Re: [E-devel] status of evas_pipe.c
On Fri, 3 Jun 2011 08:31:59 +0200 (CEST) Vincent Torri said: the README: --enable-pipe-render **DISABLED DUE TO BUGS** this enables a multiple-thread renderer that divides the rendering into N regions (1 per core) to speed up rendering in software when you have multiple cpu cores. as such its disabled by default but u can enable it.. but at your own risk README documentation pretty much tells you not to unless you love bugs :) eventually that code will go away, so worry about it then. > > > On Fri, 3 Jun 2011, Carsten Haitzler (The Rasterman) wrote: > > > On Fri, 3 Jun 2011 02:53:20 +0300 Tom Hacohen said: > > > >> On Thu, Jun 2, 2011 at 10:36 PM, Vincent Torri wrote: > >> > >>> > >>> Hey > >>> > >>> in evas_pipe.c, it is said that the code is deprecated and will go > >>> eventually. What about removing it from the build (but letting it in svn) > >>> ? > >>> > >> > >> i.e make it disabled by default (configure)? isn't it already? > > > > it is already off by default. has been for a while > > but it can be enabled. My remark was more to remove it completely (by > removing the configure option, for example) > > Vincent > -- - Codito, ergo sum - "I code, therefore I am" -- The Rasterman (Carsten Haitzler)ras...@rasterman.com -- Simplify data backup and recovery for your virtual environment with vRanger. Installation's a snap, and flexible recovery options mean your data is safe, secure and there when you need it. Discover what all the cheering's about. Get your free trial download today. http://p.sf.net/sfu/quest-dev2dev2 ___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
Re: [E-devel] status of evas_pipe.c
On Fri, 3 Jun 2011, Carsten Haitzler (The Rasterman) wrote: > On Fri, 3 Jun 2011 02:53:20 +0300 Tom Hacohen said: > >> On Thu, Jun 2, 2011 at 10:36 PM, Vincent Torri wrote: >> >>> >>> Hey >>> >>> in evas_pipe.c, it is said that the code is deprecated and will go >>> eventually. What about removing it from the build (but letting it in svn) >>> ? >>> >> >> i.e make it disabled by default (configure)? isn't it already? > > it is already off by default. has been for a while but it can be enabled. My remark was more to remove it completely (by removing the configure option, for example) Vincent -- Simplify data backup and recovery for your virtual environment with vRanger. Installation's a snap, and flexible recovery options mean your data is safe, secure and there when you need it. Discover what all the cheering's about. Get your free trial download today. http://p.sf.net/sfu/quest-dev2dev2 ___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
Re: [E-devel] status of evas_pipe.c
On Fri, 3 Jun 2011, Tom Hacohen wrote: > On Thu, Jun 2, 2011 at 10:36 PM, Vincent Torri wrote: > >> >> Hey >> >> in evas_pipe.c, it is said that the code is deprecated and will go >> eventually. What about removing it from the build (but letting it in svn) >> ? >> > > i.e make it disabled by default (configure)? isn't it already? afaics, no : evas_pipe.c is in the sources in Makefile.am, it is compiled if BUILD_PIPE_RENDER is defined, and that macro can be defined with --enable-pipe-render (and if pthread is available). Vincent -- Simplify data backup and recovery for your virtual environment with vRanger. Installation's a snap, and flexible recovery options mean your data is safe, secure and there when you need it. Discover what all the cheering's about. Get your free trial download today. http://p.sf.net/sfu/quest-dev2dev2 ___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
Re: [E-devel] E SVN: raster trunk/evas/src/lib/cache
On Thu, Jun 2, 2011 at 4:09 AM, Cedric BAIL wrote: > On Wed, Jun 1, 2011 at 9:27 PM, Rafael Antognolli > wrote: >> On Thu, Sep 9, 2010 at 4:08 AM, Enlightenment SVN >> wrote: >>> Log: >>> lots of cleanups... actually tested! :) >>> >>> >>> Author: raster >>> Date: 2010-09-09 00:08:10 -0700 (Thu, 09 Sep 2010) >>> New Revision: 52030 >> >>> static void >>> -_evas_preload_thread_end(Evas_Preload_Pthread_Data *pth) >>> +_evas_preload_thread_end(void *data) >>> { >>> - Evas_Preload_Pthread_Data *p; >>> + Evas_Preload_Pthread_Data *pth = data; >>> + Evas_Preload_Pthread_Data *p = NULL; >>> >>> - if (pthread_join(pth->thread, (void **) &p) != 0) >>> - return ; >>> - >>> - _evas_preload_thread = eina_list_remove(_evas_preload_thread, pth); >>> + if (pthread_join(pth->thread, (void **)&p) != 0) free(p); >>> } >> >> Hello raster, I know this is old, but I can't understand it... why do >> you only free the data struct on pthread_join error? Shouldn't you >> free it always? >> >> I can always get leaks on almost every call that I'm doing to >> edje_object_preload() on my program. I can send you a test case if >> necessary... > > Test case ! Test case ! Send it ! Send it ! Test case attached. It leaks for every thread started for preload. Just give images as arguments in the command line. It should leak something like this: ==10047== 64 bytes in 8 blocks are definitely lost in loss record 44 of 78 ==10047==at 0x4C267CE: malloc (in /usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so) ==10047==by 0x5302327: evas_preload_thread_run (evas_preload.c:197) ==10047==by 0x52FE624: _evas_cache_image_entry_preload_add (evas_cache_image.c:475) ==10047==by 0x5300199: evas_cache_image_preload_data (evas_cache_image.c:1270) ==10047==by 0xAE06CB6: ??? ==10047==by 0x52AEF16: evas_object_image_preload (evas_object_image.c:1013) ==10047==by 0x400EAF: _image_preloaded_cb (test-evas-preload.c:29) ==10047==by 0x52A06E6: evas_object_event_callback_call (evas_callbacks.c:222) ==10047==by 0x52BC270: evas_object_inform_call_image_preloaded (evas_object_inform.c:68) ==10047==by 0x52FE3EA: _evas_cache_image_async_end (evas_cache_image.c:419) ==10047==by 0x5301FC1: _evas_preload_thread_done (evas_preload.c:70) ==10047==by 0x52F72D1: evas_async_events_process (evas_async_events.c:90) -- Rafael Antognolli ProFUSION embedded systems http://profusion.mobi #include #include #include #define WIDTH 800 #define HEIGHT 600 static Evas_Object **pictures = NULL; static int npictures = 0; static int pic = 0; static void _image_preloaded_cb(void *data, Evas *e, Evas_Object *o, void *event_info) { evas_object_show(o); printf("loading pic %d of %d\n", pic + 1, npictures); if (++pic >= npictures) { ecore_main_loop_quit(); return; } evas_object_image_preload(pictures[pic], EINA_FALSE); } int main(int argc, const char *argv[]) { Ecore_Evas *ee; Evas *e; Ecore_Timer *timer = NULL; Evas_Object *picture; int i; if (argc <= 1) return 0; evas_init(); ecore_init(); ecore_evas_init(); ee = ecore_evas_new(NULL, 0, 0, WIDTH, HEIGHT, NULL); e = ecore_evas_get(ee); ecore_evas_show(ee); printf("trying to load %d images\n bytes = %zd\n", argc - 1, sizeof(Evas_Object *) * (argc - 1)); pictures = malloc(sizeof(Evas_Object *) * (argc - 1)); npictures = argc - 1; for (i = 1; i < argc; i++) { picture = evas_object_image_filled_add(e); evas_object_move(picture, 0, 0); evas_object_resize(picture, WIDTH, HEIGHT); evas_object_hide(picture); printf("loading: '%s'\n", argv[i]); evas_object_image_file_set(picture, argv[i], ""); evas_object_event_callback_add(picture, EVAS_CALLBACK_IMAGE_PRELOADED, _image_preloaded_cb, NULL); pictures[i - 1] = picture; } evas_object_image_preload(pictures[pic], EINA_FALSE); ecore_main_loop_begin(); ecore_evas_free(ee); free(pictures); ecore_evas_shutdown(); ecore_shutdown(); evas_shutdown(); return 0; } -- Simplify data backup and recovery for your virtual environment with vRanger. Installation's a snap, and flexible recovery options mean your data is safe, secure and there when you need it. Discover what all the cheering's about. Get your free trial download today. http://p.sf.net/sfu/quest-dev2dev2 ___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
Re: [E-devel] status of evas_pipe.c
On Fri, 3 Jun 2011 02:53:20 +0300 Tom Hacohen said: > On Thu, Jun 2, 2011 at 10:36 PM, Vincent Torri wrote: > > > > > Hey > > > > in evas_pipe.c, it is said that the code is deprecated and will go > > eventually. What about removing it from the build (but letting it in svn) > > ? > > > > i.e make it disabled by default (configure)? isn't it already? it is already off by default. has been for a while -- - Codito, ergo sum - "I code, therefore I am" -- The Rasterman (Carsten Haitzler)ras...@rasterman.com -- Simplify data backup and recovery for your virtual environment with vRanger. Installation's a snap, and flexible recovery options mean your data is safe, secure and there when you need it. Discover what all the cheering's about. Get your free trial download today. http://p.sf.net/sfu/quest-dev2dev2 ___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
Re: [E-devel] status of evas_pipe.c
On Thu, Jun 2, 2011 at 10:36 PM, Vincent Torri wrote: > > Hey > > in evas_pipe.c, it is said that the code is deprecated and will go > eventually. What about removing it from the build (but letting it in svn) > ? > i.e make it disabled by default (configure)? isn't it already? -- Tom. -- Simplify data backup and recovery for your virtual environment with vRanger. Installation's a snap, and flexible recovery options mean your data is safe, secure and there when you need it. Discover what all the cheering's about. Get your free trial download today. http://p.sf.net/sfu/quest-dev2dev2 ___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
[E-devel] status of evas_pipe.c
Hey in evas_pipe.c, it is said that the code is deprecated and will go eventually. What about removing it from the build (but letting it in svn) ? Vincent -- Simplify data backup and recovery for your virtual environment with vRanger. Installation's a snap, and flexible recovery options mean your data is safe, secure and there when you need it. Discover what all the cheering's about. Get your free trial download today. http://p.sf.net/sfu/quest-dev2dev2 ___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
Re: [E-devel] E SVN: cedric IN trunk/edje: . src/lib
On Tue, 31 May 2011 09:51:48 -0700 "Enlightenment SVN" said: cedric: just reverted this in svn. causes font classs bug. segv in e17. :) > Log: > edje: improve speed when deleting text member class. > > > Author: cedric > Date: 2011-05-31 09:51:48 -0700 (Tue, 31 May 2011) > New Revision: 59850 > Trac: http://trac.enlightenment.org/e/changeset/59850 > > Modified: > trunk/edje/ChangeLog trunk/edje/src/lib/edje_main.c > trunk/edje/src/lib/edje_private.h trunk/edje/src/lib/edje_util.c > > Modified: trunk/edje/ChangeLog > === > --- trunk/edje/ChangeLog 2011-05-31 16:43:40 UTC (rev 59849) > +++ trunk/edje/ChangeLog 2011-05-31 16:51:48 UTC (rev 59850) > @@ -110,3 +110,7 @@ > evas freeze and thaw calls. May help in certain pathological > situations. > > +2011-05-27 Cedric Bail > + > + * Improve performance by avoiding walking eina_list for nothing > + when deleting member_class. > > Modified: trunk/edje/src/lib/edje_main.c > === > --- trunk/edje/src/lib/edje_main.c2011-05-31 16:43:40 UTC (rev 59849) > +++ trunk/edje/src/lib/edje_main.c2011-05-31 16:51:48 UTC (rev 59850) > @@ -221,6 +221,8 @@ > eina_stringshare_del(cb->part); > free(cb); > } > + > + eina_hash_free(ed->members); > free(ed); > } > > > Modified: trunk/edje/src/lib/edje_private.h > === > --- trunk/edje/src/lib/edje_private.h 2011-05-31 16:43:40 UTC (rev > 59849) +++ trunk/edje/src/lib/edje_private.h 2011-05-31 16:51:48 UTC > (rev 59850) @@ -1004,6 +1004,8 @@ > int table_programs_size; > unsigned int table_parts_size; > > + Eina_Hash*members; > + > Edje_Perspective *persp; > > struct { > > Modified: trunk/edje/src/lib/edje_util.c > === > --- trunk/edje/src/lib/edje_util.c2011-05-31 16:43:40 UTC (rev 59849) > +++ trunk/edje/src/lib/edje_util.c2011-05-31 16:51:48 UTC (rev 59850) > @@ -3674,33 +3674,38 @@ > /* Get members list */ > members = eina_hash_find(_edje_text_class_member_hash, text_class); > > - /* Remove members list */ > - if (members) > - eina_hash_del(_edje_text_class_member_hash, text_class, members); > - > /* Update the member list */ > members = eina_list_prepend(members, ed); > > - /* Add the member list back */ > + /* Don't loose track of members list */ > + if (!ed->members) > + ed->members = eina_hash_string_small_new(NULL); > + eina_hash_set(ed->members, text_class, members); > + > + /* Reset the member list to the right pointer */ > if (!_edje_text_class_member_hash) > _edje_text_class_member_hash = eina_hash_string_superfast_new(NULL); > - eina_hash_add(_edje_text_class_member_hash, text_class, members); > + eina_hash_set(_edje_text_class_member_hash, text_class, members); > } > > void > _edje_text_class_member_del(Edje *ed, const char *text_class) > { > Eina_List *members; > + Eina_List *lookup; > > if ((!ed) || (!text_class)) return; > members = eina_hash_find(_edje_text_class_member_hash, text_class); > if (!members) return; > > - eina_hash_del(_edje_text_class_member_hash, text_class, members); > + lookup = eina_hash_find(ed->members, text_class); > > - members = eina_list_remove(members, ed); > - if (members) > - eina_hash_add(_edje_text_class_member_hash, text_class, members); > + if (!lookup) return ; > + > + eina_hash_del(ed->members, text_class, lookup); > + members = eina_list_remove_list(members, lookup); > + > + eina_hash_set(_edje_text_class_member_hash, text_class, members); > } > > void > > > -- > Simplify data backup and recovery for your virtual environment with vRanger. > Installation's a snap, and flexible recovery options mean your data is safe, > secure and there when you need it. Data protection magic? > Nope - It's vRanger. Get your free trial download today. > http://p.sf.net/sfu/quest-sfdev2dev > ___ > enlightenment-svn mailing list > enlightenment-...@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/enlightenment-svn > -- - Codito, ergo sum - "I code, therefore I am" -- The Rasterman (Carsten Haitzler)ras...@rasterman.com -- Simplify data backup and recovery for your virtual environment with vRanger. Installation's a snap, and flexible recovery options mean your data is safe, secure and there when you need it. Discover what all the cheering's about. Get your free trial download today. http://p.sf.net/sfu/quest-de
Re: [E-devel] [Patch] elm_progressbar patch
On Thu, 02 Jun 2011 12:44:12 + (GMT) PRINCE KUMAR DUBEY said: in elm_widget - see the signal emit hooks and signal emit handling. the infra to "rore and retrieve" the strings can be in elm_widget, with the text set hook handling directing the real text set to the appropriate text object and part based on specified part name. > Samsung Enterprise Portal mySingle > > how we are going to manage theme changability ? i mean to say how texts are > going to reset again on theme changes if we do not handle it at each widget > level. > > can you please provide your suggestions over it. > > > > --- Original Message --- > > Sender : Carsten Haitzler > > Date : Jun 02, 2011 14:28 (GMT+05:30) > > Title : Re: [E-devel] [Patch] elm_progressbar patch > > On Wed, 01 Jun 2011 09:47:01 + (GMT) PRINCE KUMAR DUBEY > said: > > yeah... all i can imagine being better is this being generic like signal emit > hooks > > > > >Hi Mr Park, > > > > > >Thanks for your comment. > > > > > >I already had discussion with Rasterman about this new API. > > > >Please visit this link for reference > >[1]http://sourceforge.net/mailarchive/message.php?msg_id=27444784. > > > > > >and this API would not make widget to break theme changing feature as we > > are handling this in _theme_hook(). > > > > > >Regards, > > > >Prince > > > > > > > >--- Original Message --- > > > >Sender : ChunEon Park > > > >Date : Jun 01, 2011 14:19 (GMT+05:30) > > > >Title : RE: RE: [E-devel] [Patch] elm_progressbar patch > > > > > >Hi, prince. > > > > > >I checked your patch right before. > > > > > >But I dont like these new APIs. > > > > > >Currently progressbar provides label_set API for normal case. > > > > > >and people could add extra text with text-part in separted layout or > >elm_label or something else if they need. > > > > > >This kind of API would make the widget break theme changing. > > > > > >Doesn't it? > > > > > >Thanks. > > > > > > > >Let's run together for the best moment! > >-Regards, Hermet- > > > > > >-Original Message- > >From: "ChunEon Park" > >To: "'ChunEon Park'"; > >"'EFL'" > >Cc: prince.du...@samsung.com > >Sent: 11-05-26(목) 11:17:05 > >Subject: RE: [E-devel] [Patch] elm_progressbar patchOops, I missed patch > >file. > >Try it again! :p > >-Original Message- > >From: ChunEon Park [mailto:her...@naver.com] > >Sent: Thursday, May 26, 2011 11:14 AM > >To: EFL > >Cc: prince.du...@samsung.com > >Subject: [E-devel] [Patch] elm_progressbar patch > >Hi, > >This patch is sent by Prince Kumar Dubey (prince.du...@samsung.com) > >Could somebody review this patch? > >If nobody review this, then I will later. > >Thanks. > > > >Let's run together for the best moment! > >-Regards, Hermet- > > > > > >-- > >vRanger cuts backup time in half-while increasing security. > >With the market-leading solution for virtual backup and recovery, > >you get blazing-fast, flexible, and affordable data protection. > >Download your free trial now. > >[2]http://p.sf.net/sfu/quest-d2dcopy1 > >___ > >enlightenment-devel mailing list > >enlightenment-devel@lists.sourceforge.net > >https://lists.sourceforge.net/lists/listinfo/enlightenment-devel > > > > > > [azUrp6pvpx30FrF0FAUdtH3Xp6UqFSdZ74lRhoRmb405MrE5pHkSpB0gbrJRbXk5pZloWrd=.gi > >f] > > > > > > > > > >[cid:Z5JE7EUABGFC@namo.co.kr] > > > >[SeenTimeChecker? > > do=705252d8974533167eb145fa4cffa29fcda6ab03264f2db82f03c9b1 > > 9f0b6f57ba777c355c197185c465c2cf80a2b7ef9aba4bb3b2b5ca43ddd7e184e0604d958075 > > b6b33f32d245b7f8aafe245478a5f1d21d5ebee74427cf878f9a26ce15a0] > > > > References > > > >1. http://sourceforge.net/mailarchive/message.php?msg_id=27444784 > >2. http://p.sf.net/sfu/quest-d2dcopy1 > > > -- > - Codito, ergo sum - "I code, therefore I am" -- > The Rasterman (Carsten Haitzler)ras...@rasterman.com > > > > > > > > > -- - Codito, ergo sum - "I code, therefore I am" -- The Rasterman (Carsten Haitzler)ras...@rasterman.com -- Simplify data backup and recovery for your virtual environment with vRanger. Installation's a snap, and flexible recovery options mean your data is safe, secure and there when you need it. Data protection magic? Nope - It's vRanger. Get your free trial download today. http://p.sf.net/sfu/quest-sfdev2dev ___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/li
Re: [E-devel] [Patch][CtxPopup] Add Arrow on the drag-able area of Base
Dear Mr. Hermet, Hello, I have attached more accurate patch based on latest elementary. The last patch already there. All your review items are reflected on the attached patch. Please review this again, sorry for any inconvenient. Dear Mr. Raster, Hi, Would you please let me know why you don't like the idea of swallowing? Someday the arrow group could be a part of base group.. Previously, I had already made as you mentioned, but Hermet did not prefer that the arrow is part of the ctxpopup(base). Thanks! Sincerely, Shinwoo Kim. 2011/6/2 Carsten Haitzler > On Fri, 27 May 2011 19:08:44 +0900 cnook said: > > this one is fine. though in general i dont like the idea of swallowing the > arrows at all - should just be part of the ctxpopup design/obj directly, > but > thats already there and not specific to this patch :) > > thanks! > > > Dear All, > > > > Hi~ > > > > If CtxPopup has elm_slider for content, the elm_slider indicator is > covered > > by CtxPopup arrow. > > So, I have made a patch for resolving this issue. As the title the arrow > > will be swallowed in the base. > > Even though user uses his/her own style, the attached patch will work > > properly. > > > > Sincerely, > > Shinwoo Kim. > > > -- > - Codito, ergo sum - "I code, therefore I am" -- > The Rasterman (Carsten Haitzler)ras...@rasterman.com > > Index: src/lib/elc_ctxpopup.c === --- src/lib/elc_ctxpopup.c (revision 59898) +++ src/lib/elc_ctxpopup.c (working copy) @@ -484,79 +484,55 @@ { case ELM_CTXPOPUP_DIRECTION_LEFT: edje_object_signal_emit(wd->arrow, "elm,state,left", "elm"); - - // if user does not use dragable part - arrow_size.y = (y - (arrow_size.h * 0.5)); - arrow_size.x = x; - - edje_object_part_swallow(wd->base, "elm.swallow.arrowLeft", wd->arrow); + edje_object_part_swallow(wd->base, "elm.swallow.arrow_left", wd->arrow); if (base_size.h > 0) { - if (y - base_size.y < 0) y = 0; - else if (y > base_size.y + base_size.h) y = base_size.h; - else y = y - base_size.y; - edje_object_part_drag_value_set(wd->base, "elm.swallow.arrowLeft", 0.5, - (double) (y) / (double) (base_size.h)); + if (y < (arrow_size.h * 0.5) + base_size.y) y = 0; + else if (y > base_size.y + base_size.h - (arrow_size.h * 0.5)) y = base_size.h - arrow_size.h; + else y = y - base_size.y - (arrow_size.h * 0.5); + edje_object_part_drag_value_set(wd->base, "elm.swallow.arrow_left", 0.5, + (double) (y) / (double) (base_size.h - arrow_size.h)); } break; case ELM_CTXPOPUP_DIRECTION_RIGHT: edje_object_signal_emit(wd->arrow, "elm,state,right", "elm"); - - // if user does not use dragable part - arrow_size.y = (y - (arrow_size.h * 0.5)); - arrow_size.x = (x - arrow_size.w); - - edje_object_part_swallow(wd->base, "elm.swallow.arrowRight", wd->arrow); + edje_object_part_swallow(wd->base, "elm.swallow.arrow_right", wd->arrow); if (base_size.h > 0) { - if (y -base_size.y < 0) y = 0; - else if (y > base_size.y + base_size.h) y = base_size.h; - else y = y - base_size.y; - edje_object_part_drag_value_set(wd->base, "elm.swallow.arrowRight", 0.5, - (double) (y) / (double) (base_size.h)); + if (y < (arrow_size.h * 0.5) + base_size.y) y = 0; + else if (y > base_size.y + base_size.h - (arrow_size.h * 0.5)) y = base_size.h - arrow_size.h; + else y = y - base_size.y - (arrow_size.h * 0.5); + edje_object_part_drag_value_set(wd->base, "elm.swallow.arrow_right", 0.5, + (double) (y) / (double) (base_size.h - arrow_size.h)); } break; case ELM_CTXPOPUP_DIRECTION_UP: edje_object_signal_emit(wd->arrow, "elm,state,top", "elm"); - - // if user does not use dragable part - arrow_size.x = (x - (arrow_size.w * 0.5)); - arrow_size.y = y; - - edje_object_part_swallow(wd->base, "elm.swallow.arrowUp", wd->arrow); + edje_object_part_swallow(wd->base, "elm.swallow.arrow_up", wd->arrow); if (base_size.w > 0) { - if (x - base_size.x < 0) x = 0; - else if (x > base_size.x + base_size.w) x = base_size.w; - else x = x - base_size.x; - edje_object_part_drag_value_set(wd->base, "elm.swallow.arrowUp", - (double) (x) / (double) (base_size.w), 0.5); + if (x < (arrow_size.w * 0.5) + base_size.x) x = 0; + else if (x > base_size.x + base_s
Re: [E-devel] [PATCH] Add some APIs for supporting virtual keyboard in Ecore_IMF
On Thu, 02 Jun 2011 18:01:37 +0900 Jihoon Kim said: yes please! :) > Thanks, raster. > I missed Changelog patch in ecore. > I attached the ecore changelog patch. > Would you please apply in svn? > > -Original Message- > From: Carsten Haitzler (The Rasterman) [mailto:ras...@rasterman.com] > Sent: Thursday, June 02, 2011 4:58 PM > To: Jihoon Kim > Cc: enlightenment-devel@lists.sourceforge.net > Subject: Re: [E-devel] [PATCH] Add some APIs for supporting virtual keyboard > in Ecore_IMF > > On Mon, 30 May 2011 22:14:27 +0900 Jihoon Kim > said: > > approved! looks good! well done! thanks!. in svn it is. > > > Hi, EFL developers. > > > > For supporting virtual keyboard, I'd like to add some APIs. > > The detail description of each API is included in the patch file as > doxygen > > format. > > > > In the attached patch, the reason why we add the subprefix 'input_panel_' > > related to virtual keyboard is > > that input method can be soft keyboard or voice input or image captured by > > camera. > > > > Would you please review this patch? > > > > > -- > - Codito, ergo sum - "I code, therefore I am" -- > The Rasterman (Carsten Haitzler)ras...@rasterman.com -- - Codito, ergo sum - "I code, therefore I am" -- The Rasterman (Carsten Haitzler)ras...@rasterman.com -- Simplify data backup and recovery for your virtual environment with vRanger. Installation's a snap, and flexible recovery options mean your data is safe, secure and there when you need it. Data protection magic? Nope - It's vRanger. Get your free trial download today. http://p.sf.net/sfu/quest-sfdev2dev ___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
Re: [E-devel] Question about eina_inline_lock_posix.x error.
On Thu, 2 Jun 2011 09:45:32 +0200 (CEST) Vincent Torri said: > On Thu, 2 Jun 2011, Carsten Haitzler (The Rasterman) wrote: > > > On Mon, 30 May 2011 14:57:16 +0200 Cedric BAIL said: > > > > cedirc is right. the real problem is someone including pthread.h before > > Eina.h and this basically clashing with somethign ending up not being > > defined. the _np version also isnt universal so your fix isnt going to > > always work either. > > > > the REAL big problem is actually eina putting CODE inside headers. its > > actually imho a nasty habit of eina's with all its inlined code. it leads > > to these kind of problems . there is no real fix really. 2 pthread.h > > includes clash as they enable different enums depending on the unix98 > > define. > > it's nothing compared to the mess of libjpeg. you include jpeglib.h before > windows.h, error, you do the contrary, still an error... being better than a smelly turd isn't exactly being a creme brulee :) -- - Codito, ergo sum - "I code, therefore I am" -- The Rasterman (Carsten Haitzler)ras...@rasterman.com -- Simplify data backup and recovery for your virtual environment with vRanger. Installation's a snap, and flexible recovery options mean your data is safe, secure and there when you need it. Data protection magic? Nope - It's vRanger. Get your free trial download today. http://p.sf.net/sfu/quest-sfdev2dev ___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
Re: [E-devel] [Patch] make elm_genlist contatin multiline entries
On Wed, 01 Jun 2011 18:07:05 +0900 Sohyun Kim said: there is a downside of this patch resizes of the genlist horizontally lead to horrible flickering as the items are deleted and re-realized AFTER a redraw happened... this isn't too good :( btw - attached updated patch as i had to merge changes after i put in other patches in upstream efl. :( > Hello, this is Sohyun. > > I made a patch for elm_genlist. > Elm_genlist cannot contain multiline entries because the height of multiline > entry is decided after calculating elm_genlist_item height. > > I added changed_size_hint event callback for elm_genlist icons. > When an icon size is changed, elm_genlist recalculates the item height. > In other words, if multiline entry height is decided, elm_genlist_item > height is recalculated. > > Elm_genlist may have many multiline entries, so I added recalculating > routine as a job. > If many multiline entries are resized, the elm_genlist recalculating routine > will be run once. > > You can check elementary_test > genlist 6 > > Thanks > -- - Codito, ergo sum - "I code, therefore I am" -- The Rasterman (Carsten Haitzler)ras...@rasterman.com Index: src/lib/elm_genlist.c === --- src/lib/elm_genlist.c (revision 59889) +++ src/lib/elm_genlist.c (working copy) @@ -205,7 +205,7 @@ Eina_List*group_items; Pan *pan; Evas_Coordpan_x, pan_y, old_pan_y, w, h, minw, minh, realminw, prev_viewport_w; - Ecore_Job*calc_job, *update_job; + Ecore_Job*calc_job, *update_job, *changed_job; Ecore_Idle_Enterer *queue_idle_enterer; Ecore_Idler*must_recalc_idler; Eina_List*queue, *selected; @@ -377,6 +377,8 @@ static void _calc_job(void *data); static void _on_focus_hook(void*data, Evas_Object *obj); +static void _changed_size_hints(void *data, Evas *e, Evas_Object *obj, void *event_info); +static void _changed_job(void *data); static Eina_Bool _item_multi_select_up(Widget_Data *wd); static Eina_Bool _item_multi_select_down(Widget_Data *wd); static Eina_Bool _item_single_select_up(Widget_Data *wd); @@ -704,6 +706,7 @@ _item_cache_zero(wd); if (wd->calc_job) ecore_job_del(wd->calc_job); if (wd->update_job) ecore_job_del(wd->update_job); + if (wd->changed_job) ecore_job_del(wd->changed_job); if (wd->queue_idle_enterer) ecore_idle_enterer_del(wd->queue_idle_enterer); if (wd->must_recalc_idler) ecore_idler_del(wd->must_recalc_idler); if (wd->multi_timer) ecore_timer_del(wd->multi_timer); @@ -1875,6 +1878,7 @@ edje_object_part_swallow(target, key, ic); evas_object_show(ic); elm_widget_sub_object_add(it->base.widget, ic); + evas_object_event_callback_add(ic, EVAS_CALLBACK_CHANGED_SIZE_HINTS, _changed_size_hints, it); if (it->disabled) elm_widget_disabled_set(ic, EINA_TRUE); } @@ -2438,6 +2442,17 @@ } static void +_changed_size_hints(void *data, Evas *e __UNUSED__, Evas_Object *obj __UNUSED__, void *event_info __UNUSED__) +{ + Elm_Genlist_Item *it = data; + if (!it) return; + it->mincalcd = EINA_FALSE; + it->block->updateme = EINA_TRUE; + if (it->wd->changed_job) ecore_job_del(it->wd->changed_job); + it->wd->changed_job = ecore_job_add(_changed_job, it->wd); +} + +static void _group_items_recalc(void *data) { Widget_Data *wd = data; @@ -2675,6 +2690,72 @@ } static void +_changed_job(void *data) +{ + Widget_Data *wd = data; Eina_List *l2; + Item_Block *itb; + int num, num0, position = 0, recalc = 0; + if (!wd) return; + wd->changed_job = NULL; + num = 0; + EINA_INLIST_FOREACH(wd->blocks, itb) + { + Evas_Coord itminw, itminh; + Elm_Genlist_Item *it; + + if (!itb->updateme) +{ + num += itb->count; + if (position) + _item_block_position(itb, num); + continue; +} + num0 = num; + recalc = 0; + EINA_LIST_FOREACH(itb->items, l2, it) + { + if (!it->mincalcd) + { + if (it->realized) +{ + Evas_Coord mw = -1, mh = -1; + itminw = it->minw; + itminh = it->minh; + + if (it->wd->height_for_width) mw = it->wd->w; + if (!it->display_only) + elm_coords_finger_size_adjust(1, &mw, 1, &mh); + if (it->wd->height_for_width) mw = it->wd->prev_viewport_w; + edje_object_size_min_restricted_calc(it->base.view, &mw, &mh, mw, mh); + if (!it->display_only) + elm_coords_finger_size_adjust(1, &mw, 1, &mh); + it->w = it->minw = mw; + it->h = it->minh = mh; + i
Re: [E-devel] [Patch] elm_progressbar patch
On Wed, 01 Jun 2011 09:47:01 + (GMT) PRINCE KUMAR DUBEY said: ok some review of patch itself: + EAPI void*elm_progressbar_part_label_set(Evas_Object *obj, const char *part, const char *label) EINA_ARG_NONNULL(1); why does it return void *? actualyl the function itself RETURNS nothing. so this is wrong on 2 counts. i think you meant it to just be void not void *. + EINA_LIST_FOREACH(wd->parts, l, info) + { +if (info->label) eina_stringshare_del(info->label); +if (info->name) eina_stringshare_del(info->name); +wd->parts = eina_list_remove(wd->parts, info); + +free(info); + } use EINA_LIST_FREE here. its less code (no need to remove, and no "l" needed). +EAPI const char * +elm_progressbar_part_label_get(const Evas_Object *obj, const char *part) ... + return edje_object_part_text_get(wd->progressbar, part); umm you want to return the text from the wd->parts list, not from edje. why? what if u set text for a part that doesnt exist? now you cant get what you set... :) can you send an updated patch? > >Hi Mr Park, > > >Thanks for your comment. > > >I already had discussion with Rasterman about this new API. > >Please visit this link for reference >[1]http://sourceforge.net/mailarchive/message.php?msg_id=27444784. > > >and this API would not make widget to break theme changing feature as we > are handling this in _theme_hook(). > > >Regards, > >Prince > > > >--- Original Message --- > >Sender : ChunEon Park > >Date : Jun 01, 2011 14:19 (GMT+05:30) > >Title : RE: RE: [E-devel] [Patch] elm_progressbar patch > > >Hi, prince. > > >I checked your patch right before. > > >But I dont like these new APIs. > > >Currently progressbar provides label_set API for normal case. > > >and people could add extra text with text-part in separted layout or >elm_label or something else if they need. > > >This kind of API would make the widget break theme changing. > > >Doesn't it? > > >Thanks. > > > >Let's run together for the best moment! >-Regards, Hermet- > > >-Original Message- >From: "ChunEon Park" >To: "'ChunEon Park'"; >"'EFL'" >Cc: prince.du...@samsung.com >Sent: 11-05-26(목) 11:17:05 >Subject: RE: [E-devel] [Patch] elm_progressbar patchOops, I missed patch >file. >Try it again! :p >-Original Message- >From: ChunEon Park [mailto:her...@naver.com] >Sent: Thursday, May 26, 2011 11:14 AM >To: EFL >Cc: prince.du...@samsung.com >Subject: [E-devel] [Patch] elm_progressbar patch >Hi, >This patch is sent by Prince Kumar Dubey (prince.du...@samsung.com) >Could somebody review this patch? >If nobody review this, then I will later. >Thanks. > >Let's run together for the best moment! >-Regards, Hermet- > > >-- >vRanger cuts backup time in half-while increasing security. >With the market-leading solution for virtual backup and recovery, >you get blazing-fast, flexible, and affordable data protection. >Download your free trial now. >[2]http://p.sf.net/sfu/quest-d2dcopy1 >___ >enlightenment-devel mailing list >enlightenment-devel@lists.sourceforge.net >https://lists.sourceforge.net/lists/listinfo/enlightenment-devel > > > [azUrp6pvpx30FrF0FAUdtH3Xp6UqFSdZ74lRhoRmb405MrE5pHkSpB0gbrJRbXk5pZloWrd=.gi >f] > > > > >[cid:Z5JE7EUABGFC@namo.co.kr] > >[SeenTimeChecker? > do=705252d8974533167eb145fa4cffa29fcda6ab03264f2db82f03c9b1 > 9f0b6f57ba777c355c197185c465c2cf80a2b7ef9aba4bb3b2b5ca43ddd7e184e0604d958075 > b6b33f32d245b7f8aafe245478a5f1d21d5ebee74427cf878f9a26ce15a0] > > References > >1. http://sourceforge.net/mailarchive/message.php?msg_id=27444784 >2. http://p.sf.net/sfu/quest-d2dcopy1 -- - Codito, ergo sum - "I code, therefore I am" -- The Rasterman (Carsten Haitzler)ras...@rasterman.com -- Simplify data backup and recovery for your virtual environment with vRanger. Installation's a snap, and flexible recovery options mean your data is safe, secure and there when you need it. Data protection magic? Nope - It's vRanger. Get your free trial download today. http://p.sf.net/sfu/quest-sfdev2dev ___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
Re: [E-devel] [PATCH] Add some APIs for supporting virtual keyboard in Ecore_IMF
Thanks, raster. I missed Changelog patch in ecore. I attached the ecore changelog patch. Would you please apply in svn? -Original Message- From: Carsten Haitzler (The Rasterman) [mailto:ras...@rasterman.com] Sent: Thursday, June 02, 2011 4:58 PM To: Jihoon Kim Cc: enlightenment-devel@lists.sourceforge.net Subject: Re: [E-devel] [PATCH] Add some APIs for supporting virtual keyboard in Ecore_IMF On Mon, 30 May 2011 22:14:27 +0900 Jihoon Kim said: approved! looks good! well done! thanks!. in svn it is. > Hi, EFL developers. > > For supporting virtual keyboard, I'd like to add some APIs. > The detail description of each API is included in the patch file as doxygen > format. > > In the attached patch, the reason why we add the subprefix 'input_panel_' > related to virtual keyboard is > that input method can be soft keyboard or voice input or image captured by > camera. > > Would you please review this patch? > -- - Codito, ergo sum - "I code, therefore I am" -- The Rasterman (Carsten Haitzler)ras...@rasterman.com Index: ChangeLog === --- ChangeLog (revision 59894) +++ ChangeLog (working copy) @@ -210,3 +210,7 @@ * Add ecore_pipe_freeze/thaw to suspend and restart watching the pipe inside the main loop. + +2011-06-02 Jihoon Kim + +* Ecore_IMF: Added APIs for setting layout, language and controlling control panel -- Simplify data backup and recovery for your virtual environment with vRanger. Installation's a snap, and flexible recovery options mean your data is safe, secure and there when you need it. Data protection magic? Nope - It's vRanger. Get your free trial download today. http://p.sf.net/sfu/quest-sfdev2dev___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
Re: [E-devel] [Patch] elm_progressbar patch
On Wed, 01 Jun 2011 09:47:01 + (GMT) PRINCE KUMAR DUBEY said: yeah... all i can imagine being better is this being generic like signal emit hooks > >Hi Mr Park, > > >Thanks for your comment. > > >I already had discussion with Rasterman about this new API. > >Please visit this link for reference >[1]http://sourceforge.net/mailarchive/message.php?msg_id=27444784. > > >and this API would not make widget to break theme changing feature as we > are handling this in _theme_hook(). > > >Regards, > >Prince > > > >--- Original Message --- > >Sender : ChunEon Park > >Date : Jun 01, 2011 14:19 (GMT+05:30) > >Title : RE: RE: [E-devel] [Patch] elm_progressbar patch > > >Hi, prince. > > >I checked your patch right before. > > >But I dont like these new APIs. > > >Currently progressbar provides label_set API for normal case. > > >and people could add extra text with text-part in separted layout or >elm_label or something else if they need. > > >This kind of API would make the widget break theme changing. > > >Doesn't it? > > >Thanks. > > > >Let's run together for the best moment! >-Regards, Hermet- > > >-Original Message- >From: "ChunEon Park" >To: "'ChunEon Park'"; >"'EFL'" >Cc: prince.du...@samsung.com >Sent: 11-05-26(목) 11:17:05 >Subject: RE: [E-devel] [Patch] elm_progressbar patchOops, I missed patch >file. >Try it again! :p >-Original Message- >From: ChunEon Park [mailto:her...@naver.com] >Sent: Thursday, May 26, 2011 11:14 AM >To: EFL >Cc: prince.du...@samsung.com >Subject: [E-devel] [Patch] elm_progressbar patch >Hi, >This patch is sent by Prince Kumar Dubey (prince.du...@samsung.com) >Could somebody review this patch? >If nobody review this, then I will later. >Thanks. > >Let's run together for the best moment! >-Regards, Hermet- > > >-- >vRanger cuts backup time in half-while increasing security. >With the market-leading solution for virtual backup and recovery, >you get blazing-fast, flexible, and affordable data protection. >Download your free trial now. >[2]http://p.sf.net/sfu/quest-d2dcopy1 >___ >enlightenment-devel mailing list >enlightenment-devel@lists.sourceforge.net >https://lists.sourceforge.net/lists/listinfo/enlightenment-devel > > > [azUrp6pvpx30FrF0FAUdtH3Xp6UqFSdZ74lRhoRmb405MrE5pHkSpB0gbrJRbXk5pZloWrd=.gi >f] > > > > >[cid:Z5JE7EUABGFC@namo.co.kr] > >[SeenTimeChecker? > do=705252d8974533167eb145fa4cffa29fcda6ab03264f2db82f03c9b1 > 9f0b6f57ba777c355c197185c465c2cf80a2b7ef9aba4bb3b2b5ca43ddd7e184e0604d958075 > b6b33f32d245b7f8aafe245478a5f1d21d5ebee74427cf878f9a26ce15a0] > > References > >1. http://sourceforge.net/mailarchive/message.php?msg_id=27444784 >2. http://p.sf.net/sfu/quest-d2dcopy1 -- - Codito, ergo sum - "I code, therefore I am" -- The Rasterman (Carsten Haitzler)ras...@rasterman.com -- Simplify data backup and recovery for your virtual environment with vRanger. Installation's a snap, and flexible recovery options mean your data is safe, secure and there when you need it. Data protection magic? Nope - It's vRanger. Get your free trial download today. http://p.sf.net/sfu/quest-sfdev2dev ___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
Re: [E-devel] [Patch] elm_progressbar patch
On Wed, 01 Jun 2011 10:24:31 + (GMT) RAJEEV RANJAN said: correct. as long as theme provides same text parts it will be happy. if it provides new ones you didnt set, they will be empty or have default text, and if it doesnt provide some labels, you lose text. i'd call it a theme bug if its not providing all the text elements its expected to. > >Hi Hermet, > > Adding a text part using edje text part in edc is possible however if we >use this approach then we won't be able to provide i18n support as this >support is not there at edje level as of now. > >As far as breaking of themability is concerned, as the name of parts along >with label for those in the list are being maintained, so on theme change >the widget can apply these parts label if those parts are present in the > new theme. > >An application using progressbar anyway will be aware of the text parts >which it is interesting in adding, so it should provide those parts in all >themes or extensions/overlays. > >Thanks. > >Regards, > >Rajeev > > >--- Original Message --- > >Message: 1 >Date: Wed, 01 Jun 2011 17:49:35 +0900 >From: ChunEon Park >Subject: Re: [E-devel] [Patch] elm_progressbar patch >To: ChunEon Park >Cc: prince.du...@samsung.com, 'EFL' >Message-ID: <7677c7613f53ca7088b6efde4e7ce...@tweb36-2.nm> >Content-Type: text/plain; charset="EUC-KR" >Hi, prince. >I checked your patch right before. >But I dont like these new APIs. >Currently progressbar provides label_set API for normal case. >and people could add extra text with text-part in separted layout or >elm_label or something else if they need. >This kind of API would make the widget break theme changing. >Doesn't it? >Thanks. > >Let's run together for the best moment! >-Regards, Hermet- >-Original Message- >From: "ChunEon Park" >To: "'ChunEon Park'"; >"'EFL'" >Cc: prince.du...@samsung.com >Sent: 11-05-26(?) 11:17:05 >Subject: RE: [E-devel] [Patch] elm_progressbar patchOops, I missed patch >file. >Try it again! :p >-Original Message- >From: ChunEon Park [mailto:her...@naver.com] >Sent: Thursday, May 26, 2011 11:14 AM >To: EFL >Cc: prince.du...@samsung.com >Subject: [E-devel] [Patch] elm_progressbar patch >Hi, >This patch is sent by Prince Kumar Dubey (prince.du...@samsung.com) >Could somebody review this patch? >If nobody review this, then I will later. >Thanks. > >Let's run together for the best moment! >-Regards, Hermet- > > >-- >vRanger cuts backup time in half-while increasing security. >With the market-leading solution for virtual backup and recovery, >you get blazing-fast, flexible, and affordable data protection. >Download your free trial now. >http://p.sf.net/sfu/quest-d2dcopy1 >___ >enlightenment-devel mailing list >enlightenment-devel@lists.sourceforge.net >https://lists.sourceforge.net/lists/listinfo/enlightenment-devel >[SeenTimeChecker? > do=705252d897453316b22ebdd87d25aaf3c8f3520e2c71f17ff1d2345d > 123b7a3c770255183a8f81f75f2c6c48df01bb6e980b3871b38ac2d74ed859d628c32758 > cd62c5cf0684a728c55b39cc59eacf878f9a26ce15a0] > -- > Simplify data backup and recovery for your virtual environment with vRanger. > Installation's a snap, and flexible recovery options mean your data is safe, > secure and there when you need it. Data protection magic? > Nope - It's vRanger. Get your free trial download today. > http://p.sf.net/sfu/quest-sfdev2dev > ___ > enlightenment-devel mailing list > enlightenment-devel@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/enlightenment-devel > -- - Codito, ergo sum - "I code, therefore I am" -- The Rasterman (Carsten Haitzler)ras...@rasterman.com -- Simplify data backup and recovery for your virtual environment with vRanger. Installation's a snap, and flexible recovery options mean your data is safe, secure and there when you need it. Data protection magic? Nope - It's vRanger. Get your free trial download today. http://p.sf.net/sfu/quest-sfdev2dev ___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
Re: [E-devel] e17 todo
On Mon, 30 May 2011 11:40:20 -0300 Gustavo Sverzut Barbieri said: > >> no, it can't. He says if you're doing development or other corner-case > >> task (I do have to agree here, it's not a regular use case) you can > >> start "connmand -I eth1" to ignore eth1. > >> > >> He said if this is about host x gadget mode on USB, then he said it > >> already detects USB gadget side, this is used to enable USB tethering. > > > > well i swear theres some connman-gnome util GUI that lets me select specific > > devices to ignore. so it's somehow possible... > > it was never per-device, but per network type. You have switches to > turn off bluetooth, 3g, ethernet and wifi. But nothing to change just > the first ethernet. but its still something that e's connman doesnt let us turn on/off/auto/whatever - cant remember the options. > > but my point is that when you have a network that doesnt even tell you what > > proxy to use and you need to look up manual ip addresses for the PAC url > > location .. .for your browser, and then for simpler things like wget, apt > > have to set http and https proxy env vars etc... it's be nice just to have > > 1 stop to manually set it up.. much like manual ip address setup. :) > > I know, that was my idea as well, but it was rejected. Thus either we > do this hack in e17 and talk to iptables, or we rely on tools being > patched to talk to this proxy configuration daemon using dbus > (wget/curl comes to mind). i'd add the hack in e as its simple to store the proxy info together per "connection" and when connman says "i connected to net X" we look up http proxy config for net X - if any, set http/https, or unset it. or hell.. yeah as u say.. talk to the proxy daemon ourselves and based on that result set env vars. > OTOH we can have ecore-con to talk to proxy daemon itself, so at least > our code is safe. Problem here is the reorganization, we need e_dbus > to talk to dbus daemon, but we're inside ecore-con. Likely we can > solve this by delegating using callbacks in ecore-con, create a e_dbus > module that talks to the daemon and hooking them in some higher level > library or the application itself. > (ecore_con_proxy_for_url_resolver_set(ctxt, cb, del_cb)) yeah - glue in a callback thing, tho ecore_con_url will respect http proxy env vars (as curl does) > > hmm all fine and great to do http req... if your proxy setup is right :) > > does it manage to just detect lan connectivity? eg if it can ping the > > router etc.? i mulled this too but it all assumes that the user just wants > > connectivity to the "big bad internet" and not "just the local network". > > does it have "levels of success" eg. succeeded in getting address, > > succeeded in getting address AND can ping/contact LAN hosts (gateway, dns > > server), got address, can access LAN hosts ANd can get to the big bad > > interweebs? thats what the ready and online states are? ready being step #2 > > above, online is #3? > > Yes, it does have levels of success. I don't know exactly which one is > the "final", but it will do a sequence of steps to determine you're > online, including emitting a signal that login is required so you can > present a browser with the login URL. ok. i guess we need to display this then when it changes :) > > cool. i guess we msotly need to have handling for all the statuses (ready > > and online, loginrequired etc.) and display appropriate icons/status. for > > loginrequired state in theory we should launch a browser when u click on it > > or offer a menu with "login now?" or something. it'd be nice to use eve at > > some point... maybe e18. worry about that then. > > Yes, for this one you want a very minimal chrome for the browser. > iPhone shows just a webview with back/forward buttons, that we can do > using webkit-efl directly and ship as e18 helper binary. yup. agreed there. no need for a full browser, just chrome... THOUGH... a good browser can do passowrd/login remembering, and a simple webkit chrome thing wont without help. if we implement enough to DO form content remembering between sessions... then we'd be good to go. > >> that is done already, data is persistent across sessions. > > > > cool. then all we need to do is.. be able to display and otherwise "control > > it" (turn it on and off, clear/delete sessions etc.) > > Anyway, problem is who is working on all of that? :-) lucas volunteered! :) -- - Codito, ergo sum - "I code, therefore I am" -- The Rasterman (Carsten Haitzler)ras...@rasterman.com -- Simplify data backup and recovery for your virtual environment with vRanger. Installation's a snap, and flexible recovery options mean your data is safe, secure and there when you need it. Data protection magic? Nope - It's vRanger. Get your free trial download today. http://p.sf.net/sfu/quest-sfdev2dev ___ enlightenment-devel m
Re: [E-devel] about packaging/debian directories
On Wed, 1 Jun 2011 12:03:14 +0200 michael bouchaud said: > The probleme of debianization of this packages aren't the stability of > libraries or softwares but migration between each update. And this can made > a lot of work. For exemple edje have break the file format before the > release. So Lutin had to done a patch to verify all theme loaded and make > sure it was compatible. If the theme aren't compatible convert it on the > fly. This is not trivial. > So I agree to upload all stable and released packages to debian but not > snapshoted packages for moment. For information Lutin has made all this work > except eeze. > I prefer support our own 'ppa' as raster said, and provide this packages > here. agreed. for those who are happy to live on bleeding edge - use our ppa/repo/whatever. those who want stable and really well, will have to wait longer and go for debian etc. upstream well tested stuff. -- - Codito, ergo sum - "I code, therefore I am" -- The Rasterman (Carsten Haitzler)ras...@rasterman.com -- Simplify data backup and recovery for your virtual environment with vRanger. Installation's a snap, and flexible recovery options mean your data is safe, secure and there when you need it. Data protection magic? Nope - It's vRanger. Get your free trial download today. http://p.sf.net/sfu/quest-sfdev2dev ___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
Re: [E-devel] [PATCH] Add some APIs for supporting virtual keyboard in Ecore_IMF
On Mon, 30 May 2011 22:14:27 +0900 Jihoon Kim said: approved! looks good! well done! thanks!. in svn it is. > Hi, EFL developers. > > For supporting virtual keyboard, I'd like to add some APIs. > The detail description of each API is included in the patch file as doxygen > format. > > In the attached patch, the reason why we add the subprefix 'input_panel_' > related to virtual keyboard is > that input method can be soft keyboard or voice input or image captured by > camera. > > Would you please review this patch? > -- - Codito, ergo sum - "I code, therefore I am" -- The Rasterman (Carsten Haitzler)ras...@rasterman.com -- Simplify data backup and recovery for your virtual environment with vRanger. Installation's a snap, and flexible recovery options mean your data is safe, secure and there when you need it. Data protection magic? Nope - It's vRanger. Get your free trial download today. http://p.sf.net/sfu/quest-sfdev2dev ___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
Re: [E-devel] evas and locks
On Thu, 2 Jun 2011, Vincent Torri wrote: > > hey, > > currently, on Windows, i build evas without async preload (expedite > segfault in evas_object_text, i don't know why at all). expedite segfault if i build evas *with* async preload. it works without but there are the messages... Vincent -- Simplify data backup and recovery for your virtual environment with vRanger. Installation's a snap, and flexible recovery options mean your data is safe, secure and there when you need it. Data protection magic? Nope - It's vRanger. Get your free trial download today. http://p.sf.net/sfu/quest-sfdev2dev ___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
Re: [E-devel] Question about eina_inline_lock_posix.x error.
On Thu, 2 Jun 2011, Carsten Haitzler (The Rasterman) wrote: On Mon, 30 May 2011 14:57:16 +0200 Cedric BAIL said: cedirc is right. the real problem is someone including pthread.h before Eina.h and this basically clashing with somethign ending up not being defined. the _np version also isnt universal so your fix isnt going to always work either. the REAL big problem is actually eina putting CODE inside headers. its actually imho a nasty habit of eina's with all its inlined code. it leads to these kind of problems . there is no real fix really. 2 pthread.h includes clash as they enable different enums depending on the unix98 define. one can do a big fat hack in eina_inline_lock_posix.x : #ifndef PTHREAD_COND_INITIALIZER # include #endif or something like that If we use 'classic' functions (no inline) will the code be really slower ? Vincent hello, On Mon, May 30, 2011 at 2:31 PM, WooHyun Jung wrote: If I include "pthread.h" first and then "Elementary.h", following error message is shown. (__USE_UNIX98 = 0) /usr/include/eina-1/eina/eina_inline_lock_posix.x:101:41: error: 'PTHREAD_MUTEX_ERRORCHECK' undeclared (first use in this function) Should eina_inline_lock_posix.x file be modified as below ? #ifdef __USE_UNIX98 if (pthread_mutexattr_settype(&attr, PTHREAD_MUTEX_ERRORCHECK) != 0) return EINA_FALSE; #else if (pthread_mutexattr_settype(&attr, PTHREAD_MUTEX_ERRORCHECK_NP) != 0) return EINA_FALSE; #endif The issue is in fact more related with double inclusion of pthread.h. And we don't have a real proper work around. I move that piece of code to be used only in debug mode. We could as suggested on irc add -D__USE_UNIX98 in the CFLAGS of eina pkgconfig file. In general, eina should provide all the portable layer for mutex and ecore the layer to sanely manipulate threads. So at the ends, we should not use pthread.h directly in an EFL apps. Right now, I know that it's not completly the case as some feature are still needed. But if you can identify what are the feature that would be usefull first, I am interested to know. -- Cedric BAIL -- vRanger cuts backup time in half-while increasing security. With the market-leading solution for virtual backup and recovery, you get blazing-fast, flexible, and affordable data protection. Download your free trial now. http://p.sf.net/sfu/quest-d2dcopy1 ___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel -- - Codito, ergo sum - "I code, therefore I am" -- The Rasterman (Carsten Haitzler)ras...@rasterman.com -- Simplify data backup and recovery for your virtual environment with vRanger. Installation's a snap, and flexible recovery options mean your data is safe, secure and there when you need it. Data protection magic? Nope - It's vRanger. Get your free trial download today. http://p.sf.net/sfu/quest-sfdev2dev ___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel -- Simplify data backup and recovery for your virtual environment with vRanger. Installation's a snap, and flexible recovery options mean your data is safe, secure and there when you need it. Data protection magic? Nope - It's vRanger. Get your free trial download today. http://p.sf.net/sfu/quest-sfdev2dev___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
[E-devel] evas and locks
hey, currently, on Windows, i build evas without async preload (expedite segfault in evas_object_text, i don't know why at all). The problem of building evas without async preload is that there is a plethora of messages that are sidplayed in the terminal. The reason is: 1) in evas_common.h, LKT is defined as 0 2) in evas_image_scalecache.c, the code #if 1 ret = LKT(im->cache.lock); if (ret == EINA_FALSE) /* can't get image lock */ is the problem. ret is always 0, hence all those messages in the terminal (see the body of the check) What can be done to avoid that problem ? Vincent -- Simplify data backup and recovery for your virtual environment with vRanger. Installation's a snap, and flexible recovery options mean your data is safe, secure and there when you need it. Data protection magic? Nope - It's vRanger. Get your free trial download today. http://p.sf.net/sfu/quest-sfdev2dev ___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
Re: [E-devel] Question about eina_inline_lock_posix.x error.
On Thu, 2 Jun 2011, Carsten Haitzler (The Rasterman) wrote: On Mon, 30 May 2011 14:57:16 +0200 Cedric BAIL said: cedirc is right. the real problem is someone including pthread.h before Eina.h and this basically clashing with somethign ending up not being defined. the _np version also isnt universal so your fix isnt going to always work either. the REAL big problem is actually eina putting CODE inside headers. its actually imho a nasty habit of eina's with all its inlined code. it leads to these kind of problems . there is no real fix really. 2 pthread.h includes clash as they enable different enums depending on the unix98 define. it's nothing compared to the mess of libjpeg. you include jpeglib.h before windows.h, error, you do the contrary, still an error... Vincent hello, On Mon, May 30, 2011 at 2:31 PM, WooHyun Jung wrote: If I include "pthread.h" first and then "Elementary.h", following error message is shown. (__USE_UNIX98 = 0) /usr/include/eina-1/eina/eina_inline_lock_posix.x:101:41: error: 'PTHREAD_MUTEX_ERRORCHECK' undeclared (first use in this function) Should eina_inline_lock_posix.x file be modified as below ? #ifdef __USE_UNIX98 if (pthread_mutexattr_settype(&attr, PTHREAD_MUTEX_ERRORCHECK) != 0) return EINA_FALSE; #else if (pthread_mutexattr_settype(&attr, PTHREAD_MUTEX_ERRORCHECK_NP) != 0) return EINA_FALSE; #endif The issue is in fact more related with double inclusion of pthread.h. And we don't have a real proper work around. I move that piece of code to be used only in debug mode. We could as suggested on irc add -D__USE_UNIX98 in the CFLAGS of eina pkgconfig file. In general, eina should provide all the portable layer for mutex and ecore the layer to sanely manipulate threads. So at the ends, we should not use pthread.h directly in an EFL apps. Right now, I know that it's not completly the case as some feature are still needed. But if you can identify what are the feature that would be usefull first, I am interested to know. -- Cedric BAIL -- vRanger cuts backup time in half-while increasing security. With the market-leading solution for virtual backup and recovery, you get blazing-fast, flexible, and affordable data protection. Download your free trial now. http://p.sf.net/sfu/quest-d2dcopy1 ___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel -- - Codito, ergo sum - "I code, therefore I am" -- The Rasterman (Carsten Haitzler)ras...@rasterman.com -- Simplify data backup and recovery for your virtual environment with vRanger. Installation's a snap, and flexible recovery options mean your data is safe, secure and there when you need it. Data protection magic? Nope - It's vRanger. Get your free trial download today. http://p.sf.net/sfu/quest-sfdev2dev ___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel -- Simplify data backup and recovery for your virtual environment with vRanger. Installation's a snap, and flexible recovery options mean your data is safe, secure and there when you need it. Data protection magic? Nope - It's vRanger. Get your free trial download today. http://p.sf.net/sfu/quest-sfdev2dev___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
Re: [E-devel] Question about eina_inline_lock_posix.x error.
On Mon, 30 May 2011 14:57:16 +0200 Cedric BAIL said: cedirc is right. the real problem is someone including pthread.h before Eina.h and this basically clashing with somethign ending up not being defined. the _np version also isnt universal so your fix isnt going to always work either. the REAL big problem is actually eina putting CODE inside headers. its actually imho a nasty habit of eina's with all its inlined code. it leads to these kind of problems . there is no real fix really. 2 pthread.h includes clash as they enable different enums depending on the unix98 define. > hello, > > On Mon, May 30, 2011 at 2:31 PM, WooHyun Jung wrote: > > If I include "pthread.h" first and then "Elementary.h", following error > > message is shown. (__USE_UNIX98 = 0) > > > > /usr/include/eina-1/eina/eina_inline_lock_posix.x:101:41: error: > > 'PTHREAD_MUTEX_ERRORCHECK' undeclared (first use in this function) > > > > Should eina_inline_lock_posix.x file be modified as below ? > > > > #ifdef __USE_UNIX98 > > > > if (pthread_mutexattr_settype(&attr, PTHREAD_MUTEX_ERRORCHECK) != 0) > > > > return EINA_FALSE; > > > > #else > > > > if (pthread_mutexattr_settype(&attr, PTHREAD_MUTEX_ERRORCHECK_NP) != > > 0) > > > > return EINA_FALSE; > > > > #endif > > The issue is in fact more related with double inclusion of pthread.h. > And we don't have a real proper work around. I move that piece of code > to be used only in debug mode. We could as suggested on irc add > -D__USE_UNIX98 in the CFLAGS of eina pkgconfig file. > > In general, eina should provide all the portable layer for mutex and > ecore the layer to sanely manipulate threads. So at the ends, we > should not use pthread.h directly in an EFL apps. Right now, I know > that it's not completly the case as some feature are still needed. But > if you can identify what are the feature that would be usefull first, > I am interested to know. > -- > Cedric BAIL > > -- > vRanger cuts backup time in half-while increasing security. > With the market-leading solution for virtual backup and recovery, > you get blazing-fast, flexible, and affordable data protection. > Download your free trial now. > http://p.sf.net/sfu/quest-d2dcopy1 > ___ > enlightenment-devel mailing list > enlightenment-devel@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/enlightenment-devel -- - Codito, ergo sum - "I code, therefore I am" -- The Rasterman (Carsten Haitzler)ras...@rasterman.com -- Simplify data backup and recovery for your virtual environment with vRanger. Installation's a snap, and flexible recovery options mean your data is safe, secure and there when you need it. Data protection magic? Nope - It's vRanger. Get your free trial download today. http://p.sf.net/sfu/quest-sfdev2dev ___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
Re: [E-devel] EFL Windows Installer
On Thu, 2 Jun 2011, Cedric BAIL wrote: > On Thu, Jun 2, 2011 at 8:05 AM, Vincent Torri wrote: >> On Thu, 2 Jun 2011, Lionel Orry wrote: >>> Vincent Torri univ-evry.fr> writes: 1) on my computer, if you install the EFL in c:\a\path, then the $PATH env var is still appended with c:\Program Files\Efl\bin. I don't understand why. You have to change that manually for now if it happens. >>> >>> Would you agree to share (on svn repo or anywhere else) your installer >>> project >>> files (nsis, innosetup or whatever it is)? >> >> no problem. Here they are. I use MSYS+MinGW and NSIS >> >> in a subdir (for me, ~/svnroot): >> >> 1) checkout the EFL in it, so you must have in ~/svnroot the following >> subdirs: >> >> evil >> eina >> eet >> evas >> expedite >> ecore >> embryo >> edje >> elementary >> >> 2) put in ~/svnroot the files test.sh, Efl.nsi and license.txt >> >> 3) install in /usr/local the dependencies (the wiki is not up to date : i >> have to add the installation of libcurl (with all the dependencies...) for >> ecore_con >> >> I want to provide a "dependency" zip file. But only for the 1.1 release, not >> before. >> >> 4) in ~/svnroot, run: >> >> ./test.sh debug >> ./test.sh package >> >> It will build the EFL in debug mode and create the installer (note that NSIS >> must have been installed in c:\Program Files >> >> comments, remarks, ideas, fixes are welcome of course. > > Maybe it could be included in our svn package directory ? for the 1.1 release, yes, but not now. Vincent -- Simplify data backup and recovery for your virtual environment with vRanger. Installation's a snap, and flexible recovery options mean your data is safe, secure and there when you need it. Data protection magic? Nope - It's vRanger. Get your free trial download today. http://p.sf.net/sfu/quest-sfdev2dev ___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
Re: [E-devel] EFL Windows Installer
On Thu, Jun 2, 2011 at 8:05 AM, Vincent Torri wrote: > On Thu, 2 Jun 2011, Lionel Orry wrote: >> Vincent Torri univ-evry.fr> writes: >>> >>> 1) on my computer, if you install the EFL in c:\a\path, then the $PATH >>> env >>> var is still appended with c:\Program Files\Efl\bin. I don't understand >>> why. You have to change that manually for now if it happens. >> >> Would you agree to share (on svn repo or anywhere else) your installer >> project >> files (nsis, innosetup or whatever it is)? > > no problem. Here they are. I use MSYS+MinGW and NSIS > > in a subdir (for me, ~/svnroot): > > 1) checkout the EFL in it, so you must have in ~/svnroot the following > subdirs: > > evil > eina > eet > evas > expedite > ecore > embryo > edje > elementary > > 2) put in ~/svnroot the files test.sh, Efl.nsi and license.txt > > 3) install in /usr/local the dependencies (the wiki is not up to date : i > have to add the installation of libcurl (with all the dependencies...) for > ecore_con > > I want to provide a "dependency" zip file. But only for the 1.1 release, not > before. > > 4) in ~/svnroot, run: > > ./test.sh debug > ./test.sh package > > It will build the EFL in debug mode and create the installer (note that NSIS > must have been installed in c:\Program Files > > comments, remarks, ideas, fixes are welcome of course. Maybe it could be included in our svn package directory ? -- Cedric BAIL -- Simplify data backup and recovery for your virtual environment with vRanger. Installation's a snap, and flexible recovery options mean your data is safe, secure and there when you need it. Data protection magic? Nope - It's vRanger. Get your free trial download today. http://p.sf.net/sfu/quest-sfdev2dev ___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
Re: [E-devel] E SVN: raster trunk/evas/src/lib/cache
On Wed, Jun 1, 2011 at 9:27 PM, Rafael Antognolli wrote: > On Thu, Sep 9, 2010 at 4:08 AM, Enlightenment SVN > wrote: >> Log: >> lots of cleanups... actually tested! :) >> >> >> Author: raster >> Date: 2010-09-09 00:08:10 -0700 (Thu, 09 Sep 2010) >> New Revision: 52030 > >> static void >> -_evas_preload_thread_end(Evas_Preload_Pthread_Data *pth) >> +_evas_preload_thread_end(void *data) >> { >> - Evas_Preload_Pthread_Data *p; >> + Evas_Preload_Pthread_Data *pth = data; >> + Evas_Preload_Pthread_Data *p = NULL; >> >> - if (pthread_join(pth->thread, (void **) &p) != 0) >> - return ; >> - >> - _evas_preload_thread = eina_list_remove(_evas_preload_thread, pth); >> + if (pthread_join(pth->thread, (void **)&p) != 0) free(p); >> } > > Hello raster, I know this is old, but I can't understand it... why do > you only free the data struct on pthread_join error? Shouldn't you > free it always? > > I can always get leaks on almost every call that I'm doing to > edje_object_preload() on my program. I can send you a test case if > necessary... Test case ! Test case ! Send it ! Send it ! -- Cedric BAIL -- Simplify data backup and recovery for your virtual environment with vRanger. Installation's a snap, and flexible recovery options mean your data is safe, secure and there when you need it. Data protection magic? Nope - It's vRanger. Get your free trial download today. http://p.sf.net/sfu/quest-sfdev2dev ___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel
Re: [E-devel] [Patch] elm_gengrid - added new feature : gengrid reorder mode
On Mon, 30 May 2011 19:41:24 +0900 Jeonghyun Yun said: this seems good. no theme stuff to handle the reorder mode change on the item tho.. shame :( > Dear All. > > I added a new feature "reorder mode" of gengrid refer to genlist reorder > mode. > > After set this reorder API, if user press an item (long press), the item can > be moved another position. > Then if the user releases the item, the item will be reordered in gengrid. > > Please review this patch. > > Thanks. > > [API] > == > /** > * Set the reorder mode > * > * @param obj The Gengrid object > * @param reorder_mode The reorder mode > * (EINA_TRUE = on, EINA_FALSE = off) > * > * @ingroup Gengrid > */ > EAPI void > elm_gengrid_reorder_mode_set(Evas_Object *obj, Eina_Bool reorder_mode) > > /** > * Get the reorder mode > * > * @param obj The Gengrid object > * @return The reorder mode > * (EINA_TRUE = on, EINA_FALSE = off) > * > * @ingroup Gengrid > */ > EAPI Eina_Bool > elm_gengrid_reorder_mode_get(const Evas_Object *obj) > == -- - Codito, ergo sum - "I code, therefore I am" -- The Rasterman (Carsten Haitzler)ras...@rasterman.com -- Simplify data backup and recovery for your virtual environment with vRanger. Installation's a snap, and flexible recovery options mean your data is safe, secure and there when you need it. Data protection magic? Nope - It's vRanger. Get your free trial download today. http://p.sf.net/sfu/quest-sfdev2dev ___ enlightenment-devel mailing list enlightenment-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-devel