November 2011 - webkit-efl - lists.webkit.org

Re: [webkit-efl] RFC: Changes to the ewk_settings_cache_* API
by Ryuan Choi 01 Jan '12

01 Jan '12

I like what you want. I have one minor points for the renaming. IMO, ewk_view_setting_offline_app_cache_{get,set} also should be considered. Best Regards, Ryuan Choi ------- Original Message ------- Sender : Raphael Kubo da Costa<kubo(a)profusion.mobi> Date : 2011-11-08 01:17 (GMT+09:00) Title : [webkit-efl] RFC: Changes to the ewk_settings_cache_* API The ewk_settings_cache_* family of functions is quite messy right now. Some calls deal with Offline Web Applications (ewk_settings_cache_directory_path_{get,set}) and others deal with in-memory cache settings (ewk_settings_cache_enable_{get,set} and ewk_settings_cache_capacity_set). I see the following issues with the current situation: 1) Offline Web Applications are completely unrelated to the in-memory cache. The former handles the appcache part of the HTML5 spec, wheareas the latter just caches stylesheets, images and scripts in memory. Yet they both share the same function prefix. The current apidox also does not help at all (the term "offline web applications" is not even mentioned). 2) The exposed API to control the memory cache is redundant and insufficient. - It is insufficient because one cannot pass all parameters accepted by MemoryCache::setCapacities(). It also does not clear other in-memory caches, such as the page and the font caches. - It is redundant because if the user could specify all the parameters accepted by MemoryCache::setCapacities() there would be no need to have the getter and setter for MemoryCache::disabled(). Please tell me what you think of the following proposal: - Rename ewk_settings_cache_directory_path_{get,set} to ewk_settings_offline_web_applications_cache_directory_{get,set}. - Follow more closely what the Qt API does and have: - ewk_settings_cache_clear(Ewk_Cache_Clear_Mode mode) Ewk_Cache_Clear_Mode is an enum which currently has only one value: enum Ewk_Cache_Clear_Mode { EWK_CACHE_CLEAR_MODE_MEMORY_CACHE }; This one clears the memory cache, the page cache, the font cache and the cross-origin preflight cache, like Qt does. Once we have the necessary code to have a network disk-cache, we can add another item to the enum to make it possible to clear it as well. - ewk_settings_object_cache_capacity_set(minDeadBytes, maxDeadBytes, totalBytes) Pass the received parameters to MemoryCache::setCapacities(). If they are all 0, disable the in-memory object cache. -- Raphael Kubo da Costa ProFUSION embedded systems http://profusion.mobi _______________________________________________ webkit-efl mailing list webkit-efl(a)lists.webkit.org http://lists.webkit.org/mailman/listinfo.cgi/webkit-efl

2 2

WebKit EFL Coding Style Change
by Gyuyoung Kim 30 Nov '11

30 Nov '11

1 0

Re: [webkit-efl] RFC: Changes to the ewk_settings_cache_* API
by Ryuan Choi 10 Nov '11

10 Nov '11

> > I like what you want. > > > > I have one minor points for the renaming. > > IMO, ewk_view_setting_offline_app_cache_{get,set} also should be considered. > > The abbreviation kind of makes sense, but there should be "path" or > "directory" somewhere in the name. I mean that ewk_view_setting_offline_app_cache_{get,set} should be changed properly when you change ewk_settings_cache_directory_path_{get,set} If we changed as you suggested at below, we don't need to touch ewk_view_setting_offline_app_cache_{get,set}. So please ignore this. For the list which you suggested, It's good to me. > What do you guys think of having the following set of calls in > ewk_settings from now on: > > == General purpose == > * ewk_settings_repaint_throttling_set > * ewk_settings_default_timer_interval_get > > == Web Database (aka WebSQL) == > * ewk_settings_web_database_default_quota_get > * ewk_settings_web_database_path_set > * ewk_settings_web_database_path_get > > * ewk_settings_web_database_default_quota_set > * ewk_settings_web_database_clear > > The last two are new functions. > > == Icon Database == > * ewk_settings_icon_database_path_set > * ewk_settings_icon_database_path_get > * ewk_settings_icon_database_clear > * ewk_settings_icon_database_icon_surface_get > * ewk_settings_icon_database_icon_object_add > > == Offline Web Apps == > * ewk_settings_offline_app_cache_path_get > * ewk_settings_offline_app_cache_path_set > > These two replace ewk_settings_cache_directory_path_{get,set}. > > * ewk_settings_offline_app_cache_max_quota_get > * ewk_settings_offline_app_cache_max_quota_set > > These functions are new. > > == In-Memory Cache == > * ewk_settings_object_cache_capacity_set > * ewk_settings_memory_cache_clear > > The former is a modified version of ewk_settings_cache_capacity_set, > and the latter is a new function. > > ewk_settings_cache_enable_{get,set} would be removed.

1 0

RFC: Changes to the ewk_settings_cache_* API
by Raphael Kubo da Costa 07 Nov '11

07 Nov '11

The ewk_settings_cache_* family of functions is quite messy right now. Some calls deal with Offline Web Applications (ewk_settings_cache_directory_path_{get,set}) and others deal with in-memory cache settings (ewk_settings_cache_enable_{get,set} and ewk_settings_cache_capacity_set). I see the following issues with the current situation: 1) Offline Web Applications are completely unrelated to the in-memory cache. The former handles the appcache part of the HTML5 spec, wheareas the latter just caches stylesheets, images and scripts in memory. Yet they both share the same function prefix. The current apidox also does not help at all (the term "offline web applications" is not even mentioned). 2) The exposed API to control the memory cache is redundant and insufficient. - It is insufficient because one cannot pass all parameters accepted by MemoryCache::setCapacities(). It also does not clear other in-memory caches, such as the page and the font caches. - It is redundant because if the user could specify all the parameters accepted by MemoryCache::setCapacities() there would be no need to have the getter and setter for MemoryCache::disabled(). Please tell me what you think of the following proposal: - Rename ewk_settings_cache_directory_path_{get,set} to ewk_settings_offline_web_applications_cache_directory_{get,set}. - Follow more closely what the Qt API does and have: - ewk_settings_cache_clear(Ewk_Cache_Clear_Mode mode) Ewk_Cache_Clear_Mode is an enum which currently has only one value: enum Ewk_Cache_Clear_Mode { EWK_CACHE_CLEAR_MODE_MEMORY_CACHE }; This one clears the memory cache, the page cache, the font cache and the cross-origin preflight cache, like Qt does. Once we have the necessary code to have a network disk-cache, we can add another item to the enum to make it possible to clear it as well. - ewk_settings_object_cache_capacity_set(minDeadBytes, maxDeadBytes, totalBytes) Pass the received parameters to MemoryCache::setCapacities(). If they are all 0, disable the in-memory object cache. -- Raphael Kubo da Costa ProFUSION embedded systems http://profusion.mobi

1 0

Does anyone use EDJ based custom cursor ?
by Ryuan Choi 03 Nov '11

03 Nov '11

Hello, WebKit/Efl guys. I'd like to know whether we keep an edj based cursor support. Now, WebKit/Efl support EDJ based cursor and Ecore_X based cursor. But, eve and our applications use default.edj which doesn't have cursor related code. IMO, providing basic cursor-sets is the role of platform library. and ecore provide Ecore_X_Cursor for X and Ecore_Win32_Cursor for Win32. So, I'd like to remove EDJ based cursor for the code clean up. If someone want to use different backend, I think that we'd better to support related ecore APIs like Ecore_Win32_Cursor. How do you think about it? If you are using EDJ based cursor or want to keep it, please let me know. Thanks Best Regards, Ryuan Choi ----------------------------------------------------------------------- Listen and listen and listen -----------------------+

2 1

HEADS-UP: ewk_view_setting_scripts_window_open_{get, set} have been renamed
by Raphael Kubo da Costa 01 Nov '11

01 Nov '11

Hey there. After SVN commit 98908, the ewk_view_setting_scripts_window_open_* family of functions has been renamed. - ewk_view_setting_scripts_window_open_get is now ewk_view_setting_scripts_can_open_windows_get. - ewk_view_setting_scripts_window_open_set is now ewk_view_setting_scripts_can_open_windows_set. Since r93833 we had the API in a weird state where the calls to (dis)allow scripts to close windows did not match the naming scheme of the calls to (dis)allow scripts to open windows, which was pretty weird. The naming scheme of the former made much more sense, so I've chosen to adapt the latter, older functions. -- Raphael Kubo da Costa ProFUSION embedded systems http://profusion.mobi

1 0