GType.xs, Gtk2/examples/histogramplot.pl: a bunch of code,
borrowed from pygtk, to add signals to a derived class.
altered the histogramplot example to use a new signal.
* Gtk2/xs/GtkTooltips.xs: allow tip_private to default to NULL
2003/05/02 00:30 muppetman
* Gtk2/Gtk2.pm, Gtk2/gtk2perl.h, Gtk2/examples/scribble.pl,
Gtk2/gtk-demo/drawingarea.pl, Gtk2/gtk-demo/pixbufs.pl, Gtk2/xs/Gdk.xs,
Gtk2/xs/GdkTypes.xs, Gtk2/xs/GtkCellRenderer.xs, Gtk2/xs/GtkWidget.xs:
reverted the whole GdkRectangle mess. it's a boxed type again.
this clears up several bugs to do with GdkAllocation, and in fact
makes the whole shebang more efficient because we don't just create
an array any time the GdkRectangle is needed, in which case it is
often thrown away or only one element is used. added a ->values
function, like in gtk2-perl, which returns the members in a list
in the order you'd want for passing to several important gdk
functions. Gtk2::Gdk::Rectangle->new is good for creating new
rectangles.
* Gtk2/xs/GtkTextView.xs, Gtk2/xs/GtkTreeView.xs: implemented some
functions found to be missing when looking for GdkRectangle returns
that needed to be marked _copy.
2003/05/01 23:17 joered
* Gtk2/xs/GtkCombo.xs: added GtkCombo->entry and GtkCombo->list
returning the correspondent widgets
* Gtk2/xs/GtkHBox.xs, Gtk2/xs/GtkVBox.xs: default homogenous is 0
and default spacing is 5, as in gtk-perl
* Gtk2/xs/GtkWidget.xs: widget flags can now be set with
Widget->flag_name(1) resp. unset with Widget->flag_name(0);
Widget->flag_name() still returns current state
2003/05/01 08:00 rwmcfa1
* Gtk2/xs/GtkStyle.xs: removed deprecated functions, there was a lot
of them.
2003/04/31 01:00 muppetman
* Gtk2/xs/GtkFrame.xs: properly allow undef in Gtk2::Frame->new (was
adding an empty string instead of passing NULL)
* Gtk2/xs/Gdk.xs, Gtk2/xs/GdkPixbufLoader.xs, Gtk2/xs/GdkRegion.xs,
Gtk2/xs/PangoContext.xs, Gtk2/xs/PangoLayout.xs: newly implemented
* Gtk2/examples/histogramplot.pl: new code to test drive drawing
primitives, pango text handling and drawing, and subclassing. lots
of stuff in here; was ported from a working C class library i've
been writing.
* G/GType.xs, G/GObject.xs, Gtk2/xs/GtkObject.xs: support for
pure-perl GObject subclasses.
added G::Type->register to create a new GType (basically wraps
g_type_register_static), G::Object->_new, to be called from perl
constructors for things inheriting GObject, and Gtk2::Object->new,
which MUST be used for things inheriting GtkObject (to handle the
floating ref situation properly).
this allows the perl developer to create new widgets without writing
C code! adding signals and properties is currently not implemented.
* Gtk2/CodeGen.pm: more correct handling of undef --- previous code
was allowing a variable containing undef to pass, which would cause
a croak in the wrapper-reader function.
* Gtk2/Gtk2.pm, Gtk2/xs/GdkTypes.xs: stopgap solution, simple lvalue
subs to get members from a rectangle list
* Gtk2/gtk-demo/drawingarea.pl: revert to named member method syntax
for rectangles
* Gtk2/gtk-demo/pixbufs.pl: actually works now. you need the images,
which i don't think are in CVS yet.
* Gtk2/gtk-demo/stock_browser.pl: cleanup
* Gtk2/xs/GdkPixmap.xs: implemented create_from_xpm_d and
colormap_create_from_xpm_d
* Gtk2/xs/GdkWindow.xs: allow NULL for cursor in set_cursor
* Gtk2/xs/GtkWidget.xs: implemented create_pango_layout
2003/04/29 21:55 joered
* Gtk2/xs/GtkListStore.xs: bugfix: gtk_list_store_append and
gtk_list_store_prepend were swapped
2003/04/29 23:44 muppetman
* Gtk2/xs/GdkDrawable.xs: implemented gdk_draw_polygon,
gdk_draw_points, gdk_draw_lines, gdk_draw_segments,
gdk_draw_layout, and gdk_draw_layout_with_colors
* G/GObject.xs: removed some very old and very broken commented-out
code (wrapper instance caching). added new methods for dealing with
foreign objects in perl: G::Object->new_from_pointer(VAL, NOINC) (a
direct wrapper around gperl_new_object), and $object->get_pointer.
2003/04/29 18:10 muppetman
* Gtk2/xs/GtkWindow.xs: icon list stuff
* Gtk2/gtk-demo/images.pl, Gtk2/gtk-demo/pixbufs.pl: the last two
pieces of gtk-demo (need some image files and such, though), and
these two don't work correctly.
* Gtk2/gtk-demo/appwindow.pl, Gtk2/gtk-demo/button_box.pl,
Gtk2/gtk-demo/changedisplay.pl, Gtk2/gtk-demo/colorsel.pl,
Gtk2/gtk-demo/dialog.pl, Gtk2/gtk-demo/editable_cells.pl,
Gtk2/gtk-demo/item_factory.pl, Gtk2/gtk-demo/list_store.pl,
Gtk2/gtk-demo/main.pl, Gtk2/gtk-demo/menus.pl,
Gtk2/gtk-demo/panes.pl, Gtk2/gtk-demo/sizegroup.pl,
Gtk2/gtk-demo/stock_browser.pl, Gtk2/gtk-demo/textview.pl,
Gtk2/gtk-demo/tree_store.pl: gtk-demo runs! lots of cleanup in the
pieces, and changed each one to us a single entry point name, defined
in a package with the same name as the file; this bit of subterfuge
was necessary because of the differences between C and perl, and the
fact that the app is designed as a C program. NOTE: drawingarea.pl
is broken, because my copy has other changes that won't work with
the current state of CVS.
2003/04/29 16:16 rwmcfa1
* Gtk2/xs/GdkGC.xs: impelemted gdk_gc_set_dashes
2003/04/29 15:10 muppetman
* Gtk2/xs/GtkSizeGroup.xs: implemented size groups
* Depends.pm, G/Depends.pm, G/Makefile.PL, Gnome2/Makefile.PL,
GnomePrint2/Makefile.PL, Gtk2/Makefile.PL, GtkSpell/Makefile.PL,
G/PkgConfig.pm, Gtk2/CodeGen.pm, helpers/genstuff.pl,
helpers/genboot.pl: build system hacks.
moved Depends.pm under G, so that G can install it.
made a module of some boilerplate to handle dealing with pkgconfig,
and converted the code in helpers/genstuff.pl and helpers/genboot.pl
into Gtk2/CodeGen.pm, stuff that can be called from Makefile.PLs.
hacked up a the Makefile.PLs to reflect these changes.
this makes it possible to use the autogen build tools outside the
source tree --- the first step towards breaking up the source tree
into separately distributable modules.
2003/04/29 11:14 muppetman
* Gtk2/gtk-demo/sizegroup.pl: another little piece of my heart
* Gtk2/xs/GtkDialog.xs: use alias to ensure that new_with_buttons
exists for those who seek it.
2003/04/28 23:25 muppetman
* Gtk2/xs/GtkListStore.xs, Gtk2/xs/GtkTreeStore.xs: work around
bizarre stack behavior by not using a helper function. trying to
read the stack in a helper function called from an xsub was resulting
in the stack showing the wrong number of items.
2003/04/28 18:00 muppetman
* Gtk2/gtk-demo/panes.pl: another piece of the gtk-demo pie
* G/G.pm G/GObject.xs: overload the == operator, for more natural
object comparisons
* Gtk2/xs/GtkPaned.xs: struct member access
* Gtk2/Makefile.PL Gtk2/genkeysyms.pl: create a big hash of key
symbols in Gtk2::Gdk::Keysyms, a la gtk-perl
* Gtk2/xs/GtkRadioButton.xs, Gtk2/xs/GtkRadioMenuItem.xs,
Gtk2/xs/GtkButton.xs, Gtk2/xs/GtkCheckButton.xs,
Gtk2/xs/GtkCheckMenuItem.xs, Gtk2/xs/GtkImageMenuItem.xs,
Gtk2/xs/GtkMenuItem.xs, Gtk2/xs/GtkToggleButton.xs: consolidate
constructors with ALIAS to avoid copying code. make sure that
group isn't a valid SV pointing to undef
* Gtk2/xs/GtkWidget.xs: implemented queue_draw, add_accelerator,
remove_accelerator, and get_display
2003/04/27 08:52- rwmcfa1
* Gtk2/t/16.GtkMenu-etc.t: Added a first pass at the testing of
GtkMenu and friends.
* Gtk2/t/15.GtkHandleBox.t: Added a decently complete test of
GtkHandleBox
* Gtk2/t/14.GtkToolbar.t: Added a decently complete test of GtkToolbar
* Gtk2/xs/GtkToolbar.xs: Fixed a bug where GtkToolbarChildType was
wrongly being used as a gtype, caused all _element functions to fail
* Gtk2/t/13.GtkTooltips.t: Added a decently complete test of
GtkTooltips
* Gtk2/xs/GtkTooltips.xs: implemented an attempt at
gtk_tooltips_data_get which returns what is in the GtkTooltipsData
struct as a hash.
* Gtk2/t/12.GtkDialog.t: Added a decently complete test of GtkDialog
* Gtk2/xs/GtkDialog.xs: added gtk_dialog_add_butttons with multiple
calls to gtk_dialog_add_button. combined the vbox and action_area get
functions into one aliased function. also changed a char* to a gchar *
* Gtk2/maps: hand added entry for Gtk2::Progress, removed
* Gtk2/xs/GtkProgressBar.xs: added a BOOT section with a isa call to
tell GtkProgressBar that it's a GtkWidget dependant.
2003/04/26 09:53- rwmcfa1
* GtkSpell/Spell.pm: first pass of documentation
* GtkSpell/GtkSpell.xs: gtkspell_get_from_text_view now accepts and
ignores a class
* Gtk2/t/10.GtkProgressBar.t, Gtk2/t/11.GtkStatusBar.t: first passes
at testing the two modules
* Gtk2/maps: hand added a (maybe temporary) entry for Gtk2::Progress
* Gtk2/xs/GtkProgressBar.xs: removed a bunch of deprecated functions
* Gtk2/xs/GtkMenuItem.xs, Gtk2/xs/GtkRadioButton.xs,
Gtk2/xs/GtkRadioMenuItem.xs, Gtk2/xs/GtkToggleButton.xs: now new with
string new('string') uses mnemonic instead of label. seems like a good
idea. you can use new_with_label if you don't want this behavior. also
impelented news with aliases for new and new_with_mnemonic (saves code
space/copying).
* Gtk2/xs/GtkCheckButton.xs, Gtk2/xs/GtkCheckMenuItem.xs,
Gtk2/xs/GtkImageMenuItem.xs: same as ^ with the addition of: wrappers
for _new_with_label were invalid they're now fixed.
2003/04/25 23:01 rwmcfa1
* GtkSpell/: Initial import of working (for me anyway) GtkSpell
mappings, give them a try.
2003/04/25 18:17 muppetman
* Gtk2/gtk-demo/changedisplay.pl, Gtk2/gtk-demo/drawingarea.pl,
Gtk2/gtk-demo/editable_cells.pl: more demo pieces. changedisplay.pl
is completely untested, because i do not have Gtk+ 2.2 on my
development machine.
* G/GClosure.xs: always copy SVs that are to be stored. see the
perlcall manpage. this fixes some intermittent bugs that happen
when reusing the same variable for various objects.
* Gtk2/xs/GdkWindow.xs: implemented gdk_window_invalidate_rect
* new-gtk2-perl.html, G/GBoxed.xs, helpers/genstuff.pl, G/gperl.h,
G/GObject.xs: renamed gperl_register_class to gperl_register_object
to be more consistent (with gperl_register_fundamental and
gperl_register_boxed). also added
gperl_object_set_no_warn_unreg_subclass, made gperl_get_object
honor it.
* Gtk2/xs/GdkGC.xs, Gtk2/xs/GtkStyle.xs: set 'no warn for unregistered
subclasses on GtkStyle and GdkGC. causes the type system to stop
spewing messages on stderr about unregistered types from theme
engines and gdk backends.
* Gtk2/gtk2perl.h, Gtk2/examples/scribble.pl, Gtk2/xs/GdkEvent.xs,
Gtk2/xs/GdkTypes.xs, Gtk2/xs/GtkCellRenderer.xs: completely reworked
the handling of GdkRectangle. it's now treated as a perl list
instead of an opaque type; this is consistent with gtk-perl and
makes life easier for manipulating the rectangles in perl.
* Gtk2/xs/GtkWidget.xs: changed handling of GtkAllocation, since
handling of GdkRectangle changed. invisible from the perl side.
also implemented gtk_widget_get_events.
* Gtk2/xs/GtkImage.xs: implemented gtk_image_new_from_pixmap,
gtk_image_set_from_pixmap, and gtk_image_get_pixmap
* Gtk2/xs/GtkLabel.xs: allow Label->new to default to NULL for
creating empty labels.
* Gtk2/xs/GtkTreeModel.xs: implemented gtk_tree_path_get_indices
2003/04/25 12:40 rwmcfa1
* Gtk2/t/1.GtkWindow.t: corrected check for gtk >= 2.2
* Gtk2/t/5.GtkListStore-etc.t: if 2.2 then try the reorder function
* Gtk2/xs/GtkCurve.xs: re-did set_vector to be clearer and removed a
unused param name from prototype.
* Gtk2/xs/GtkFileSelection.xs: clarified the a for loop's operations
in get_selections
* Gtk2/xs/GtkTreeStore.xs, Gtk2/xs/GtkListStore.xs: added itital pass
at _store_reorder
* Gtk2/maps: added stuff new as of gtk2.2.1
* Gtk2/xs/GtkPlug.xs Gtk2/xs/GtkWindow.xs Gtk2/xs/GdkCursor.xs
Gtk2/xs/GdkDrawable.xs Gtk2/xs/GtkInvisible.xs Gtk2/xs/GtkMenu.xs:
uncommented stuff new to gtk 2.2 now that maps is up to date for 2.2.1
2003/04/24 18:24 muppetman
* Gtk2/gtk-demo/appwindow.pl, Gtk2/gtk-demo/button_box.pl,
Gtk2/gtk-demo/colorsel.pl, Gtk2/gtk-demo/dialog.pl: more pieces of
the demo
* Gtk2/xs/GtkToolbar.xs: implemented all the append/prepend/insert
functions that were left out because they require callbacks
* Gtk2/xs/GtkStock.xs: implemented gtk_stock_add
* G/GSignal.xs, G/gperl.h: export gperl_signal_connect, the actual
workhorse, so other XSubs can use it
* Gtk2/xs/GtkColorSelection.xs: fixed get_current_color and
get_previous_color
* Gtk2/xs/GtkColorSelectionDialog.xs: member access
* TODO: high-level things (i forgot to commit this last week)
2003/04/24 17:55 rwmcfa1
* Gtk2/t/5.GtkListStore-etc.t: fixed type-o, datam -> data that was
preventing entries being made into the list
* Gtk2/xs/GtkMessagedialog.xs: removed a TODO, it is probably better
to just pass the message as you want it rather than use the varargs
stuff anyway
2003/04/24 13:30 muppetman
* Gnome2/xs/GnomeProgram.xs, Gnome2/druid.pl: implemented object
properties on Gnome2::Program->init (was a FIXME)
* G/GObject.xs: created alias get_property for get and set_property
for set, since some objects mask the G::Object-level method with
their own. in list_properties, don't die if the descr isn't set.
2003/04/24 12:55 muppetman
* helpers/genstuff.pl: wrap generated code in #ifdefs to support
versioning (typemaps must be generated even if the code isn't, so
we generate everything but only use part of it)
2003/04/23 10:44 rwmcfa1
* Gtk2/xs/GtkHButtonBox.xs, Gtk2/xs/GtkVButtonBox.xs:
get_spacing_defaults should accept and ignore class
* Gtk2/gtk2perl.h: type-o newSVGdkModiferType -> newSVGdkModifierType
* Gtk2/xs/GtkFontSelection.xs: type functions should not be in XS, it's
all automagical
* Gtk2/xs/GtkWindow.xs: a first pass at set_icon_list_function added
2003/04/23 10:37 muppetman
* ChangeLog: new change log, from the cvs log on sourceforge.
please keep it updated.
2003/04/22 muppetman
* G/GBoxed.xs: updates to debugging output
* G/GClosure.xs: don't keep the supplemental arguments array in
gperl_closure_marshal --- just put mortal values on the stack and
everything works out fine. simplifies the code quite a bit, and
removes some subtle and nasty bugs.
* G/GType.xs: GPERL_TYPE_SV, a boxed wrapper for perl scalars, mapped
to the package G::Scalar. this is handy for storing hashes and other
perl data structures in a TreeModel.
* G/GValue.xs: special handling for GPERL_TYPE_SV. implement handling
for G_TYPE_INTERFACE.
* Gtk2/gtk-demo/stock_browser.pl, Gtk2/xs/GtkIconFactory.xs: stock
browser demo and some supporting code
* Gtk2/gtk2perl.h, Gtk2/xs/GdkTypes.xs: special handling for
GdkModifierType flags... GDK_MODIFIER_MASK matches all of the flag
values, and causes nasty problems when you try to convert the SV flags
wrapper *back* to C. so, handle it separately, and don't allow
GDK_MODIFIER_MASK to make it into perl from C. (can still go the other
way, though)
* Gtk2/xs/GtkItemFactory.xs: bracket callback with ENTER/SAVETMPS &
FREETMPS/LEAVE
* Gtk2/xs/GtkMenu.xs: implemented gtk_menu_popup
* Gtk2/xs/GtkTreeModel.xs: it's possible for gtk_tree_model_iter_next
to return NULL
* Gtk2/xs/GtkTreeView.xs: implemented
gtk_tree_view_insert_column_with_data_func
* Gtk2/xs/GtkTreeViewColumn.xs: implemented
gtk_tree_view_column_set_cell_data_func
* Gtk2/xs/GtkWidget.xs: default params on set_size_request, and
implemented render_icon
2003/04/22 rwmcfa1
* Gtk2/t/5.GtkListStore-etc.t, Gtk2/t/6.GtkLabel.t,
Gtk2/t/7.GtkBoxes.t, Gtk2/t/8.GtkCombo.t: initial import
* Gtk2/t/9.GtkRadioButton.t: test out a little more throughly
* Gtk2/xs/GtkFileSelection.xs: first pass at
gtk_file_selection_get_selections, needs to be utf8
tested/implemented maybe
* Gtk2/xs/GtkRadioButton.xs: crash bug, need to make sure that svp
exists
* Gtk2/xs/GtkRadioMenuItem.xs: first pass at implementing this class,
not tested yet
2003/04/21 muppetman
* G/G.xs, G/gperl.h: added gperl_alloc_temp
* G/GClosure.xs: added GPerlCallback, and made some robustness fixes
for GPerlClosure. (there was bizarre stuff happening in the
marshaller, wrong number of items in the supplemental arguments
array.)
* G/gperl.h: add GPerlCallback, with docs, and some other noise
* Gtk2/gtk-demo/item_factory.pl: gtk-demo driver for GtkItemFactory
* Gtk2/xs/GtkItemFactory.xs: implemented GtkItemFactory
* Gtk2/xs/GtkWidget.xs: set_flags and unset_flags
2003/04/21 rwmcfa1
* Gtk2/xs/GtkRadioButton.xs: implement all of the functions now with a
first pass at how to deal with GSList/group stuff. (notice: it's
subject to change)
2003/04/20 rwmcfa1
* just about every file: added cvs Header: tags
* Gtk2/xs/GtkWindow.xs: missing functions added; version 2.2 functions
added (some commented out); decorated_window functions added, but
commented out (how should we go about these?)
2003/04/18 gthyni
added .cvsignore files for cleaner updates
2003/04/18 muppetman
* Gnome2/druid.pl: add a button to test out Gnome2::About
* Gnome2/xs/GnomeAbout.xs: patch from Chas Owens to implement the
authors and documenters parameter lists, and allow defaults on
parameters following authors. (authors is required by gnome_about_new)
* Gtk2/xs/GtkAccelGroup.xs: implemented gtk_accelerator_parse and
gtk_accelerator_name
* Gtk2/xs/GtkFileSelection.xs: patch from Chas Owens giving access
to member widgets
* Gtk2/xs/GtkImage.xs: implemented new_from_pixbuf and set_from_pixbuf
* Gtk2/xs/GtkStock.xs: implemented some stock handling stuff, enough
to get the stock_browser demo working
* Gtk2/xs/GtkTreeSelection.xs: implemented
gtk_tree_selection_get_selected
2003/04/18 rwmcfa1
* Gtk2/xs/GtkWindow.xs: add a header tag and see if i can commit,
header tags need to be added to everything before to long.
2003/04/17 muppetman
massive commit of changes made by muppetman and rwmcfa1 since the
last pre-sourceforge snapshot.
* G/G.xs, G/gperl.h: added gperl_croak_gerror, takes care of properly
freeing a GError before croaking with the message it contains
* G/GObject.xs: don't allow non-RVs in gperl_get_object;
added G::Object->eq
* G/gperl.h: added gperl_croak_gerror, takes care of properly freeing
a GError before croaking with the message it contains
* Gtk2/gdk.typemap, Gtk2/gtk2perl.h: custom handling for GdkBitmap
* Gtk2/examples/layout.pl: new example (ported from C by ross)
* Gtk2/examples/socket.pl, Gtk2/t/0.Gtk2.t, Gtk2/t/1.GtkWindow.t,
Gtk2/t/2.GtkButton.t, Gtk2/t/3.GtkGammaCurve.t: updates since snapshot
* Gtk2/gtk-demo/textview.pl, Gtk2/gtk-demo/floppybuddy.gif,
Gtk2/gtk-demo/gtk-logo-rgb.gif: another piece of gtk-demo, and
accessories
* Gtk2/xs/GdkPixbuf.xs: added lots of missing functionality
* Gtk2/xs/GdkPixmap.xs: can now create GdkBitmaps. warning, GdkBitmap
may be in the wrong class, there are some complaints at runtime about
it (porting C code passing a GdkBitmap where a GdkPixmap was wanted,
i had to re-bless in perl to achieve that).
* Gtk2/xs/Gtk2.xs: added version information functions
* Gtk2/xs/GtkCurve.xs: work around a C bug that the Gtk+ maintainers
say won't be fixed (because the widget is to be removed in future
versions)
* Gtk2/xs/GtkHScale.xs, Gtk2/xs/GtkHScrollBar.xs, Gtk2/xs/GtkVScale.xs,
Gtk2/xs/GtkVScrollBar.xs, Gtk2/xs/GtkScrolledWindow.xs: allow default
parameters
* Gtk2/xs/GtkListStore.xs: place TreeModel at the beginning of ISA so
Gtk2::TreeModel::get is found before G::Object::get. remove some
warn()s.
* Gtk2/xs/GtkNotebook.xs: patch from Emmanuele Bassi, which was a
double commit, because goran had already fixed it and i didn't
notice in time
* Gtk2/xs/GtkTextBuffer.xs: implemented insert_with_tags_by_name
* Gtk2/xs/GtkTextIter.xs: allow NULL return from get_child_anchor
* Gtk2/xs/GtkTextView.xs: oops, bad signature
* Gtk2/xs/GtkTreeView.xs: implemented missing new_with_model
* Gtk2/xs/GtkTreeViewColumn.xs: stack randomly had the wrong number
of items. converted to a macro to avoid the use of dXSARGS, and
suddenly the list_store.pl portion of gtk-demo works. dXSARGS is
supposed to work anywhere; i don't understand why it didn't.
* Gtk2/xs/GtkWidget.xs: implemented several _modify_* methods
* Gtk2/xs/GtkWindow.xs: allow window type to default to 'toplevel'
on ->new; all NULL for several other functions' parameters.
* Gtk2/xs/PangoFont.xs: pango constants.
2003/04/17 gthyni
imported 20030415 snapshot into CVS