[Dillo-dev]Dillo for the ipaq From: Eduardo Marcel Macan - 2001-06-29 17:48 Guys thank you for the *wonderful* work you are doing with dillo. I am working with the guys of the familiar distribution (familiar.handhelds.org) and dillo is becoming the best choice for use in a limited storage device like the ipaq. I have just made a package for dillo out of today's snapshot, and I must say It is just *GREAT* . here is the .ipk package of dillo for the ipaq using the familiar GNU/Linux distribution. http://www.metainfo.org/macan/ipaq/dillo_20010629-1_arm.ipk And here is the screenshot of dillo on an iPAQ displaying slashdot: http://www.metainfo.org/macan/ipaq/10-ipaq-dillo_slashdot.png I am leaving for a 10 day break now, but when I am back I hope to start contributing more to the dillo web browser! :) Thank you all! -- "If you have an apple and I have an apple and we exchange apples then you and I will still each have one apple. But if you have an idea and I have an idea and we exchange these ideas, then each of us will have two ideas." George Bernard Shaw [Dillo-dev]Feature From: Sebastian Geerken - 2001-06-27 13:45 Hi iPAQ users (and others). Since tables may be used to set the width at which words are wrapped (either explicitly by the "width" attribute, or implicitly by images and long words), the text may become unreadable on small screens. Dillo could change the rendering to prevent this, one simple possibility is new option, "limit_text_width", which limits the word wrap width (but not necessary the final page width) to the viewport width. I've implemented this, it is available via CVS. I've tested it a bit, browsing with a small window size. It is probably a bit better than before, but don't expect too much, you loose the overview over the page easily. There may be new options (e.g. factors for size specification and images), but they will not be part of the (soon released) version 0.6.0. Please post your comments. Does this feature make sense? Any further ideas? Should dillo try to render such pages at all in a "readable" way? Etc. Sebastian [Dillo-dev]Dillo-ssl (patch) From: - 2001-06-26 21:27 Attachments: dillo-ssl.patch.gz https.c.gz Hi all. Nice work with dillo! About the only browser that is lightweight enough for my iPAQ. However... i kinda need cookies, https and auth for it to be really useable. I hacked together some kind of SSL support for it, and tested against apache/mod_ssl. Seems to be functional. Included are two files, a cvs diff against yesterdays cvs, and a new file "https.c" which goes in IO/. Add the -I's and -l's for openssl (in my case 0.9.5a, yeah, it's old, i know ;) Since it's my first attempt at both dillo and ssl hacking, the approach might be a bit off. Basically it's a combination of about: and http: backends with some SSL inbetween. Now all i gotta do is to find a place in the GUI for a padlock ;) ...and ofcourse implement cookies and auth. Best regards, T-Å Re: [Dillo-dev]locale From: Martynas Jocius - 2001-06-23 16:50 Hi, Sebastian On Sat, Jun 23, 2001 at 03:35:33PM +0200, Sebastian Geerken wrote: > gtk_set_locale is intended for the user interface, but it probably > works for this purpose, if your "system" language ($LANG and $LC_...) > and the language of the documents you are viewing are the same. > I am not talking about the documents. Bug 173 was inserted by me :). I am from Lithuania, small country in Europe near the Baltic Sea. Native charset is iso-8859-13. I have no problems on viewing any documents with lithuanian characters. But I was talking about Dillo. For example, I bookmark a page and set a title with lithuanian characters. I can't see them. Or I have to type some text with my native characters into the entry in the html page. I can't do that too... Well, now it's ok, I fixed this :) > BTW, there has been a discussion on the list, you can find it in the > archive at: [some addresses] Yes, discussian was about documents with Korean/Russian characters. As I said, I have no problems at this point ;) > Perhaps we can already implement a simple scheme dealing with 8-bit > fonts, looking at what Gtk+ 1.2 already provides. Extending > DwStyleFont by an encoding attribute might works. (Sorry, I'm no > expert on this topic at all.) So what about gtk_set_locale()? I think it's a good way. Maybe someday we will make Dillo to understand i18n for the translations. -- Martynas Re: [Dillo-dev]locale From: Sebastian Geerken - 2001-06-23 13:30 On Fri, Jun 22, Martynas Jocius wrote: > Bug #173. Can't see/write native characters. Insert gtk_set_locale() into > src/dillo.c line 80. :-) gtk_set_locale is intended for the user interface, but it probably works for this purpose, if your "system" language ($LANG and $LC_...) and the language of the documents you are viewing are the same. BTW, there has been a discussion on the list, you can find it in the archive at: http://www.geocrawler.com/archives/3/702/2001/5/0/5805536/ http://www.geocrawler.com/archives/3/702/2001/5/0/5806440/ http://www.geocrawler.com/archives/3/702/2001/5/0/5811325/ http://www.geocrawler.com/archives/3/702/2001/5/0/5812833/ http://www.geocrawler.com/archives/3/702/2001/5/0/5830229/ http://www.geocrawler.com/archives/3/702/2001/5/0/5833416/ Perhaps we can already implement a simple scheme dealing with 8-bit fonts, looking at what Gtk+ 1.2 already provides. Extending DwStyleFont by an encoding attribute might works. (Sorry, I'm no expert on this topic at all.) Sebastian [Dillo-dev]DwStyle memory leaks From: Sebastian Geerken - 2001-06-22 16:13 Hi, just a hint on warnings like "** WARNING **: 1 styles left": It seems that the memory management works, except if the HTML parser is interrupted. In this case, Html_close is not called, which will normally unref all styles left in DilloHtml. Sebastian Re: [Dillo-dev]HTML/CSS question From: Sebastian Geerken - 2001-06-22 16:13 On Fri, Jun 22, Jörgen Viksell wrote: > >does someone know how to replace the NOSHADE attribute of
by CSS > >attributes? In HTML4, it's deprecated, but there is no hint how to get > >the same effect using CSS. > > This would make it look similar: > border: none; background-color: grey I believe the ruler itself is the content, so it won't work. Test the following in Amaya (I've installed 3.1):
The first two look equally, the third has a gray rectangle *around* the ruler. Sebastian Re: [Dillo-dev]HTML/CSS question From: Viksell - 2001-06-22 12:23 >does someone know how to replace the NOSHADE attribute of
by CSS >attributes? In HTML4, it's deprecated, but there is no hint how to get >the same effect using CSS. This would make it look similar: border: none; background-color: grey // J=F6rgen [Dillo-dev]locale From: Martynas Jocius - 2001-06-22 12:07 Hi, Bug #173. Can't see/write native characters. Insert gtk_set_locale() into src/dillo.c line 80. :-) -- Martynas [Dillo-dev]HTML/CSS question From: Sebastian Geerken - 2001-06-22 11:29 Hi, does someone know how to replace the NOSHADE attribute of
by CSS attributes? In HTML4, it's deprecated, but there is no hint how to get the same effect using CSS. (I'm working on bug #168 (colors of
's), this could be solved simple by DwStyle boxes, but I thought of making it already compliant with a future CSS implementation, which will be based on DwStyle.) Sebastian Re: [Dillo-dev][patch] File -> Close Window From: Sam Dennis - 2001-06-21 16:12 On Thu, Jun 21, 2001 at 02:04:48PM +0300, Martynas Jocius wrote: > I made another close window patch, maybe identical with last one. But I saw, > that the memory is freed when destroying the window. Interface_quit() > function is called when I do gtk_widget_destroy(). How does this work? There's this little line in interface.c which reads: gtk_signal_connect(GTK_OBJECT(bw->main_window), "destroy", GTK_SIGNAL_FUNC(Interface_quit), bw); So when the window is destroyed, gtk calls Interface_quit. Re: [Dillo-dev][patch] File -> Close Window From: Martynas Jocius - 2001-06-21 12:00 Attachments: dillo-cvs-06-21-close_window.patch On Mon, Jun 18, 2001 at 09:22:38AM -0400, Jorge Arellano Cid wrote: > Is very encouraging to see new people trying to get involved > into dillo development, please don't get discouraged if you fail > at your first attempts; it was the same for everyone of us ;-) Thanks Jorge :) I made another close window patch, maybe identical with last one. But I saw, that the memory is freed when destroying the window. Interface_quit() function is called when I do gtk_widget_destroy(). How does this work? -- Martynas [Dillo-dev]RGBA data (was: Misc) From: Sebastian Geerken - 2001-06-20 15:48 On Tue, Jun 19, I wrote: > - Image transparency 2: If you have something like > > > >
> >
> > the first image will, correctly, have have a red background, but > the second, incorrectly, too, since they share the same dicache > buffer. So the transparency should somehow stored in the dicache > buffer, instead of replaced by a background color. If there are no objections, I'll start to modify the image code to use RGBA data instead of RGB. Sebastian [Dillo-dev]Misc From: Sebastian Geerken - 2001-06-19 16:45 Hi all, I have been working on borders and colors, now they work for table (just uploaded), although a bit polishing is needed. However, some notes: - What happened to the implementation? I've now activated (but not face), of course inside an "if (!prefs.force_my_colors)", since this makes some colored tables more readable. - Image transparency 1: I've modified Png_datarow_callback to use the background color passed to a_Image_new, by replacing prefs.bg_color by png->Image->bg_color. I don't know very much about image processing, but this seems the way this should be handled (?). - Image transparency 2: If you have something like
the first image will, correctly, have have a red background, but the second, incorrectly, too, since they share the same dicache buffer. So the transparency should somehow stored in the dicache buffer, instead of replaced by a background color. Sebastian Re: [Dillo-dev][patch] File -> Close Window From: Jorge Arellano Cid - 2001-06-18 13:52 Martynas, > There was a bug #87 about closing window. There's a small patch which fixes > this. Is very encouraging to see new people trying to get involved into dillo development, please don't get discouraged if you fail at your first attempts; it was the same for everyone of us ;-) I ask, at the web page, to be sent information about personal skills, but I understand if they don't send any. That's just to know in which part of dillo their knowledge would fit better, and also to guide or hint their progress. Jorge.- Re: [Dillo-dev]tables: COLSPAN From: Sebastian Geerken - 2001-06-14 22:58 On Thu, Jun 14, Sam Dennis wrote: > On Thu, Jun 14, 2001 at 06:59:05PM +0200, Sebastian Geerken wrote: > [...] > > Fixed and committed. > > > > Sebastian > > I'm sorry but it doesn't look very fixed to me, I'm getting almost exactly the > same results here. What images do you use? I tested it with
and got +---------+ +---+ | | | | +---------+ | | +---+ +---+ | | | | | | | | | | +---+ +---+ | | +---------+ | | | | +---+ +---------+ With some certain image sizes, the result is indeed not perfect, see example in doc/DwTable.txt. Sebastian Re: [Dillo-dev]tables: COLSPAN From: Sam Dennis - 2001-06-14 21:35 On Thu, Jun 14, 2001 at 06:59:05PM +0200, Sebastian Geerken wrote: > On Wed, Jun 13, Sam Dennis wrote: > > It may have been a while since I wrote any real HTML, but I think I'd be > > correct in thinking that the following table shouldn't be rendered as > > illustrated: > > [...] > > Fixed and committed. > > Sebastian I'm sorry but it doesn't look very fixed to me, I'm getting almost exactly the same results here. Re: [Dillo-dev]tables: COLSPAN From: Sebastian Geerken - 2001-06-14 16:54 On Wed, Jun 13, Sam Dennis wrote: > It may have been a while since I wrote any real HTML, but I think I'd be > correct in thinking that the following table shouldn't be rendered as > illustrated: > [...] Fixed and committed. Sebastian Re: [Dillo-dev][patch] File -> Close Window From: Sam Dennis - 2001-06-14 11:59 On Thu, Jun 14, 2001 at 01:39:05PM +0300, Martynas Jocius wrote: > On Thu, Jun 14, 2001 at 11:56:38AM +0100, Sam Dennis wrote: > > > When a window is closed, significantly more needs to be done than simply > > destroying the window. Interface_quit performs all the necessary actions and > > should almost certainly be called when the close window function is used. > > Ok, sorry. I'll try to fix it. Just I'm very new in Dillo sources. > What about adding: > if (bw->main_window != NULL) > gtk_widget_destroy(bw->main_window); > > into Interface_quit? And then call Interface_quit when the menu item is > clicked? This will require a lot more changes in sources. Sorry about this, but ignore everything I said, I'm not at my sharpest today :) Interface_quit is the handler for destroying the main window, so you don't have to worry about calling it. Re: [Dillo-dev][patch] File -> Close Window From: Martynas Jocius - 2001-06-14 11:35 On Thu, Jun 14, 2001 at 11:56:38AM +0100, Sam Dennis wrote: > When a window is closed, significantly more needs to be done than simply > destroying the window. Interface_quit performs all the necessary actions and > should almost certainly be called when the close window function is used. Ok, sorry. I'll try to fix it. Just I'm very new in Dillo sources. What about adding: if (bw->main_window != NULL) gtk_widget_destroy(bw->main_window); into Interface_quit? And then call Interface_quit when the menu item is clicked? This will require a lot more changes in sources. -- Martynas Re: [Dillo-dev][patch] File -> Close Window From: Sam Dennis - 2001-06-14 10:47 On Thu, Jun 14, 2001 at 12:28:57PM +0300, Martynas Jocius wrote: > Hi, > > There was a bug #87 about closing window. There's a small patch which fixes > this. > > -- > Martynas > /* > - * ? > + * Close single window. Exit Dillo if only one window is. > */ > void a_Commands_close_callback(GtkWidget *widget, gpointer client_data) > { > + BrowserWindow *bw = (BrowserWindow *)client_data; > + Interface_destroy_window(bw->main_window); > } When a window is closed, significantly more needs to be done than simply destroying the window. Interface_quit performs all the necessary actions and should almost certainly be called when the close window function is used. [Dillo-dev][patch] File -> Close Window From: Martynas Jocius - 2001-06-14 10:25 Attachments: dillo-cvs-close_window.atthis.patch Hi, There was a bug #87 about closing window. There's a small patch which fixes this. -- Martynas Re: [Dillo-dev]table algorithm (was: tables: COLSPAN) From: Sam Dennis - 2001-06-14 10:10 On Fri, Jan 05, 1990 at 01:34:08AM -0900, Jon Bradley wrote: > > It is impossible to access the CVS. I have done it before, back when > 0.3.0 was release. But it appears all information about the CVS is gone > from the website AND the knownledge base. > > Have you closed down the CVS to the public? I would like to see how > dillo is looking with frames. AFAIK, anonymous CVS is working perfectly, and I'm afraid that you won't be seeing frames yet, only tables. Although my frames code should be ready in the not-so-distant future... Re: [Dillo-dev]tables: COLSPAN From: Sebastian Geerken - 2001-06-14 09:15 On Wed, Jun 13, Sam Dennis wrote: > I was going to congratulate you on actually making that system work, but now, > having actually seen its output, I've decided that it needs a bit more work. This is what bug reports are for. :-) > It may have been a while since I wrote any real HTML, but I think I'd be > correct in thinking that the following table shouldn't be rendered as > illustrated: > [...] This one is fixed, but I've not yet uploaded the change, there are still a few more bugs. Sebastian Re: [Dillo-dev]dillo CVS From: Jon Bradley - 2001-06-13 21:58 Please disgrad the last message I just sent about an impossible CVS. =-=-=-=-=-=-=-=-=-=-=-= Email: kreator@gc... bbs: telnet://toga.cx =-=-=-=-=-=-=-=-=-=-=-= On 13 Jun 2001, Sebastian Geerken wrote: > On Thu, Jan 04, Jon Bradley wrote: > > > > Hi, > > > > I am trying to login to the dillo CVS, so I searched the dillo website, > > where I found no info on the CVS. I then got to the knowledge base of > > Mail Archives, where I found a message showing how to login to the CVS, > > although, the knowledge base has some type of filter for email addresses? > > So again, I was not able to login. > > > > I am a novice at CVS, can someone please describe what I am suppose to do, > > or atleast what my CVSROOT var is suppose to be? Thanks. > > The commands are: > > cvs -d:pserver:anonymous@cv...:/cvsroot/dillo login > cvs -z8 -d:pserver:anonymous@cv...:/cvsroot/dillo co dillo > > There is also information at the SourceForge side, at > http://so....net/cvs/?group_id=603. Perhaps this should added to > home page, Jorge? > > Sebastian > > _______________________________________________ > Dillo-dev mailing list > Dillo-dev@li... > http://lists.so....net/lists/listinfo/dillo-dev > Re: [Dillo-dev]table algorithm (was: tables: COLSPAN) From: Jon Bradley - 2001-06-13 21:56 It is impossible to access the CVS. I have done it before, back when 0.3.0 was release. But it appears all information about the CVS is gone from the website AND the knownledge base. Have you closed down the CVS to the public? I would like to see how dillo is looking with frames. =-=-=-=-=-=-=-=-=-=-=-= Email: kreator@gc... bbs: telnet://toga.cx =-=-=-=-=-=-=-=-=-=-=-= On 13 Jun 2001, Sebastian Geerken wrote: > On Tue, Jun 12, I wrote: > > Optimal := what you would have as a result, according to the > > autolayout algorithm suggested in Appendix B of the HTML > > spec, for tables *without* COLSPAN. > > Well, sorry, I was a bit sloppy in reading the text. My code is (as > most other web browsers) based on an other definition: > > 1. All columns have have the same width W, except: > > - W is less than the minimal column width, or > - W is greater than the maximal column width. > > Furthermore, W is > > - less than all minimal widths of columns not having W as > width, and > - greater than all maximal widths of columns not having W as > width. > > 2. The table tries to use exactly the whole available width, except > if it is not possible, because the available width is > less/greater than the minimal/maximal table width. > > (BTW, there is a bug in the CVS, but I've fixed this, and will upload > it soon.) > > A simple example: The table > > > >
> Blah Blah Blah Blah Blah Blah Blah Blah Blah Blah > > Blah Blah Blah Blah Blah Blah Blah Blah Blah Blah > Blah Blah Blah Blah Blah Blah Blah Blah Blah Blah >
> > would, by the W3C algorithm, rendered as: > > +------------+----------------------+ > | Blah Blah | Blah Blah Blah Blah | > | Blah Blah | Blah Blah Blah Blah | > | Blah Blah | Blah Blah Blah Blah | > | Blah Blah | Blah Blah Blah Blah | > | Blah Blah | Blah Blah Blah Blah | > +------------+----------------------+ > > while most browsers render this: > > +-----------------+-----------------+ > | Blah Blah Blah | Blah Blah Blah | > | Blah Blah Blah | Blah Blah Blah | > | Blah Blah Blah | Blah Blah Blah | > | Blah | Blah Blah Blah | > | | Blah Blah Blah | > | | Blah Blah Blah | > | | Blah Blah | > +-----------------+-----------------+ > > Note that the W3C algorithm may have very "unaesthetical" results when > the columns differ quite much in there maximal length, while having > comparable content length (number of characters), e.g. if one column > consists of many small paragraphs, whereas another consists of few > long paragraphs. > > I think we should follow the common implementations, which authors may > have in mind, instead of the W3C recommendation. Different opinions? > > > Sebastian > > _______________________________________________ > Dillo-dev mailing list > Dillo-dev@li... > http://lists.so....net/lists/listinfo/dillo-dev > Re: [Dillo-dev]tables: COLSPAN From: Sam Dennis - 2001-06-13 19:33 On Wed, Jun 13, 2001 at 08:29:08PM +0200, Sebastian Geerken wrote: > > Currently, I'm working on COLSPAN. [...] > > Just uploaded, together with some bugfixes. > > Sebastian I was going to congratulate you on actually making that system work, but now, having actually seen its output, I've decided that it needs a bit more work. It may have been a while since I wrote any real HTML, but I think I'd be correct in thinking that the following table shouldn't be rendered as illustrated:
+-------+ +---+ | | | | +-------+ | | +---++---+ | | | || | +---+ | |+---+ | |+-------+ +---+| | +-------+ But don't take this the wrong way, I'm impressed that you managed to get that approach working at all! Re: [Dillo-dev]tables: COLSPAN From: Sebastian Geerken - 2001-06-13 18:22 > Currently, I'm working on COLSPAN. [...] Just uploaded, together with some bugfixes. Sebastian Re: [Dillo-dev]dillo CVS From: Jorge Arellano Cid - 2001-06-13 14:27 Hi! > [CVS] > The commands are: > > cvs -d:pserver:anonymous@cv...:/cvsroot/dillo login > cvs -z8 -d:pserver:anonymous@cv...:/cvsroot/dillo co dillo > > There is also information at the SourceForge side, at > http://so....net/cvs/?group_id=603. Perhaps this should added to > home page, Jorge? Done, in [Links] section. Jorge. [Dillo-dev]table algorithm (was: tables: COLSPAN) From: Sebastian Geerken - 2001-06-13 13:34 On Tue, Jun 12, I wrote: > Optimal := what you would have as a result, according to the > autolayout algorithm suggested in Appendix B of the HTML > spec, for tables *without* COLSPAN. Well, sorry, I was a bit sloppy in reading the text. My code is (as most other web browsers) based on an other definition: 1. All columns have have the same width W, except: - W is less than the minimal column width, or - W is greater than the maximal column width. Furthermore, W is - less than all minimal widths of columns not having W as width, and - greater than all maximal widths of columns not having W as width. 2. The table tries to use exactly the whole available width, except if it is not possible, because the available width is less/greater than the minimal/maximal table width. (BTW, there is a bug in the CVS, but I've fixed this, and will upload it soon.) A simple example: The table
Blah Blah Blah Blah Blah Blah Blah Blah Blah Blah Blah Blah Blah Blah Blah Blah Blah Blah Blah Blah Blah Blah Blah Blah Blah Blah Blah Blah Blah Blah
would, by the W3C algorithm, rendered as: +------------+----------------------+ | Blah Blah | Blah Blah Blah Blah | | Blah Blah | Blah Blah Blah Blah | | Blah Blah | Blah Blah Blah Blah | | Blah Blah | Blah Blah Blah Blah | | Blah Blah | Blah Blah Blah Blah | +------------+----------------------+ while most browsers render this: +-----------------+-----------------+ | Blah Blah Blah | Blah Blah Blah | | Blah Blah Blah | Blah Blah Blah | | Blah Blah Blah | Blah Blah Blah | | Blah | Blah Blah Blah | | | Blah Blah Blah | | | Blah Blah Blah | | | Blah Blah | +-----------------+-----------------+ Note that the W3C algorithm may have very "unaesthetical" results when the columns differ quite much in there maximal length, while having comparable content length (number of characters), e.g. if one column consists of many small paragraphs, whereas another consists of few long paragraphs. I think we should follow the common implementations, which authors may have in mind, instead of the W3C recommendation. Different opinions? Sebastian Re: [Dillo-dev]dillo CVS From: Sebastian Geerken - 2001-06-13 13:34 On Thu, Jan 04, Jon Bradley wrote: > > Hi, > > I am trying to login to the dillo CVS, so I searched the dillo website, > where I found no info on the CVS. I then got to the knowledge base of > Mail Archives, where I found a message showing how to login to the CVS, > although, the knowledge base has some type of filter for email addresses? > So again, I was not able to login. > > I am a novice at CVS, can someone please describe what I am suppose to do, > or atleast what my CVSROOT var is suppose to be? Thanks. The commands are: cvs -d:pserver:anonymous@cv...:/cvsroot/dillo login cvs -z8 -d:pserver:anonymous@cv...:/cvsroot/dillo co dillo There is also information at the SourceForge side, at http://so....net/cvs/?group_id=603. Perhaps this should added to home page, Jorge? Sebastian [Dillo-dev]dillo CVS From: Jon Bradley - 2001-06-13 01:38 Hi, I am trying to login to the dillo CVS, so I searched the dillo website, where I found no info on the CVS. I then got to the knowledge base of Mail Archives, where I found a message showing how to login to the CVS, although, the knowledge base has some type of filter for email addresses? So again, I was not able to login. I am a novice at CVS, can someone please describe what I am suppose to do, or atleast what my CVSROOT var is suppose to be? Thanks. =-=-=-=-=-=-=-=-=-=-=-= Email: kreator@gc... bbs: telnet://toga.cx =-=-=-=-=-=-=-=-=-=-=-= Re: [Dillo-dev]tables: COLSPAN From: Sebastian Geerken - 2001-06-12 18:49 On Tue, Jun 12, Sam Dennis wrote: > It seems to me that your algorithm may have some problems with tables using > combinations of ROWSPAN and COLSPAN, to take an extreme example: > > +-------+---+ > | A | | > +---+---+ B + > | | D | | > | C +---+---+ > | | E | > +---+-------+ > > Would your algorithm handle this well? The description you've given doesn't > suggest so. Tested ... works. :-) Table rendering is (in both implementations) done in three steps: 1. Calculating minimal and maximal column widths, 2. calculating the column widths, and 3. calculating the column heights. Only step 1 and 2 (which are changed) deal with COLSPAN, and only step 3 deals with ROWSPAN, so there is no problem. Note that when rendering documents (as opposed to GUIs), both dimensions are not symmetric, most is determined by the viewport width, the height plays a much less important role. > Also, there is perhaps the issue of what `optimal' means in table layout > anyhow. Optimal := what you would have as a result, according to the autolayout algorithm suggested in Appendix B of the HTML spec, for tables *without* COLSPAN. (BTW, this is what led me to the idea of recursive tables: if COLSPAN is a problem, try to find subtables and so eliminate COLSPAN.) E.g., the following layout is optimal: +------+-------------------------------+ | Blah | Blah Blah Blah Blah Blah Blah | | | Blah Blah Blah Blah | +------+-------------------------------+ |------------ viewport width ------------| while the following, trivial, is not: +------------------+-------------------+ | Blah | Blah Blah Blah | | | Blah Blah Blah | | | Blah Blah Blah | | | Blah | +------------------+-------------------+ This should also work for COLSPAN: +--------------------------------------+ |BlahBlahBlahBlahBlahBlahBlahBlahBlah | +------+-------------------------------+ | Blah | Blah Blah Blah Blah Blah Blah | | | Blah Blah Blah Blah | +------+-------------------------------+ instead of: +--------------------------------------+ |BlahBlahBlahBlahBlahBlahBlahBlahBlah | +------------------+-------------------+ | Blah | Blah Blah Blah | | | Blah Blah Blah | | | Blah Blah Blah | | | Blah | +------------------+-------------------+ Note that my algorithm isn't optimal in some cases, an alternative would be "real" optimization, but this will be even more complicated and slow. > Finally, your solution looks _very_ awkward to code. I don't like the idea of > a recursive algorithm for table layout in the first place and it might present > more problems than it solves, mightn't a little thought go towards a simpler > algorithm? (Not that I've got one...) I don't think that recursion is complex per se, often problems can solved much simpler that way. In this case, the code (it is nearly complete) isn't much more complex than before (and not slower), although I must admit that the idea might be a bit strange. Maybe a look at other browsers might be useful, but I'll complete it now, it can anyway changed in the future. Sebastian Re: [Dillo-dev]tables: COLSPAN From: Sam Dennis - 2001-06-12 15:03 It seems to me that your algorithm may have some problems with tables using combinations of ROWSPAN and COLSPAN, to take an extreme example: +-------+---+ | A | | +---+---+ B + | | D | | | C +---+---+ | | E | +---+-------+ Would your algorithm handle this well? The description you've given doesn't suggest so. Also, there is perhaps the issue of what `optimal' means in table layout anyhow. Finally, your solution looks _very_ awkward to code. I don't like the idea of a recursive algorithm for table layout in the first place and it might present more problems than it solves, mightn't a little thought go towards a simpler algorithm? (Not that I've got one...) Re: [Dillo-dev]tables: COLSPAN From: Sebastian Geerken - 2001-06-11 18:02 On Mon, Jun 11, Sam Dennis wrote: > Sorry if this may mean that a lot of time has been wasted in reinventing the > wheel, but there are quite extensive notes on tables in general and layout algorithms in > particular in the HTML specs. I use, more or less, the algorithm described in the spec, but I handle cells spanning several columns in a different way, since the described algorithm is sufficient for tabled data, but not for "layout tables". Not that that tables are the best way for page layout, but some pages would be unreadable otherwise. This is the part in the spec about the problem: For cells that span multiple columns, a simple approach consists of apportioning the min/max widths evenly to each of the constituent columns. A slightly more complex approach is to use the min/max widths of unspanned cells to weight how spanned widths are apportioned. Experiments suggest that a blend of the two approaches gives good results for a wide range of tables. But I'm aiming towards perfect, not just good, results. ;-) Suppose the following table:
Links:
... 		A long text over several lines blah blah blah ...
An optimal layout (what my algorithm does indeed :-) would result in: ,----------------------------------------. | Banner | `----------------------------------------' Links: A long text over several lines blah blah blah ... blah blah ... |--------------------- viewport width ---------------------| The result of the simple algorithm in the spec: ,--------------------------------------------------. | Banner | `--------------------------------------------------' Links: A long text over several lines blah ... blah blah ... |--------------------- viewport width ---------------------| The complex approach would have a similar effect, since the minimal widths of both text cells are quite similar, this would make both weights for the minimum nearly equal, and so (in this case) determine the same column widths. Sebastian Re: [Dillo-dev]tables: COLSPAN From: Sam Dennis - 2001-06-11 14:41 On Fri, Jun 08, 2001 at 02:14:47PM +0200, Sebastian Geerken wrote: > Hi, > > first, FYI: there have been some updates of the table implementation, > image maps and COLSPAN now work. For finding text, the best solution > is probably to add a new method DwWidget::find_text, which searches > recursively in DwContainer's. > > Currently, I'm working on COLSPAN. As you may have noticed, the > current implementation does not work well for large spanning > columns. The approach I'm using is to divide the table recursively > into subtables. This is not always possible, so sometimes the cells > have to be "divided", which makes the rendering not optimal, but > reasonable, the effect should be small in most cases. An optimal > solution would otherwise be complicated and slow. Sorry if this may mean that a lot of time has been wasted in reinventing the wheel, but there are quite extensive notes on tables in general and layout algorithms in particular in the HTML specs. http://www.w3.org/TR/html4/appendix/notes.html#notes-table might be worth a look. Re: [Dillo-dev]tables: COLSPAN From: Sebastian Geerken - 2001-06-08 12:43 On Fri, Jun 08, I wrote: > first, FYI: there have been some updates of the table implementation, > image maps and COLSPAN now work. ^^^^^^^ Of course, I meant ROWSPAN. Sebastian [Dillo-dev]tables: COLSPAN From: Sebastian Geerken - 2001-06-08 12:13 Hi, first, FYI: there have been some updates of the table implementation, image maps and COLSPAN now work. For finding text, the best solution is probably to add a new method DwWidget::find_text, which searches recursively in DwContainer's. Currently, I'm working on COLSPAN. As you may have noticed, the current implementation does not work well for large spanning columns. The approach I'm using is to divide the table recursively into subtables. This is not always possible, so sometimes the cells have to be "divided", which makes the rendering not optimal, but reasonable, the effect should be small in most cases. An optimal solution would otherwise be complicated and slow. I'm not sure if this is the simplest approach, I've also had other ideas, but did not succeed in developing them further to an algorithm. I've appended an extract of the new documentation, I'm now working on the implementation. If someone knows a better approach, let me know it! Sebastian ---------------------------------------------------------------------- A table is divided into subtables, until single columns are left. Cells spanning the whole width are removed before dividing further. Example: +---+-------+---+ | A | B | C | +---+-------+---+ | D | E | +---+-------+---+ | F | G | H | +---+-------+---+ ' ' ` ` ' ' ` ` +---+-------+ +---+ | A | B | | C | +---+-------+ +---+ removed --> | D | | E | +---+-------+ +---+ | F | G | | H | +---+-------+ +---+ ' ' ` ` final ' ' ` ` +---+ +-------+ | A | | B | <-. +---+ +-------+ >- removed | F | | G | <-' +---+ +-------+ final ' ' ` ` ' ' ` ` [empty] [empty] final final If table cells overlap horizontally, they are (virtually) divided (resulting in a non optimal layout): +-------+---+---+ | A | B | C | +---+---+---+---+ | D | E | +---+-----------+ ' ' ` ` ' ' ` ` +-------+ +---+---+ | A | | B | C | +---+---+ +---+---+ | D |1/3| | 2/3 E | | | E | | | +---+---+ +-------+ Calculating extremes -------------------- minimal/maximal width (sub table) = - for single column: maximum of all minimal/maximal widths - otherwise: maximum of - all minimal/maximal widths of cells spanning the whole width, and - the sum of the minimal/maximal widths of the sub tables Calculating column widths ------------------------- The calculation is based on an available width, which is, at the top, the width set by a_Dw_widget_set_width, corrected by the minimal and maximal width of the whole table. At each level, the available width is always between the minimal and the maximal width of the subtable. For single columns, the width is the available width. Otherwise, use the simple iteration [described in the old DwTable.txt] on the subtables, excluding the columns spanning the whole subtable width, to get the width of the subtables. After this, apply this recursively on all subtables, with the subtable width as available width. Re: [Dillo-dev]can not visit dillo.so....net From: Jorge Arellano Cid - 2001-06-04 19:05 Li, > [...] > Why can I not visit http://dillo.so....net now? You SHOULD be able to see it!!! (I can) If you can't visit it for an extended period, you can blame your Internet Service Provider (ISP). Just as Livio said, source forge has been under some troubles lately, but AFAIK that's over. Jorge.- Re: [Dillo-dev]can not visit dillo.so....net From: Livio Baldini Soares - 2001-06-04 14:31 Hi liqx! Qiaoxia,Li writes: > Hi, > Where can I download dillo-0.5.1,Why can I not visit http:// > dillo.so....net now? Sourceforge is sometimes down... specially nowadays... But when it's not you can get dillo-0.5.1 at: http://download.so....net/dillo/dillo-0.5.1.tar.gz best regards, -- Livio [Dillo-dev]can not visit dillo.so....net From: Qiaoxia,Li - 2001-06-04 00:46 Attachments: Message as HTML SGksDQpXaGVyZSBjYW4gSSBkb3dubG9hZCBkaWxsby0wLjUuMSxXaHkgY2FuIEkgbm90IHZpc2l0 IGh0dHA6Ly9kaWxsby5zb3VyY2Vmb3JnZS5uZXQgbm93Pw0KcmVnYXJkcywNCmxpcXgNCg0K [Dillo-dev]File selection patch (again :) From: Livio Baldini Soares - 2001-06-03 04:54 Hello! I'm sending (again) my updated patch for file selection while saving pages or links. As Holger Rapp suggested, I have removed the modal type windows, now they're like any dialog window. The patch is simple (against 02-Jun CVS), and it's at: http://www.linux.ime.usp.br/~livio/dillo/file_select_new.diff Also, I think I'm willing to implement the Downloads scheme described in [Project Notes], becasue in my opinion it's another great TODO... Has anybody started anything on this? PS: Sebastian, nice work on the tables! I was really happy when I saw sourforge.net being rendered in a "readable way! I know it's still a it buggy, but we'll work it out with time... Thanks! best regards to all, -- Livio Re: [Dillo-dev]authentification From: Jorge Arellano Cid - 2001-06-01 09:44 Hi! On Thu, 31 May 2001, Didier Bretin wrote: > Hello, > > Dillo looks good :o). Go on !!!! Thanks! > Have you planned to deal with authentification ? Sure, I have an ancient patch from Marcos Ram=EDrez that was waiting for DilloUrl... If you're willing to work on it, I can send you the patch so you can start from there. Jorge.- Re: [Dillo-dev]authentification From: Jorge Arellano Cid - 2001-06-01 04:22 Hi! On Thu, 31 May 2001, Alexey F. Bekish wrote: > How from all it nsubscribe???? ??? How to unsubscribe? Go to dillo's home page, follow the [Mailing List] link, click on the "Visit subscribers list" button, click your entry, enter your password and click "unsubscribe" button! Maybe sending an email with "unsubscribe" as subject could also do the trick. Jorge.-