- Dillo-dev - lists.mailman3.com

[Dillo-dev]FOSDEM 2005
by Jorge Arellano Cid March 3, 2005

March 3, 2005

Hi there, I'm in Europe these days! At this time FOSDEM 2005 is history, and Dillo was there (embedded track). We had a lot of work with Sebastian and presented Dillo succesfully! The conference was informative and entertaining so people liked it very much. Now I'm in Dortmund and we're working hard with Sebastian on a CVS based prototype with the new Abstraction Layer, so don't expect much emails these days. All the pending work is stalled as we're 100% on this, so patches, extensions, bug-fixes etc will have to wait. Eventually we will come with a renewed Dillo design that provides for stronger basis for the future (think of FLTK, CSS, utf8, text antialiasing, better rendering, etc). -- Cheers Jorge.-

1 0

[Dillo-dev]Patch to handle "td width" issues
by sam+dillo＠chaosring.org March 3, 2005

March 3, 2005

It would be nice if we lived in a world without MSIE (Microsoft Internet Explorer) In particular, it would be nice if we lived in a world without MSIE's broken CSS. For example, there is a wonderful CSS attribute, called max-width, that lets a user specify the maximum width for a given element. Should the user have a narrow window (can you say 640x480), the element in question will just become narrower to fit this window. However, MSIE does not have support for this attribute. This is a feature that only 10% of the browsing population can make use of. The workaround looks like this: <table><tr> <td width=700> The content of a web site. </td></tr></table> The way all the mainstream browsers (MSIE, Opera, Konqueror/Safari, and all the Gecko browsers) handle this is as follows: * They consider the <td width=123> to be a suggestion. This is the maximum possible width for the element in question, as long as the table cell does not have any single elements (such as very long words) that are wider. * If the browser window (viewport) is wide enough to fit this element with the suggested width, the width is unchanged. * If the browser window is not wide enough to fit this element, the table cell will be made narrower to accomodate the narrow browser window. Dillo, however, does not support this /de facto/ standard. The following patch gives Dillo support for this standard. This patch doesn't do the rendering the correct way (which is to make sure all such TDs are not wider than the browser viewport window, barring long table elements), but it does make sure that a single table cell will not be wider than the Dillo browser window. This patch does not handle dynamic resizing, but will ensure that the TD hack will render (mostly) correctly in Dillo. - Sam --- dillo-0.8.4.orig/src/dw_gtk_viewport.c 2004-10-21 15:35:00.000000000 -0700 +++ dillo-0.8.4/src/dw_gtk_viewport.c 2005-03-02 19:04:07.954913184 -0800 @@ -56,6 +56,8 @@ static void Dw_gtk_viewport_adj_changed (GtkAdjustment *adj, GtkDwViewport *viewport); +int ext_window_width = -1; + /* * Standard Gtk+ function */ @@ -302,6 +304,7 @@ widget->allocation.height, viewport->depth); viewport->back_width = widget->allocation.width; + ext_window_width = viewport->back_width; viewport->back_height = widget->allocation.height; DEBUG_MSG (1, " Creating new pixmap, size = %d x %d\n", widget->allocation.width, widget->allocation.height); @@ -1165,6 +1168,7 @@ viewport_area.x = gtk_layout_get_hadjustment(layout)->value; viewport_area.y = gtk_layout_get_vadjustment(layout)->value;; viewport_area.width = widget->allocation.width; + ext_window_width = viewport_area.width; viewport_area.height = widget->allocation.height; if (p_Dw_rectangle_intersect (&viewport->draw_areas[i], --- dillo-0.8.4.orig/src/html.c 2005-01-05 13:01:00.000000000 -0800 +++ dillo-0.8.4/src/html.c 2005-03-02 19:25:16.673038712 -0800 @@ -132,6 +132,8 @@ } TagInfo; static const TagInfo Tags[]; +extern int ext_window_width; + /* * Return the line number of the tag being processed by the parser. */ @@ -1837,6 +1839,22 @@ if ((attrbuf = Html_get_attr(html, tag, tagsize, "width"))) { style_attrs.width = Html_parse_length (html, attrbuf); + /* In Netscape 4/MSIE/Gecko there is a complicated algorithm + * that makes sure all of the td widths are no wider than the + * screen; since MSIE 5/6 doesn't have the CSS max_width + * attribute, web designers still use td width=whatever to specify + * a maximum width. The following makes sure a single table cell + * is no wider than the browser window */ + if(DW_STYLE_IS_ABS_LENGTH(style_attrs.width)) { + int wide; + wide = DW_STYLE_ABS_LENGTH_VAL(style_attrs.width); + printf("------------%d %d\n",wide,ext_window_width); + if(wide > ext_window_width - 10 && + ext_window_width > 300) { + wide = ext_window_width - 10; + style_attrs.width = DW_STYLE_CREATE_ABS_LENGTH(wide); + } + } new_style = TRUE; }

1 0

[Dillo-dev]Frame support
by dominik Feb. 27, 2005

Feb. 27, 2005

hi, does anyone know a patch for current cvs code to enable frame support ? p.s. i saw some older frame-pathces for dillo, what is the problem with it ? -- dominik.holler(a)gmx.net || PGP Key ID: FDEF2A0B

1 0

[Dillo-dev]Patch for Sun C compiler ()
by Jérémy JUST Feb. 25, 2005

Feb. 25, 2005

Hello, To compile Dillo-0.8.4 with Sun C compiler under Solaris 9, I had to apply the attached patch. It appears that Dillo code uses some GCC features that are not in the C language specifications. With my correction, it compiles and seems to work, but there are still many warnings. I haven't tested my patch with GCC or under another OS. Description of my environment: <<<<< $ /opt/SUNWspro/bin/version Machine hardware: sun4u OS version: 5.9 Processor type: sparc Hardware: SUNW,Netra-T12 The following components are installed on your system: Sun Studio 8 Sun Studio 8 C Compiler Sun Studio 8 C++ Compiler Sun Studio 8 Tools.h++ 7.1 Sun Studio 8 C++ Standard 64-bit Class Library Sun Studio 8 Garbage Collector version of "/opt/SUNWspro/bin/cc": Sun C 5.5 2003/03/12 version of "/opt/SUNWspro/bin/CC": Sun C++ 5.5 2003/03/12 >>>>> Greetings. -- Jérémy JUST <jeremy_just(a)netcourrier.com>

1 0

[Dillo-dev]pthread troubles on IRIX
by Riccardo Mottola Feb. 22, 2005

Feb. 22, 2005

Hey, I am compiling on irix 6.5 and have this trouble: gcc -I/usr/freeware/include/gtk-1.2 -I/usr/freeware/include/glib-1.2 -I/usr/freeware/lib32/glib/include -g -O2 -D_REENTRANT -D_THREAD_SAFE -Wall -Waggregate-return -L/usr/local/lib -o dillo -pthread chain.o commands.o cache.o capi.o dw.o dw_aligned_page.o dw_bullet.o dw_button.o dw_container.o dw_embed_gtk.o dw_ext_iterator.o dw_gtk_scrolled_frame.o dw_gtk_scrolled_window.o dw_gtk_statuslabel.o dw_gtk_viewport.o dw_hruler.o dw_image.o dw_list_item.o dw_marshal.o dw_page.o dw_style.o dw_table.o dw_table_cell.o dw_tooltip.o dw_widget.o findtext.o selection.o web.o progressbar.o dillo.o bookmark.o dicache.o dns.o gtk_ext_button.o gtk_ext_menu.o gtk_ext_menu_item.o gtk_menu_title.o gif.o jpeg.o png.o html.o image.o misc.o interface.o history.o nav.o plain.o menu.o prefs.o colors.o bitvec.o klist.o strbuf.o url.o cookies.o dpiapi.o IO/libDio.a -L/usr/freeware/lib32 -lgtk -lgdk -lgmodule -lglib -lintl -lXi -lXext -lX11 -lm gcc: unrecognized option `-pthread' ld32: ERROR 33: Unresolved text symbol "pthread_cancel" -- 1st referenced by IO/libDio.a(IO.o). looking in the config.log: configure:6722: checking whether threads work with -pthread configure:6749: gcc -o conftest -g -O2 -I/usr/local/include -L/usr/local/lib -pthread conftest.c >&5 gcc: unrecognized option `-pthread' configure:6755: $? = 0 I tried replacing -pthread with -lpthread in src/Makefile and in src/IO but it must be elsewhere. I found that I had to modify dillo_LDFLAGS = -lpthread too but I get at the end: -lm ld32: WARNING 84: /usr/lib32/libpthread.so is not used for resolving any symbol. ld32: ERROR 33: Unresolved text symbol "pthread_cancel" -- 1st referenced by IO/libDio.a(IO.o). Use linker option -v to see when and which objects, archives and dsos are loaded. so now I am still stuck :( Cheers, Riccardo

1 1

[Dillo-dev]help on dillo
by �� JeN �� Feb. 22, 2005

Feb. 22, 2005

good day! i would like to ask for your help. i am a computer science student and currently, doing my thesis. i, together with my group mates is implementing a system in linux that enables to add plug-ins (a plug-in can be an application or a functionality of the system) using shared objects. If we are to add, for example, an existing application in our system it has to be in the .so form. Dillo is one of the examples we considered to add in our system. It is also one way of testing if the plug-in function we are implementing really works. Regarding the dillo browser, we tried to make it as a shared object file using the command line "gcc -g -shared -Wl,-soname,dillo.so.0 \-o dillo.so dillo.o -lc" the problem is there are certain methods that cannot be recognized when we run the shared object that is produced using our system. those methods are undefined according to the system. We do not know if the paths of the libraries used by dillo have something to do with the error... or something else. we would like to ask if our way of making a shared object file is correct or applicable to dillo. if not, can u tell us how to make it as a shared object file - the proper way of making a shared object file. Also, can u let us know how to compile dillo manually, without using the make file. if you also have an advice regarding the problem/solution to the problem, we would appreciate it very much. thank you very much, sir. - jen __________________________________ Do you Yahoo!? The all-new My Yahoo! - What will yours do? http://my.yahoo.com

1 0

[Dillo-dev]Please help me with some data
by Jorge Arellano Cid Feb. 22, 2005

Feb. 22, 2005

Hi guys, This weeks have taken a hectic pace for me, finishing and taking care of a lot of issues has been the workload of each day. Now almost everything is ready for FOSDEM 2005 and a very long trip. The point: ;) I'm gathering more data to tune and enrich some points of the conference. The skeleton is: Presentation contents: * What's Dillo * Project objectives * Compatibility/Portability Dillo on PDAs: iPAQ, Yopi, Psion (familiar). uclibc * Architecture overview: Parser/Rendering/IO/Plugins * Current state of the project: Supported standards Features HTML bug meter Memory usage: Comparison table with other browsers by VSZ/RSS. CPU usage: Comparative wall-clock times for page rendering. * Near future potential The port to FLTK Antialiased fonts, UTF-8, more speed, smaller footprint, rendering to abstract devices. CSS, XHTML validation, more SSL, ... I need more data on: Memory usage: Comparison table with other browsers by VSZ/RSS. CPU usage: Comparative wall-clock times for page rendering. and uclibc This is: -------- * Comparison table with other browsers by VSZ/RSS.: On the same system, how much VSZ/RSS Dillo uses in comparison with Mozilla, Firefox, Opera, Whatever on startup and on loading a sample page. * Comparative wall-clock times for page rendering. Wall clock times for rendering a sample page. * uclibc I know Dillo has run under ucLibc. I need the system spec, RSS and compared with other browsers that may also work on the environment. That's it. Please post your answers to dillo-dev so everyone can see the collected data. A lot of TIA! -- Cheers Jorge.-

2 1

[Dillo-dev]Another refresh patch
by Horst Gutmann Feb. 22, 2005

Feb. 22, 2005

Hi everyone :-) I also needed some meta-refresh functionality because I use meta-refresh as an "offline"-alternative for 30x-return codes. Basically I just wanted to ask how ugly the code is and if I'm at least on the right way :-) Currently it simply ignores the delay-value because I don't really need it for this project. MfG, Horst

1 1

[Dillo-dev]testing https plugin gave problems
by Peter Wood Feb. 21, 2005

Feb. 21, 2005

Hi First, thanks to all for a great browser! The following applies to dillo-0.8.4 on linux with glibc2. I tried to test the https plugin by removing line 56 from dpi/https.c, as suggested. However, the linker complained about unresolved symbols when I recompiled. I solved that problem by changing the order of LIBSSL_LIBS in configure.in to "-lssl -lcrypto". Dillo then recompiled fine, however I still get the default message about the prototype plugin for https being disabled. Is there something else I should do. Regards, Peter

2 1

[Dillo-dev]dillo-0.8.4 ported to ipaq familiar-X11
by Michael Kropfberger Feb. 16, 2005

Feb. 16, 2005

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi everyone! familiar comes with dillo-0.6.6 which is even older than old, so I've cross-compiled dillo-0.8.4 with some extra patches: * when started in fullscreen, no toggle-button is visible * no cursor is visible, when the mouse is over the dillo window * the great mime-type patch for external viewers (from <http://www-fourier.ujf-grenoble.fr/~franz/dillo/>) * it reloads all urls (cache disabled for url rewriting, since it ignores the according meta tags) since dillo still is based on gtk+-1.2, this is a little bloated, especially because now there are extra X11-fonts needed... yes, I know, familiar comes with dillo-0.6.6 ported to gtk2 and uses built-in pango fonts, but this was too much work for me now :) I've also packaged a very tiny XFree86-font selection, so it really becomes usable... grab it here: http://www-itec.uni-klu.ac.at/~mike/ipaq/xfonts-tiny-0.1.tar.gz for fast installation on your familiar-0.8 ipaq, grab the font .ipk at: http://www-itec.uni-klu.ac.at/~mike/ipaq/xfonts-tiny_0.1-r0_all.ipk and the dillo .ipk at: http://www-itec.uni-klu.ac.at/~mike/ipaq/dillo_0.8.4-r0_arm.ipk for the OpenEmbedded maintainers, here are the needed bitbake .bb files: http://www-itec.uni-klu.ac.at/~mike/ipaq/my-new-dillo-oe.tgz I hope you all like it! greets, Mike -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.4 (GNU/Linux) iD8DBQFCE4gxw6IyDgGSsNIRAsl0AJ4kIbSuaU+OUAAtZmtk5/Fbo2V97ACfTnP3 A3xWwswJSz9wiqIpaqrqc5s= =Aj25 -----END PGP SIGNATURE-----

1 0