[seed] Created tag seed-0.5
- From: Robert Carr <racarr src gnome org>
- To: svn-commits-list gnome org
- Subject: [seed] Created tag seed-0.5
- Date: Thu, 16 Apr 2009 14:17:48 -0400 (EDT)
The unsigned tag 'seed-0.5' was created.
Tagger: Robert Carr <racarr svn gnome org>
Date: Thu Apr 16 14:16:49 2009 -0400
Tag 0.5 release.
Changes since the last tag 'seed-0.3.1':
Robert Carr (207):
-Wall...apparently this got switched off at some point.
Add seed_gobject_named_constructor
Named constructor rework:
Update examples to the new constructor stuff, etc...
Fix more examples, and the Gio extension.
More warnings.
Merge branch 'struct-prototypes'
Avoid defining methods on struct constructors, as this is wrong. Shaves
Add GObject propery installation support.
Refactor some of the property handling code in order to prepare for
Simple evaluation API.
Continue to pound on memory usage.
Memory...
Make lightsoff work with more versions of Clutter.
Add the basic code to handle property getter/setters.
Add support for GType arrays.
Add support for GValue arguments, with a heuristic for giving the type,
Add a simple GtkTreeView example..and some small changes to make it
Lightsoff is now hosted in gnome-games SVN.
Fix incorrect definition of an additional (protected nonetheless...)
Fix bad owner_type v. value_type in property getters.
Add clutter COGL example.
Clean up the clutter examples a bit.
More.
I could play with clutter for hours...
More example cleanups.
Base memory usage is down another 100k or so.
Fix a hypothetical base info reference leak (if one day we were to allow
Funny example of GIO + JSON to pretty print the reddit home page.
Add support for GI_TYPE_TAG_ARRAY in return.
Fix GSList/GList segfaulting with GI_TYPE_TAG_VOID
Make returned GList/GSlist real arrays as opposed to objects with
Add support for GI_TYPE_TAG_FILENAME.
Fix leak of GI_TYPE_TAG_ARRAY in.
GClosure argument releasing.
Tiiiiny leaks in builtins.
If argument type conversion fails, out_values array is not freed.
Don't try and convert null GI_TYPE_TAG_ARRAY out arguments.
Change enums to Namespace.Enum.BIGVALUEISBIG
Update tests to enum changes.
Example fixes
Fix the rest of the examples
Fix Soup example, and consistency with struct constructors/object
Rename instance_init to init
Update tests to remove some deprecated stuff.
Fix GI_TRANSFER_CONTAINER for GI_TYPE_TAG_ARRAY
Update clutter shader example to make it run again
Add support for gfloat arrays.
Refactor some array code
Add support for UINT8 arrays
Add opengl-glib girs in Seed (this will be split in to it's own magical
Add OpenGL examples.
Add async quine example
Add Ryan Pauls twitter example, with a few modifications to make the HTTP request async
Install OpenGL typelibs to the new directory
Add GLU gir/typelib
Some annotations to GLU gir
Add glxgears port
Some small fixes to gears rendering
Yay, all the gears render correctly now
Fix twitter example makefile
Add GdkPixbufLoader example
Example cleanup and maintenance
Install OpenGL girs to the correct directory
Add Profile images to twitter example
Make twitter example download images async
Update GLTypes gir to include GLhandleARB
Mode types to GLTypes-1.0.gir required for clutter-0.9 to build
Add the annotation files for the OpenGL girs
Implement seed_value_is_null seed_value_is_object and seed_object_is_function
Add seed_script_new_from_file
AHHHHH
I suck at version control
Lots of cleanup, standardize on GNU coding style
Fix readline bindings Makefile
Remove gtype-signal-gvalue.js test as it is nonsense (Note GObject.TYPE_VALUE does not exist)
Fix missing seed_union_set_property implementation.
Fix longstanding seed_gi_argument_make_js bug with handling enums, sometimes broke things in getting struct properties struct-nested-set.js passes now
seed_gobject_signal_emit needs to initialize ret_value to query.return_type, fixes two tests
hello_cb in api-js-signal-form-c was smashing the stack
Merge branch 'master' of ssh://git.gnome.org/git/seed
api-js-signal-from-c needs to use double as the return value if it wants to strictly check for equality (Given JSCore is always in double precision
Avoid initializing return GValue in case of G_TYPE_NONE (also avoid unsetting it leading to an unpleasant message on the console)
Change the two tests that use Clutter to use Pango and GTK instead
seed_print doesn't need to use printf
Some canvas updates
Some more work on Canvas...it's a little broken right now but in the middle of an essential rewrite that should let state work properly
Merge branch 'master' of ssh://git.gnome.org/git/seed
cleanup canvas tests a bit...more work on stage
Initialize default stroke/fill colors when making a new SeedCanvasStyle
Revert "Patch from Matt to add global include paths."
Support some more primitive array types
Fix potential crash (or confusion) in seed_value_to_string
Add an assertion in seed_value_to_gobject, should be helpful in catching a few memory bugs that sometimes pop up
Switch GObject reference to the JS ref to use g_object_set_qdata, low hanging optimization
Remove incredibly dubious setting of SIGCHLD
Switch seed_prepare_closure to use g_callable_info_prepare_closure...drops a lot of code
Merge branch 'master' of ssh://git.gnome.org/git/seed
Big build system reorganization.n. Also add shave for pretty builds
Fix warning in C test
Add seed_signal_connect and seed_signal_connect_full C convenience API
Fix indentation on C tests
Add seed_context_get_global function to API
Style fixes in seed-api.c
More gtk-doc
Some fixes to canvas color parsing
Merge branch 'master' of ssh://git.gnome.org/git/seed
Validate enum types in property setting
Add validation of enum types to the GArgument conversion logic
Minor cosmetic improvements to seed-types.c
Add Seed.breakpoint for debugging purposes
Add seed_init_with_context_group
Big signal rework. Should save some memory. Dynamically create signal objects, also smaller signal privates struct...fix a leak...all kinds of good stuff
Add scoped includes
Begin work on the new importer
Importer now supports enums and toplevel functions
Importer now supports GI_INFO_TYPE_OBJECT
Importer supports GI_INFO_TYPE_STRUCT
Importer supports GI_INFO_TYPE_UNION
Support GI_INFO_TYPE_CONSTANT, and some refactoring
Protect against multiple imports
Add support for versions
Fix build
Work on start of importer for normal files
Fix dumb person bug
Make importer cache imports for files properly
Free the search path
Some work on nested directory imports
Make it actually work
Merge branch 'importer'
Importer supports native modules
Add a default search path
Begin to update tests
Properly set Namespace.Object.prototype in the new importer
Make importer work for including the JS extensions to GI modules
Tests still don't all pass
Change searchPath to include '.' last
Fix the rest of the tests
Fix enum validation to not break flags
Add some more SEED_NOTE's
Start randomly updating a few examples
Update some more examples to new importer
Update more examples
Update twitter example
A bunch of more minor examples updated
Update shader example
Add SEED_DEBUG_IMPORTER and some debugging statements
Use the correct context for adding globals to new import contexts
Some comments on the new importer
Add a high level description/spec of the importer object
Fix shadowed variable declaration causing the Seed global object to be wrong sometimes with the new importer
Fix some more context confusion, and add a seed_prepare_global_context function
Some more importer cleanup stuff
Remove gratuitous protection
Remove old comment
Wheeeeeeeeeeeeeeeeeeeeeeeeeeeeee
Start porting the browser example
Start fleshing out the os module a little bit...aiming for identical to the python one as much as possible
Add ctermid
Add getegid and geteuid
Add seed_make_array to the API
Add os.getgroups
Add os.getlogin
Add os.getpgid
Add os.getpgrp
Add getpid and getppid
Add os.getenv
Add os.putenv
Have class_init callbacks use g_object_info_get_class_struct
Remove some deprecated usage of Seed.import_namespace in hardcoded strings that could trigger readline module to be imported twice
Fix the REPL script in main.c
Fix redundant requirement of GObject typelib
Add os.setegid os.setgid os.seteuid and os.setuid
Add os.strerror
Add os.umask
Add os.uname
Add os.unsetenv
Protect struct and union prototypes after putting them in the hash table...fixes a crash
Add a feature to debug builds that checks signal connection arity
seed_canvas_update_fill/stroke style should return TRUE
Fix type in canvas color parsing
Add tests for (unimplemented) compositing operators in seed canvas
Composite type tests was testing composite types that didn't exist
I don't understand canvas...some more work on handling state correctly
Add definitions of all the O_* defines to the OS module
Add os.open and os.close
Add os.dup, os.dup2, and os.fchmod
Update importer to support mutually inclusive files
Rough sketch of a port of the browser, could probably be reorganized a bit
Shader example works with new importer now
When evaluating a script in the importer, deifne the this object to be the global context of the script including the included script
Add os.fchown
Add os.fdatasync and os.fpathconf
Add os.fsync
Add os.ftruncate and os.isatty
Merge branch 'master' of ssh://git.gnome.org/git/seed
Some more tweaks to the canvas tests
Make the file passed to the seed binary available as imports.filename off the bat to prevent recursive running
Add os.lseek
Add os.openpty and os.pipe
Add os.read
Add os.write
Addos.ttyname os.tcsetpgrp and os.tcgetpgrp
Merge branch 'master' of ssh://git.gnome.org/git/seed
Some changelog updates, bump version number
Update pong example to use new imports
Add release notes
Tim Horton (238):
Big Lights Off cleanup.
Readd unnecessary show because of a clutter bug.
Continuation of last patch.
Didn't need the first part of last two patches.
Add documentation templates for the other modules (SQLite, multiprocessing, Canvas, and the example module).
Working on sqlite docs.
Finished SQLite module documentation.
Starting Canvas documentation.
Work on canvas module docs.
Enable profiling for some modules, simple SQLite test.
Put errors back into sqlite.
Added test for readline_bind.
Move tests into a subdir. Getting ready to add C tests (to test API, etc.)
One addition to last patch.
Add the new tests from earlier to automake.
Found some more /usr/local/bin/seed garbage.
64-bit build fix.
Starting to make example installation optional.
Add clutter transitions library to SVN; going to make a demo script in an hour or two.
Finished moving LO out of SVN.
Transition driver script.
Preliminary support for C tests.
Fix building.
C Tests don't disable asserts (yay!) and stuff now.
Integrate C tests into the little python test script.
Rename C tests, cleanup. Done for now.
ROBB! It's /usr/bin/env seed...
A bunch of very simple type conversion tests with the Seed API.
Quick closure tests.
Sync with seed changes.
Start of Everything.gir tests.
Robb's fixing things that make more of Everything.gir tests work.
Still working on "Everything" test.
Robb fixed lists yesterday, so this test needed to be updated.
Automatically update README version number.
Use autoconf to insert the current seed version number into the documentation.
Fix optimization flag nonsense. (We only get one -O flag now; 0 for profile or debug, 3 normally).
Something went horribly wrong with making the readme autoconf'd.
Remove broken autoconfd readme.
Readme doesn't need to be versioned, this fixes the need for autoconf magic.
Working on fixing build on Mac OS X. Still fails because it can't find libwebkit, of course.
Move readme back to autoconf'd so we get the version number (I understand my mistake now).
I give up with the autoconf readme thing. Broke matt's build again.
Fixed build on OS X. Robb has to fix the flat namespace thing, and then we're golden.
Fix seed build on OS X. All of libseed and most of the modules build, we fail on building turtle (because of lack of Gtk) now!
Only build turtle if we have Gtk.
Fix linux build.
Check for JSContextGroupCreate, and warn on too-old WebKit revisions.
Update readme for OS X installation.
Fix build on OS X, again.
Fix Gtk detection on Linux.
Move OS X specific instructions to a separate INSTALL file.
Users can now toggle the build of /any/ module with --[en|dis]able-*-module at configure time. Of course, if they don't have the required libraries, they don't get the module, too. Should leave more of a warning if readline isn't included.
Add warning if we're not building the readline module.
Some configure output cleanup.
Add ability to disable turtle example with configure.
Indentation update.
Use pkg-config to find Gtk for turtle example.
Add test for installed-from-JS signals on GTypes which have arguments.
Fix some error consistency problems in builtins, update test, add argument length test for spawn, etc.
Fix toggleable example installation.
Add makefile for clutter-transitions.
Add transitions to configure.
Remove first round of things that don't need to be in SVN
Removed one more file that doesn't need to be in SVN.
Fixing examples to match new enum typing.
Move old browser to broken, add new browser back (still has no features, though).
Build stuff for replaced browser.
More enum casing change.
Fix more enums
Add missing header to extra_dist.
Update OS X installation documentation.
Missed some macports packages.
Split new browser back into bunch of files.
Working on new browser.
Working on Browser. Due to a bug in installed properties (construction properties don't show up in init), no more for now!
Small style things.
Whoops @ Seed build. Fixes module installation, examples, etc.
added horribly dirty ugly pong in clutter. it's horrible. don't even open it. it's the worst code ever. it was more of a "rush to implement something really really fast" thing with Matt. I'll fix it tomorrow, add networking, scoring, fix accelerations and stuff, make the AI LOSE at some point, etc.
Some small style updates for examples, nothing huge.
Massive rework of pong beginning. Subclasses in your subclasses (I'm pretty sure this is the first time we've subclassed a subclass that was created in Seed...)
Pong textures!
Fix ball logic, etc. Lots of stuff is better now. Still not great. Hard to beat the AI.
Bunch of look-n-feel updates to pong, etc.
Working on making build work both on Fedora and Arch at the same time. Apparently, Arch and Gentoo don't have pc.s for libffi. Cool.
Working on Fedora build.
Fix Fedora build at the expense of Arch and Gentoo. Still working on it...
Reimplement open-in-new-window. Unfortuately installed properties don't work as construction properties (they aren't available in init) (Robb?), so this is a (hopefully) temporary hack to workaround that.
Statusbar, new setting, progressbar, reorderable tab view, etc. for browserrewrite.
Better fitting progress bar!
Fix readline from requiring complete seed rebuild each time you change seed.h
Add a missing file to makefile.
Update browser to use progress-in-GtkEntry, like Ephy/Safari. Shouldn't
Fix paths to Seed in OpenGL examples.
Oops. Pull out JSContextGroupRef check until we can fix it...
Trying to fix WebKit ContextGroupCreate check.
Only build opengl-glib stuff if we have GLUT. Should fix to be more careful, but it's a good start.
Update runtime docs to mention workaround for loading namespaces twice.
Make sure we clean typelibs.
Fix struct union enumerate test; Introspection update catches more members!
Tools folder; bring JSLint in, starting a script to run JSLint on stuff; need to find options to turn off to make it even remotely reasonable.
JSLint works with Seed code now, mostly! Rev 800!
Many of the tests now pass JSLint.
Fix one mistake I forgot to check for in last commit.
Fix one example to pass JSLint, fixed 'with', so it doesn't complain (the maintainer doesn't /believe/ in 'with', for good reason, but ... still ...)
Most of the examples pass JSLint now; JSLint now accepts "const".
ClutterCairo works now if you have both 0.8 and 0.9 installed.
Fix all of the Clutter examples to require Clutter 0.8.
Pong should also require Clutter 0.8.
More JSLint fixes. That's all, for tonight.
Fix missing semicolons in builtin-argument-length test.
Tests now give diffs, instead of just printing both outputs!
Fix object-info test on Jaunty. It's a bad test...
Add Debian packaging stuff here, since we're still maintaining it ourselves.
Update deps.
Allow --disable-opengl-glib, also check for all three headers (GL/GLU/GLUT) before building.
Remove opengl-glib, since it's now in its own GNOME SVN module.
Finally found the WebKit revno that we require.
Fix small warning with automake.
More type tests, module profiling is optional now
If you install a signal (from JS) which wants to return a GValue, and you return a JS number type, Seed segfaults (as you can see in this test. This shouldn't happen, but I'm not sure what should ...
Still segfaulting. One fault is in Seed.print, the other one will need to wait for me to get to a machine with a WebKit debug build.
Initial version of our same gnome clone (with clutter/svg/etc.)
There's a long debugging function here. It's going away but might be useful later. The game is playable now. Yay!
Game works now (forgot sliding left when you empty a row...)
Use clutter 0.9.
Same Seed. It works, but Clutter is broken. Or something. So uh. No animations at the moment.
Fix silly nested switch in seed-types, add test that doesn't work (something's broken with return types of Seed-installed signals).
Whoops forgot to add the test.
80 characters!
Oh wow Same Seed animations!
Better animation, etc.
Significantly better animation performance.
Lots and lots of Same Seed updates.
Oops forgot a file.
Fix exception in sameseed.
Woah timings are sometimes important.
Fix an exception in SS.
Much improved picking.
Picking directly after animation works now.
Horrible hack to make SS work on slower machines/with many, many tiles... will be much improved when we gain the ability to pass a timeline into the animate function.
Whoops random debugging stuff...
working on 0.4 release stuff.
Updating ChangeLog, etc.
ChangeLog - this is actually 0.5 not 0.4, and update SS to match clutter changes.
Consistent headers across the C files.
Headers have consistent "headers", too! Also, fix everything.js (forgot to update enum casing a long time ago...)
Indent session!
A note about a potential leak for Robb.
Holy crap so many changes.
Add SameSeed GtkBuilder UI file.
Didn't mean to add animate_with_timeline in the last push.
Don't depend on SVN.
Test formatting cleanup; lots of duplicated test headers.
Fix some tests I accidentally broke in the last one (line numbers in exceptions changed, etc.)
For some reason (upstream!) Clutter 0.9 GIRs don't have offset data for ClutterColor. Someone should fix this.
Use gnome-www-browser instead of gnome-open for profiling output.
Add a few tiny conventions.
Remove old copy of the browser (bad Git import, I assume...)
Add middle-click-open-in-new-tab to seed browser. This exposes a bug in something which causes a segfault, so it's disabled for now.
Add a file of open questions that I have. If anyone has answers, let me know!
Fix Twitter client so it responds to enter key as well as click.
Twitter client profile images are aligned right now.
Add seed_repl_expose C function. Call it, you're dropped to a REPL with the arguments passed to seed_repl_expose stored in the array Seed.debug_argv.
Fix seed_repl_expose output formatting.
Struct offset test change from earlier no longer applies. An extra annotation to clutter_color_parse was what actually broke things; I have a patch to fix it.
It seems like a good idea to keep the not-yet-applied patches that we create against upstream projects /somewhere/ besides in my email...
Matt rewrote Robb's Clutter "game" for Clutter 0.9, using the new animation API and such. Requires the clutter-color-out-annotation.patch as of commit time, but hopefully that'll go upstream. Now we have something that uses Clutter 0.9 without using GtkClutter (yay!!)
Whoops! Forgot to update Makefile for new clutter example.
Cleanup Matt's patch a bit.
Fix up some docs autotools stuff, add empty shell of a 'mapping' guide like Gjs has.
Add matt's clutter thing to changelog.
A patch from Matt adding persistent readline history.
Indent session!
Still slightly broken readline detection is now fixed - some things were manually linking against readline, which was wrong.
Fix two examples to conform to jslint.
More JSLint fixes.
Update changelog for today's business.
Removed open question:
Fix some autotools stuff so opengl and twitter examples get distributed in seed tarballs.
It's finally time to start working on Gtk-Doc! Here's a slim beginning.
More GtkDoc.
More GtkDoc. Also, apparently there are functions we have that aren't showing up in GtkDoc. Maybe because they're not in the seed-sections.txt or whatever? Anyway, goodnight.
Matt's type fixes, and a warning fix (use %Zd instead of %d to print size_t)
Long patch from Matt that changes lots of things to glib functions, removing most of the stdio includes so we don't accidentally use them, switching printf(size_t) to use %Zd instead of casting to an int, etc.
Docs cleanup. Added missing seed.h symbols, created a reasonable ordering, added subsections, etc.
Merge branch 'master' of git+ssh://git.gnome.org/git/seed
Patch from Matt to add global include paths. Can't set from JS, but can from C, and gets used as search path during Seed.include. Also, swapping between indent styles isn't cool.
Indentation.
Reapply cleaned up version of Matt's patch.
Fix some small lacks of use of g_*, etc
Fix matt's docs (still need documentation for the second function, too.)
Matt added documentation for seed_engine_get_search_path.
Don't put lines between parameter documentation or GtkDoc gets mad.
Updated docs (don't show JS* types in API documentation!)
Docs for seed_context_create.
Docs for seed_make_object.
Oops. seed_string_to_utf8_buffer doesn't /return/ a buffer!
seed_string_is_equal_utf8 docs.
80characters.
Don't need all these extra sections that we're simply not using. Maybe some other day, but for now we're only documenting the external API.
Lots of documentation updates. Also, do we have a typedef for JSClassDefinition?
GtkDoc is choking on long type names? Wrote a note, anyway.
protect and unprotect documentation.
Lots of *(un)ref docs.
Missed some return values.
Create .gitignore and add everything that's currently untracked. I'm pretty sure we've got stuff in the repo that shouldn't be, but that's OK for now. If we're ignoring anything we shouldn't be, ... oops.
Documentation for seed_object_call. This marks 100% of the functions in seed-api.c being documented. Still a long way to go.
I'm pretty sure that carrying the templates aroung in Git is unnecessary (the GtkDoc docs suggest so anyway...)
Starting to move documentation into different sections.
Mushing documentation around into sections, fixing headers that indent screwed up (as usual...) and stuff.
Other files that got updated by the last commit.
Some type documentation, mostly copy/paste.
Revert "Remove incredibly dubious setting of SIGCHLD"
Revert "Revert "Remove incredibly dubious setting of SIGCHLD""
Somehow I accidentally committed with 2 colors on.
Matt fixed up his Seed.set_search_path function, so we've enabled it.
Exception documentation. GtkDoc is losing some of my docs on seed_make_exception for some reason. Also, somehow some of our headers are out of sync (variable-name-wise) with implementation files.
Most of seed-types.c is documented now (most of the public part, anyway).
Patch from Matt:
Update debian packaging now that ogl-glib is no longer inside of seed.
Patch from Matt, fixes horrible crash in include path code.
One question resolved.
Patch patch to WebKit to make it introspectible. WK guys say that they're going to have it in for 1.1.4.
Yay webkit committed the patch.
Quiiick fixes to make Same Seed run again. However, there's an annotation missing in Clutter, so you'll have to wait some time to be able to run it yourself.
Update documentation to API changes.
Merge branch 'master' of git+ssh://git.gnome.org/git/seed
Update clutter 0.9 test to new imports style and match some changes in Clutter trunk.
Update changelog to lots and lots of changes
Update gitignore.
Small fix.
That release-checking script I've been talking about.
Add some files that were missing from the build.
Add them to configure.
Missing INSTALL.osx and gtype-signal-args tests in tarball.
%Zd is size_t, not %d...
Stopgap to get around the broken pipe test thing.
[
Date Prev][
Date Next] [
Thread Prev][
Thread Next]
[
Thread Index]
[
Date Index]
[
Author Index]
