- Dillo-dev - lists.mailman3.com

[patch] improve scroll performance of dillo-fltk

by Johannes Hofmann

Hi, below is an initial patch that improves scrolling performance of dillo-fltk by copying screen data instead of rerendering it. It's based on Scroll.cxx from fltk. The difference is noticeable mostly when scrolling large images especially within a remote X11 session. The newly introduced scrollDX / scrollDY could probabely be combined with the existing oldScrollX / oldScrollY variables. Regards, Johannes diff --git a/dw-testbed-0.0.43j/dw/fltkviewbase.cc b/dw-testbed-0.0.43j/dw/fltkviewbase.cc --- a/dw-testbed-0.0.43j/dw/fltkviewbase.cc +++ b/dw-testbed-0.0.43j/dw/fltkviewbase.cc @@ -54,18 +54,16 @@ void FltkViewBase::draw () Group::draw (); - if (damage () & DAMAGE_EXPOSE) { - setcolor (bgColor); - ::fltk::Rectangle rect (0, 0, w(), h()); - fillrect (rect); - - core::Rectangle area; - area.x = translateViewXToCanvasX (0); - area.y = translateViewYToCanvasY (0); - area.width = translateViewXToCanvasX (0 + w ()) - 0; - area.height = translateViewYToCanvasY (0 + h ()) - 0; - theLayout->expose (this, &area); - } + setcolor (bgColor); + ::fltk::Rectangle rect (0, 0, w(), h()); + fillrect (rect); + + core::Rectangle area; + area.x = translateViewXToCanvasX (0); + area.y = translateViewYToCanvasY (0); + area.width = translateViewXToCanvasX (0 + w ()) - 0; + area.height = translateViewYToCanvasY (0 + h ()) - 0; + theLayout->expose (this, &area); } core::ButtonState getDwButtonState () diff --git a/dw-testbed-0.0.43j/dw/fltkviewport.cc b/dw-testbed-0.0.43j/dw/fltkviewport.cc --- a/dw-testbed-0.0.43j/dw/fltkviewport.cc +++ b/dw-testbed-0.0.43j/dw/fltkviewport.cc @@ -131,8 +131,8 @@ void FltkViewport::hscrollbarChanged () { int oldScrollX = scrollX; scrollX = hscrollbar->value (); - /** \bug Inefficient */ - redraw (DAMAGE_EXPOSE); + scrollDX = scrollX - oldScrollX; + redraw (DAMAGE_SCROLL); updateCanvasWidgets (oldScrollX, scrollY); theLayout->scrollPosChanged (this, scrollX, scrollY); } @@ -141,8 +141,8 @@ void FltkViewport::vscrollbarChanged () { int oldScrollY = scrollY; scrollY = vscrollbar->value (); - /** \bug Inefficient */ - redraw (DAMAGE_EXPOSE); + scrollDY = scrollY - oldScrollY; + redraw (DAMAGE_SCROLL); updateCanvasWidgets (scrollX, oldScrollY); theLayout->scrollPosChanged (this, scrollX, scrollY); } @@ -163,6 +163,25 @@ void FltkViewport::layout () { theLayout->viewportSizeChanged (this, w(), h()); adjustScrollbarsAndGadgetsAllocation (); +} + +void FltkViewport::draw_area (void *data, const Rectangle& cr ) +{ + FltkViewport *vp = (FltkViewport*) data; + push_clip(cr); + + vp->FltkWidgetView::draw (); + + //draw_child (*hscrollbar); + //draw_child (*vscrollbar); + + for(Iterator <TypedPointer < ::fltk::Widget> > it = vp->gadgets->iterator (); + it.hasNext (); ) { + ::fltk::Widget *widget = it.getNext()->getTypedValue (); + vp->draw_child (*widget); + } + + pop_clip(); } void FltkViewport::draw () @@ -170,20 +189,16 @@ void FltkViewport::draw () int hdiff = vscrollbar->visible () ? SCROLLBAR_THICKNESS : 0; int vdiff = hscrollbar->visible () ? SCROLLBAR_THICKNESS : 0; Rectangle cr (0, 0, w () - hdiff, h () - vdiff); - push_clip(cr); - - FltkWidgetView::draw (); - - //draw_child (*hscrollbar); - //draw_child (*vscrollbar); - - for(Iterator <TypedPointer < ::fltk::Widget> > it = gadgets->iterator (); - it.hasNext (); ) { - ::fltk::Widget *widget = it.getNext()->getTypedValue (); - draw_child (*widget); - } - - pop_clip(); + int d = damage(); + + if (d == DAMAGE_SCROLL) { + scrollrect(cr, -scrollDX, -scrollDY, draw_area, this); + } else { + draw_area(this, cr); + } + + scrollDX = 0; + scrollDY = 0; } int FltkViewport::handle (int event) @@ -205,8 +220,9 @@ int FltkViewport::handle (int event) case PageUpKey: oldScrollY = scrollY; scrollY -= vscrollbar->pagesize (); + scrollDY = scrollY - oldScrollY; adjustScrollbarValues (); - redraw (DAMAGE_EXPOSE); + redraw (DAMAGE_SCROLL); updateCanvasWidgets (scrollX, oldScrollY); theLayout->scrollPosChanged (this, scrollX, scrollY); return 1; @@ -214,8 +230,9 @@ int FltkViewport::handle (int event) case PageDownKey: oldScrollY = scrollY; scrollY += vscrollbar->pagesize (); + scrollDY = scrollY - oldScrollY; adjustScrollbarValues (); - redraw (DAMAGE_EXPOSE); + redraw (DAMAGE_SCROLL); updateCanvasWidgets (scrollX, oldScrollY); theLayout->scrollPosChanged (this, scrollX, scrollY); return 1; @@ -254,12 +271,13 @@ void FltkViewport::scrollTo (int x, int void FltkViewport::scrollTo (int x, int y) { int oldScrollX = scrollX; - //int oldScrollY = scrollY; + int oldScrollY = scrollY; scrollX = x; scrollY = y; + scrollDX = scrollX - oldScrollX; + scrollDY = scrollY - oldScrollY; adjustScrollbarValues (); - /** \bug Inefficient */ - redraw (DAMAGE_EXPOSE); + redraw (DAMAGE_SCROLL); updateCanvasWidgets (oldScrollX, scrollY); } diff --git a/dw-testbed-0.0.43j/dw/fltkviewport.hh b/dw-testbed-0.0.43j/dw/fltkviewport.hh --- a/dw-testbed-0.0.43j/dw/fltkviewport.hh +++ b/dw-testbed-0.0.43j/dw/fltkviewport.hh @@ -20,6 +20,7 @@ private: enum { SCROLLBAR_THICKNESS = 15 }; int scrollX, scrollY; + int scrollDX, scrollDY; ::fltk::Scrollbar *vscrollbar, *hscrollbar; @@ -35,6 +36,7 @@ private: static void vscrollbarCallback (Widget *vscrollbar, void *viewportPtr); void updateCanvasWidgets (int oldScrollX, int oldScrollY); + static void draw_area (void *data, const Rectangle& cr); protected: int translateViewXToCanvasX (int x);

Oct. 5, 2007

Re: [Dillo-dev] Dillo FLTK2 alpha release for developers

by Vincent Thomasset

Le Fri, 5 Oct 2007 04:58:33 +0200, Vincent Thomasset <vincent.thomasset(a)skynet.be> a écrit : > Le Tue, 2 Oct 2007 22:23:19 +0000, > higuita <higuita7(a)yahoo.co.uk> a écrit : > > > On Tue, 2 Oct 2007 16:42:16 -0400, wrote: > > > > Now it starts, but when i try to load a page it crashes: > > I get the same backtrace here on x86_64 too. > > I think this link may have some clue as to why this occurs: > http://sourceware.org/ml/glibc-bugs/2007-08/msg00148.html Confirmed, copying the va_list in dStr_vsprintfa (dlib.c) fixes the crash for me and i'm now able to display pages. Cheers, Vincent

Oct. 5, 2007

Re: [Dillo-dev] Dillo FLTK2 alpha release for developers

by Vincent Thomasset

Le Tue, 2 Oct 2007 22:23:19 +0000, higuita <higuita7(a)yahoo.co.uk> a écrit : > On Tue, 2 Oct 2007 16:42:16 -0400, wrote: > > Now it starts, but when i try to load a page it crashes: I get the same backtrace here on x86_64 too. I think this link may have some clue as to why this occurs: http://sourceware.org/ml/glibc-bugs/2007-08/msg00148.html Cheers, Vincent

Oct. 5, 2007

Re: [Dillo-dev] Dillo FLTK2 alpha release for developers

by higuita

Hi again On Tue, 2 Oct 2007 18:45:07 -0400, wrote: > Yes, please send your patch to check suspicious parts. in attach are the 2 patchs, for dw-testbed and dillo thanks for the help higuita -- Naturally the common people don't want war... but after all it is the leaders of a country who determine the policy, and it is always a simple matter to drag the people along, whether it is a democracy, or a fascist dictatorship, or a parliament, or a communist dictatorship. Voice or no voice, the people can always be brought to the bidding of the leaders. That is easy. All you have to do is tell them they are being attacked, and denounce the pacifists for lack of patriotism and exposing the country to danger. It works the same in every country. -- Hermann Goering, Nazi and war criminal, 1883-1946

Oct. 3, 2007

Re: [Dillo-dev] Dillo FLTK2 alpha release for developers

by Andreas Kemnade

Hi Jorge, That are really great news. I had some compiling problems. in dillo source code src/IO/Makefile.am and dpi/Makefile.am needs a AM_CXXFLAGS = @LIBFLTK_CXXFLAGS@ Same for dw-testbed/test/Makefile.am The other problem is that I need to comment out the tablet support in fltk due to missing libXi. Ok, then I got it to compile finally for my ipaq with familar 0.8.4 on it. I runs with X-forwarding so that I get the window on my notebook display. But it segfaults without x forwarding: dillo_dns_init: Here we go! (threaded) Disabling cookies. bad font: variable Segmentation fault I think I have to check my font setup, Bitstream Vera seems to be available... Greetings Andreas Kemnade

Oct. 3, 2007

Re: [Dillo-dev] Re: Dillo FLTK2 alpha release for developers

by Alexander Becher

* Johannes Hofmann: >On Wed, Oct 03, 2007 at 02:11:56PM +0200, Alexander Becher wrote: >> I compiled dillo-fltk2 on NetBSD-3.1, implementing strndup was the only >> change needed. However, strndup needs to null-terminate its return >> value, and strncpy does not do this if the source string is longer than >> the destination, so: > >As calloc() initializes the allocated buffer with 0, this is not >needed, I think. Hm, right. I had used malloc() in my replacement function, but with calloc() it is indeed unnecessary. I have wrapped this in autoconf, see attached patch. I also made autoconf in dillo and dw-testbed check for the path to the perl binary and substitute it in dpidc and Doxyfile, respectively, see other patch. Actually, these patches can't simply be applied: first, dillo-f15/dpid/dpidc and dw-testbed-0.0.43j/Doxyfile have to be moved to *.in. Regards, Alexander -- PGP key available citator(a)jabber.ccc.de

Oct. 3, 2007

Re: [Dillo-dev] Re: Dillo FLTK2 alpha release for developers

by Johannes Hofmann

Hi Alexander, On Wed, Oct 03, 2007 at 02:11:56PM +0200, Alexander Becher wrote: > Hi, > > I compiled dillo-fltk2 on NetBSD-3.1, implementing strndup was the only > change needed. However, strndup needs to null-terminate its return > value, and strncpy does not do this if the source string is longer than > the destination, so: As calloc() initializes the allocated buffer with 0, this is not needed, I think. Johannes > > * Johannes Hofmann: > > > >* I added a own version of strndup() as it does not exist for > > DragonFly: > > > >char* strndup(const char *s, size_t size) { > > char *r = (char*) calloc(size + 1, 1); > > > > if (r) { > > strncpy(r, s, size); > r[size] = '\0'; /* null-terminate return value */ > > } > > > > return r; > >} > > Also, strndup is found in <string.h> only if _GNU_SOURCE was #define'd. > This should probably be done in configure, with something like > HAVE_STRNDUP, but I have no experience with autoconf. > > Regards, > Alexander > -- > PGP key available > > _______________________________________________ > Dillo-dev mailing list > Dillo-dev(a)dillo.org > http://lists.auriga.wearlab.de/cgi-bin/mailman/listinfo/dillo-dev

Oct. 3, 2007

Re: Dillo FLTK2 alpha release for developers

by Alexander Becher

Hi, I compiled dillo-fltk2 on NetBSD-3.1, implementing strndup was the only change needed. However, strndup needs to null-terminate its return value, and strncpy does not do this if the source string is longer than the destination, so: * Johannes Hofmann: > >* I added a own version of strndup() as it does not exist for > DragonFly: > >char* strndup(const char *s, size_t size) { > char *r = (char*) calloc(size + 1, 1); > > if (r) { > strncpy(r, s, size); r[size] = '\0'; /* null-terminate return value */ > } > > return r; >} Also, strndup is found in <string.h> only if _GNU_SOURCE was #define'd. This should probably be done in configure, with something like HAVE_STRNDUP, but I have no experience with autoconf. Regards, Alexander -- PGP key available

Oct. 3, 2007

Re: [Dillo-dev] Dillo FLTK2 alpha release for developers

by None

On Tue, Oct 02, 2007 at 10:23:19PM +0000, higuita wrote: > On Tue, 2 Oct 2007 16:42:16 -0400, wrote: > > -Wno-unused-parameter -MT ui.o -MD -MP -MF .deps/ui.Tpo -c -o ui.o ui.cc > > > ui.cc: In function 'void b1_cb(fltk::Widget*, void*)': > > > ui.cc:235: error: cast from 'void*' to 'int' loses precision > > > make[3]: *** [ui.o] Error 1 > > > > Try: > > - int bn = (int)cb_data; > > + int bn = (long)cb_data; > > > > If this works OK, the same may be used for C++ in Dw (i.e. default to > > long). > > Thanks, it works!! it allowed the build to continue > it gave several errors like this and i fix then the same way. > > I can extract a patch and send it if you want. Yes, please send your patch to check suspicious parts. -- Cheers Jorge.-

Oct. 2, 2007

https and OpenBSD

by None

Hi, > Dear Jorge & dillo fans, > > I can confirm this will run and build on OpenBSD-4.1 with minor > modifications: > > * remove the GL & glut targets from the fltk Makefile if you don't > need it Did you just "./configure --disable-gl"? Or: - DIRS = src $(LOCALIMAGES) images OpenGL fluid glut test + DIRS = src $(LOCALIMAGES) images fluid test else? > * implement a strndup() function in dw/selection.cc in the testbed > as OpenBSD doesn't have that in string.h Yes. > > I still have one question: > > I wanted to try out https (yes I know the README says it doesn't > work). And of course it did not work: It works bery basically. If you can see "https://www.rapidssl.com/test/rapidssl.htm", that's it. > > prefs: Using internal defaults. > dillo_dns_init: Here we go! (threaded) > Enabling cookies as from cookiesrc... > Nav_open_url: Url=>https://www.ccc.de< > [dpid]: dpi.c:70: a_Dpi_rd_dpi_socket_dir: fopen > Dpi_blocking_start_dpid: try 1 > [dpid]: No such file or directory > [dpid]: - /home/ckeen/.dillo/dpi_socket_dir > [dpid]: init_sockdir: The dpi_socket_dir file (null) does not exist > dpid started > a_Dpiapi_dialog: > dpip_tag: <cmd='dialog' msg='The remote certificate cannot be > verified (code 20)' alt1='Continue' alt2='Cancel' '> > mouse button 1 was pressed > a_UIcmd_stop() > > > Is the backward channel form dpis supposed to work? Yes. > If yes I will > look into it. If there is too much missing atm, I would be gladful > for a hint on where to start. The point here is that https is a "filter" dpi (i.e. one instance per request is spawned). The solution is to implement it as a "server" (e.g. like the file server). That way, when a https request has many embedded resources (i.e. many https requests to the same site), the server can make only one query and not flood the backward channel for dpis. With regard to the actual certificate-validating code, AFAIR much of it is already there, it needs some testing and enabling. It would be great if you can work on it. > Thank you very _very_ much for this nice programme and for bearing > with us! Thank you for offering your help. -- Cheers Jorge.-

Oct. 2, 2007