| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |
|
|
|
|
|
|
|
|
|
| |
This is the new name in preparation for deleting the backwards-compatible typedef.
BUG=
TEST=
Review URL: https://chromiumcodereview.appspot.com/10532162
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@143282 0039d316-1c4b-4281-b951-d872f2087c98
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
For interface Foo with versions A and B, structs will be defined for Foo_A and Foo_B, and a typedef generated from Foo_B to Foo. Versioning of IDL structs remains unchanged.
(Follow-up to discussion on CL 8931019)
*** This change breaks compatibility with C code that makes use of unversioned-named interface structs. :( ***
We need to define a versioned-named struct for each interface's current version. We could:
1. Carry on defining the current struct unversioned and typedef a versioned name for it.
The versioned type for the interface would be a typedef for the latest version and a struct for later versions, causing calling C code that uses it to break when a new version is added.
2. Define structs for all versions, and a separate unversioned struct.
This would lose type equivalence between the versioned and unversioned copies of the latest interface specification, and lead to needless duplication, especially for one-version interfaces.
3. Do this CL. We break some C code once, by change the unversioned type from struct to typedef, but we avoid these headaches in future. C++ code shouldn't be affected.
*** Contents of this CL, including notes on to-dos:
This change requires updating thunk-getters to be defined using versioned names, so that the interface structs can be forward declared; the thunk-getters now have versioned names and return values.
Changing the thunk-getter naming requires updating unit-tests to call the versioned names. It also requires some interface headers not generated from IDL to be manually updated to the new scheme (PPB_CharSet_Dev, PPB_Crypto_Dev, PPB_DirectoryReader_Dev, PPB_LayerCompositor_Dev, PPB_Graphics3D, PPB_Flash_Menu, PPB_Instance_Private, PPP_Pdf, PPB_Flash_NetConnector, PPB_GLESChromiumTextureMapping_Dev and PPB_Graphics3D_Trusted).
The proxy GetInfo() calls are updated to use versioned interface macros and thunks. Similarly, most PPBs added in interface_list.cc are now added using versioned interface macros and thunk getters. Ditto PluginModule, and some of PluginInstance. Some implementations (e.g. PPB_CharSet_Dev) needed updating to use versioned thunk getters to fetch interfaces to use.
The VideoDecoder interface size checks are for 32-bit are updated not to expect structs.
It was necessary to replace forward declarations of interface structs with includes, and remove "struct" prefixes where unversioned names were used. In most cases the affected code should really be updated to cope with versions.
PPP_Pdf has become PPP_Pdf_1. Other versionless structs that should be updated for consistenct include PPB_UMA_Private, PPB_GPU_Blacklist_Private, PPB_Proxy_Private, PPP_PDF, PPB_OpenGLES2, PPB_Flash_File_FileRef and PPB_Flash_File_ModuleLocal. Also PPP_Class_Deprecated, PPP_CursorControl_Dev, PPP_Find_Dev, PPP_NetworkState_Dev, PPP_Scrollbar_Dev, PPP_Selection_Dev, PPP_VideoCapture_Dev, PPP_Graphics3D and PPP_Instance_Private.
The Graphics2D and GLES2 examples now use unversioned interface type names without "struct" prefix. It's not clear whether examples should use versioned names, to show best practice, or unversioned.
The typedef hack in PPP_Instance IDL is gone. Yay.
BUG=107398,108379
TEST=All Pepper and NaCl tests should continue to pass on Chromium bots. NaCl SDK examples build correctly.
Review URL: http://codereview.chromium.org/8989006
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@116490 0039d316-1c4b-4281-b951-d872f2087c98
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
interfaces, and replaces it with a list of macros. When files want to know which Pepper interface names and structs there are, they define what they want to do with the macros, and then include the relevant files for the classes of interfaces they want (stable, private, dev).
This re-lands my previous change.
Original Review URL: http://codereview.chromium.org/7874002
This does not convert all the dev interfaces. I just did a few to keep the patch smaller. So there is still a lot of manual registration.
This fixes the previous design problem where we assumed one *_Proxy object == one interface. We have been hacking around this lately with duplicate GetInfo calls, but this doesn't work for PPP interfaces.
Now, a _Proxy object is just there to help keep things organized. One proxy can handle zero, one, or many interfaces, and this mapping is controlled by just one line in the interfaces file.
So for example, to add a new function to a new version of an interface with backward compatibility, you would add that function to the _api.h file, and write a thunk for the new interface. Then you only need to add one line to the interfaces_ppb_public_stable.h file and that will be hooked up with the proxy and the implementation.
This removes some _proxy objects/files that were used only to declare that the interfaces existed, since they're no longer necessary.
I folded Console into the Instance API which removed a bunch of code.
I removed FileChooser 0.4. I think everybody has converted to the new one, and I think parts of it weren't even hooked up properly anymore.
Review URL: http://codereview.chromium.org/7887001
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@100936 0039d316-1c4b-4281-b951-d872f2087c98
|
| |
|
|
| |
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@100854 0039d316-1c4b-4281-b951-d872f2087c98
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
interfaces, and replaces it with a list of macros. When files want to know which Pepper interface names and structs there are, they define what they want to do with the macros, and then include the relevant files for the classes of interfaces they want (stable, private, dev).
This re-lands my previous change.
Original Review URL: http://codereview.chromium.org/7740038
This does not convert all the dev interfaces. I just did a few to keep the patch smaller. So there is still a lot of manual registration.
This fixes the previous design problem where we assumed one *_Proxy object == one interface. We have been hacking around this lately with duplicate GetInfo calls, but this doesn't work for PPP interfaces.
Now, a _Proxy object is just there to help keep things organized. One proxy can handle zero, one, or many interfaces, and this mapping is controlled by just one line in the interfaces file.
So for example, to add a new function to a new version of an interface with backward compatibility, you would add that function to the _api.h file, and write a thunk for the new interface. Then you only need to add one line to the interfaces_ppb_public_stable.h file and that will be hooked up with the proxy and the implementation.
This removes some _proxy objects/files that were used only to declare that the interfaces existed, since they're no longer necessary.
I folded Console into the Instance API which removed a bunch of code.
I removed FileChooser 0.4. I think everybody has converted to the new one, and I think parts of it weren't even hooked up properly anymore.
Review URL: http://codereview.chromium.org/7874002
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@100851 0039d316-1c4b-4281-b951-d872f2087c98
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
for Pepper interfaces, and replaces it with a list of macros. When files want to know which Pepper interface names and structs there are, they define what they want to do with the macros, and then include the relevant files for the classes of interfaces they want (stable, private, dev).
This does not convert all the dev interfaces. I just did a few to keep the patch smaller. So there is still a lot of manual registration.
This fixes the previous design problem where we assumed one *_Proxy object == one interface. We have been hacking around this lately with duplicate GetInfo calls, but this doesn't work for PPP interfaces.
Now, a _Proxy object is just there to help keep things organized. One proxy can handle zero, one, or many interfaces, and this mapping is controlled by just one line in the interfaces file.
So for example, to add a new function to a new version of an interface with backward compatibility, you would add that function to the _api.h file, and write a thunk for the new interface. Then you only need to add one line to the interfaces_ppb_public_stable.h file and that will be hooked up with the proxy and the implementation.
This removes some _proxy objects/files that were used only to declare that the interfaces existed, since they're no longer necessary.
I folded Console into the Instance API which removed a bunch of code.
I removed FileChooser 0.4. I think everybody has converted to the new one, and I think parts of it weren't even hooked up properly anymore.
Review URL: http://codereview.chromium.org/7740038
TBR=brettw@chromium.org
Review URL: http://codereview.chromium.org/7844018
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@100754 0039d316-1c4b-4281-b951-d872f2087c98
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
interfaces, and replaces it with a list of macros. When files want to know which Pepper interface names and structs there are, they define what they want to do with the macros, and then include the relevant files for the classes of interfaces they want (stable, private, dev).
This does not convert all the dev interfaces. I just did a few to keep the patch smaller. So there is still a lot of manual registration.
This fixes the previous design problem where we assumed one *_Proxy object == one interface. We have been hacking around this lately with duplicate GetInfo calls, but this doesn't work for PPP interfaces.
Now, a _Proxy object is just there to help keep things organized. One proxy can handle zero, one, or many interfaces, and this mapping is controlled by just one line in the interfaces file.
So for example, to add a new function to a new version of an interface with backward compatibility, you would add that function to the _api.h file, and write a thunk for the new interface. Then you only need to add one line to the interfaces_ppb_public_stable.h file and that will be hooked up with the proxy and the implementation.
This removes some _proxy objects/files that were used only to declare that the interfaces existed, since they're no longer necessary.
I folded Console into the Instance API which removed a bunch of code.
I removed FileChooser 0.4. I think everybody has converted to the new one, and I think parts of it weren't even hooked up properly anymore.
Review URL: http://codereview.chromium.org/7740038
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@100748 0039d316-1c4b-4281-b951-d872f2087c98
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This is more consistent with the stuff in shared_impl, and removes a lot of
namespace using goop.
Add a unified resource tracker shared between the proxy and the impl.
This renames the old ResourceObjectBase to Resource and removes the old
PluginResource. It moves the resource tracker from the impl to the
shared_impl, and makes the proxy use it.
Some things become a little less neat because there's no proxy resource base
class. In particular GetDispatcher() is now gone. I considered whether to
add a helper base class that provides this function, but decided against it
and had individual resource classes implement this when their implementation
would find it useful. This is because ultimately I want more of this
functionality to move into the shared_impl, and it's easier to do that if
there are fewer proxy-specific things in the resources.
This changes the way that plugins are added to the tracker. Previously they
would only be in the tracker if the plugin had a reference to them, although
they could be alive if the impl had a scoped_ptr referencing an object. This
actually has the bug that if we then give the resource back to the plugin,
it wouldn't be refcounted properly and everything would get confused.
Now the tracker tracks all live resource objects whether or not the plugin
has a ref. This works basically like the var tracker (it would be nice if
the var and resource trackers shared more code, but that would further
complicate this already overcomplicated patch). The resource tracker takes an
extra ref whenever the plugin has one or more, and otherwise just tracks live
resources.
BUG=
TEST=
Review URL: http://codereview.chromium.org/7655002
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@97367 0039d316-1c4b-4281-b951-d872f2087c98
|
|
|
I added a couple of tests while I was figuring out the ref counting problem I was having. Also lots of bonus spelling fixes.
BUG=86123
TEST=ppp_messaging_proxy_test and ppapi_tests
Review URL: http://codereview.chromium.org/7531003
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@95240 0039d316-1c4b-4281-b951-d872f2087c98
|
