- Dillo-dev - lists.mailman3.com

a couple of patches: referer and image_off
by jcid＠dillo.org Feb. 2, 2008

Feb. 2, 2008

On Fri, Feb 01, 2008 at 07:38:45PM +0000, place wrote: > > In brief, using the same URI, the three options offer good > > privacy (because no new information is sent), and if we ever > > enable using the referring page, the old comments apply. > > Whew. > > See, I ran on the equivalent of "host" for > at least a year, and then have been using the > equivalent of "path" for at least a year, so > the idea that I had somehow been doing something > wrong was alarming :) :-) Well, in that case you can provide good advice on which is better as a default. -- Cheers Jorge.-

2 2

a couple of patches: referer and image_off
by jcid＠dillo.org Feb. 1, 2008

Feb. 1, 2008

Hi there, - Added a text placeholder: "[IMG]" for img_off mode. - Added a http_referer preference. See details in dillorc2. This couple plays well together. When in img_off mode, non loaded images will show "[IMG]" (or original alt text), and the partial referer helps with getting images from sites that don't allow external sources. >From the updated dillorc2: # Set the HTTP Referer information (WARNING: affects privacy) * # We use the same URI, not the refering page (to avoid cross-site tracking). # none : No referer at all (full privacy). # host : Default value. Send same URI's hostname (no sensible info is sent). # path : same URI is sent (NO PRIVACY). http_referer=host Note: please read the "*"-marked line twice. ;) -- Cheers Jorge.-

2 4

Various minor patches
by tmseck-lists＠netcologne.de Jan. 31, 2008

Jan. 31, 2008

All, I am very glad to see that dillo-fltk has now gained some momentum. While looking at it on FreeBSD (6.2) I came across some minor problems, patches for which are attached: 1) detect_libiconv.diff src/decode.c uses iconv_open but - at least on FreeBSD - libiconv is not in the default linker path, making the build abort. Add a simple test to configure.in to check for libiconv's presence and usability and a substitution for ICONV_LIBS to src/Makefile.am. 2) src-colors.h.diff I needed to include <stdint.h> or <inttypes.h> in order to have int32_t properly defined. I decided to wrap this include in a hopefully portable and correct way by instrumenting config.h. 3) src-dns.c.diff On current FreeBSD releases, netdb.h does no longer define EAI_NODATA, it had been marked as obsoleted there and #if 0'ed out since Oct 2003. It therefore cannot be checked for on a current FreeBSD release, so test whether we are either compiling on something other that FreeBSD or if it's a 4.x or older FreeBSD release. Other BSDs might have similar issues in this regard. Otherwise dillo-fltk looks really promising -- keep up the good work! Best regards, -- Thomas-Martin Seck current dillo and FLTK port maintainer on FreeBSD

4 22

[PATCH]: improved ROWSPAN handling
by onepoint＠starurchin.org Jan. 31, 2008

Jan. 31, 2008

Here's a page demonstrating some problems Dillo has when all the cells in a table have ROWSPAN > 1 . http://starurchin.org/dillo/bugs/table_rowspan.html The cause of these issues is that Dillo has (excuse me) a rather clumsy way of calculating the row heights once it has assigned column widths and calculated the cell heights. It first calculates row heights based on the cells with ROWSPAN = 1 , and then adds extra height to rows to accommodate cells with ROWSPAN > 1 . This algorithm gives poor results when no cells have ROWSPAN = 1 . It can add unnecessary vertical space (see example on the page). Worse, if there is no vertical padding in the table as well as no cells with ROWSPAN = 1 , the table has zero height after the first pass. The second pass can't cope with this (because it would be dividing by 0) so it does nothing, so the table still has no height when the layout is finished. This messes up the redisplaying of the contents and can create "dead" text that can't be selected or clicked. This patch fixes all these problems by rewriting the algorithm that calculates row heights. Nothing else in the layout code is changed. The new code scans all the cells once, top-down, and keeps track of the minimum height we must insert between the top of the table and the bottom of the cell. Once each cell is dealt with, the height assigned to its top will not change (because we work top-down) and the height assigned to its bottom can only increase, so by the end all cells have at least as much vertical height as they need. Then another pass tours the height constraints and writes the correct values into the cumHeights vector. Advantages of this patch: * Fixes the bugs noted above. * Assigns the minimal possible vertical height to the table (given the column widths that have already been assigned). * Treats all cells the same no matter what their ROWSPAN, so there are no special cases when no cells have ROWSPAN = 1 . * Doesn't do any division, so there are no special cases when the available padding is 0 . If this change is merged I'll submit a cleanup patch that removes the now-unused apportionRowSpan() method. This was a purely internal change, so I didn't need to update any docs. (Phew!) I need to look at the handling of ROWSPAN="0" attributes. Are these even legal? Regards, Jeremy Henty

2 2

[PATCH]: escape illegal characters instead of stripping them
by onepoint＠starurchin.org Jan. 31, 2008

Jan. 31, 2008

HREF attributes often contain unescaped spaces, even though this is illegal. This is unfortunate because Dillo deletes illegal URL characters, which breaks the link. See, eg. http://www.labreporter.com/Films.html This patch makes Dillo escape all illegal URL characters instead of stripping them. Regards, Jeremy Henty

2 3

Re: patch: image input as submit input
by jcid＠dillo.org Jan. 31, 2008

Jan. 31, 2008

On Wed, Jan 30, 2008 at 07:43:26PM +0000, place wrote: > In which I display my oh-so-impressive ability to, uhhh, uncomment code... :-) > It mostly just allows one to discover that there are lots > of paypal buttons out there. Half-committed. i.e. I left the outer spam_safe if, as an incentive to implement the real image input widget! ;-) -- Cheers Jorge.-

1 0

a highlighting colors proposal (with code)
by place＠gobigwest.com Jan. 31, 2008

Jan. 31, 2008

(Maybe this one deserves to be called a patch. It seems pretty clearly to be an improvement to me, at least...) If a page has, e.g., black text on a black background, it is nice to be able to highlight it to read it. So this changes the behavior from highlighted background is inverted background and highlighted text is inverted text to highlighted background is inverted background and highlighted text is normal background. And that is a good improvement in most cases, but then bg_color=#808080 inverts to 7f7f7f. So I changed SHADING_INVERSE so that it's really more of a SHADING_CONTRAST (I left the name alone because you might have chosen to apply my recent widget coloring code, which had SHADING_INVERSE in it somewhere, and it's nice to have patches compile cleanly.) The one negative point I've seen is in something like: <html><head></head><body bgcolor=#808080 text=#808080 link=blue> words words <s>strike</s> <u>underline</u> <a href=link>link</a> words. </body></html> The highlighted text will show no distinction between the underlined text and the link. It didn't seem important enough to bother with at the moment...

2 2

patch: colspan=0
by place＠gobigwest.com Jan. 30, 2008

Jan. 30, 2008

(Since Jeremy brought up the topic of spanning...) table.cc can handle colspan=0, so this just makes html.cc accept it as valid. from the spec: "The value zero ("0") means that the cell spans all columns from the current column to the last column of the column group (COLGROUP) in which the cell is defined." <table border=2> <tr> <td>A</td><td colspan=h>colspan=h</td></tr> <tr> <td>A</td><td colspan=-2>colspan=-2</td></tr> <tr> <td>A</td><td colspan=0>colspan=0</td></tr> <tr> <td>A</td><td colspan=1>colspan=1</td></tr> <tr> <td>A</td><td colspan=2>colspan=2</td></tr> <tr> <td>B</td><td>C</td><td>D</td><td>E</td><td>F</td></tr> </table>

2 1

[PATCH]: cleanup: comment out DEBUG_MSG in dw2/dw/widget.cc
by onepoint＠starurchin.org Jan. 30, 2008

Jan. 30, 2008

This patch comments out all uses of DEBUG_MSG in widget.cc . The justification is: * Consistency: all other uses of DEBUG_MSG in the DW source are already commented out. * The arguments to DEBUG_MSG use the GTK API, so they would fail to compile if they were ever used. Regards, Jeremy Henty

2 1

Q: "Yes, enough memory!", really?
by onepoint＠starurchin.org Jan. 27, 2008

Jan. 27, 2008

I'm worried about dillo2/src/url.c, line 361: p = str1 = dNew(char, strlen(url_str)); /* Yes, enough memory! */ The code after this line copies url_str to str1 , skipping illegal characters. It seems to me it should allocate one more character for the terminating null character. Does the comment "enough memory" mean that I am missing something, or is it talking about some other issue? Regards, Jeremy Henty

2 2

Work on TABLES
by jcid＠dillo.org Jan. 19, 2008

Jan. 19, 2008

On Sat, Jan 19, 2008 at 07:19:13PM +0000, place wrote: > table.cc is intimidating stuff. > > It works so far, so I don't have much to report, "It works so far" is great news for me, and certainly if it keeps stable better. Good feedback is welcomed! > but I just needed to say > that when I saw per.bak.html, I thought it was going to be something derived > from some article on http://en.wikipedia.org/wiki/Per_Bak instead of being > about, you know, percentages. Percentages is definitely less cool than frozen insulators... -- Cheers Jorge.-

1 0

Re: patch: image.cc typo one-liner
by jcid＠dillo.org Jan. 19, 2008

Jan. 19, 2008

Hi, On Sat, Jan 19, 2008 at 02:47:43AM +0000, place wrote: > It caused stripes along the sides of images with vspace and hspace. > > The line numbers of the patch are off because my image.cc is somewhat > hacked up, but who cares for a one-liner... > > diff -pur dw2/dw/image.cc dw2-cur/dw/image.cc > --- dw2/dw/image.cc 2007-12-24 06:17:14.000000000 +0000 > +++ dw2-cur/dw/image.cc 2008-01-19 02:18:15.000000000 +0000 > @@ -263,7 +263,7 @@ void Image::drawRow (int row) > > buffer->getRowArea (row, &area); > if (area.width && area.height) > - queueDrawArea (area.x + getStyle()->boxOffsetY (), > + queueDrawArea (area.x + getStyle()->boxOffsetX (), > area.y + getStyle()->boxOffsetY (), > area.width, area.height); > } Committed. Now that you're images, would you mind giving a look to these anomalies? http://www.dillo.org/test/img/ 1.- n1 renders a right column with old background (over dialup). 2.- n2 makes space inside the link, not outside, and renders different image size the first time (e.g. when from local harddisk); after back and forward size changes. 3.- n3 is obviously wrong! ;) -- Cheers Jorge.-

2 2

Work on TABLES
by jcid＠dillo.org Jan. 19, 2008

Jan. 19, 2008

Hi there, Good news! It looks like awakening the dragon wasn't a bad idea after all! It took a lot of time and debugging, but now after lots of tests, it seems much better than what we had before. This is not a simple patch, but an algorithms change + bug fixes. Beware that the html documentation is not up-to-date yet though; I prefer to let you test the new code with lots of other cases and to send feedback or bugfixes to the list before adopting the new code. I uploaded a subset of the used test cases to: http://www.dillo.org/test/tables You can start with per1.html, per2.html per3.html & per4.html. Please note that the test is not just rendering the page, but resizing its tables from MIN to MAX. It's interesting to compare what we had before, dillo1 and Firefox (Ill cases follow Firefox's example). So the bottom line is: Test, test, test!!! & send feedback. I'll be back near Jan 30; after enjoying my deserved rest period. ;-) -- Cheers Jorge.- PS: If you want to debug, uncomment table.cc:21 and define dillo_dbg_rendering to 1 in html.cc:64. PS2: I left an exit assert testing d_a and d_w >= 0. If this triggers, please send the URL.

2 1

Working on TABLES
by jcid＠dillo.org Jan. 16, 2008

Jan. 16, 2008

Hi there, As you may remember, I dared to awake the sleeping dragon of table apportion algorithms. Frankly, I don't yet know whether it was a good idea after all, but I have hope. It is impresive how complex the geometric combination of cases grows; hopefully some code will be uploaded before Jan 20. In the meanwhile, does anyone want to give a try to imagemaps? Or maybe to implement tabs? This is much simpler than it looks. Basically we can add the tab widgets and to switch the navigation stack and rendered page upon click. As dillo is quick to render this has the advantage of low memory consumption because we don't need to have all the rendered pages in memory! And it opens the possibility of multiple windows with TABS using low resources. -- Cheers Jorge.-

1 0

patch: progressive images
by place＠gobigwest.com Jan. 16, 2008

Jan. 16, 2008

(the ones that start out blurry and grow sharper) - Added code for progressive display of, well, progressive jpegs because I was tired of waiting for full download before display. jpeglib uses the word "scan" for each loop through the scanlines of the image, so I did the same. I'm not sure it's a terribly descriptive term, though, at least to me (not a graphics person). Couldn't think of a better word. - Turned out scaled ones didn't work. (And, incidentally, that was frustrating because it took some time to realize that the breaking ones were scaled ever so slightly.) Added some bit clearing to imgbuf code on the dw side so that scaled buffers would continue to copy from root buffers. - Turned out that if there was more than one copy on a page, only the first copy worked. (dillo uses image decoders on the first one and pulls the scanlines out of Dicache for the rest.) More bit clearing and a scan number to help a_Dicache_callback with its copying. - Discovered that interlaced pngs use progressive display, and that they hadn't been working for scaled or for multiple on a page. - Considered looking into animated gifs, but then my good sense prevailed :)

3 3

patch: the http header Content-Type half of charset conversion
by place＠gobigwest.com Jan. 15, 2008

Jan. 15, 2008

Dug out my old patch and ripped out the meta tag stuff, the charset-setting dialog, etc. This much should be unobjectionable. Had good fun searching for things like dillo ??????? and seeing how many sites using windows-1251 or koi8-r worked now with this much of the code. Maybe half. Greek and Hebrew were somewhat less than half. And my fonts don't have enough coverage to make it fun trying Asian sites. Of course it doesn't matter much when everybody who participates here is a native speaker of some latin-1 language... Note that View Source will not show files as converted because the conversion is up in html.cc and View Source uses the cache.

2 3

Report on fltk-2.0.x-r5990
by onepoint＠starurchin.org Jan. 13, 2008

Jan. 13, 2008

I have been running Dillo on top of the latest daily snapshot (fltk-2.0.x-r5990) and seen no problems. Has anyone else tried this? Maybe we should update the recommended FLTK snapshot? NB: fltk-2.0.x-r5990 is not binary compatible with the current recommended snapshot, so if you upgrade you have to rebuild Dillo (and any other FLTK2 applications) or you will get memory corruption. FYI my system is Linux 2.6.23.12-0 , glibc 2.3.4 , gcc 3.4.3 . Jeremy Henty

3 2

patch: multiple-line http headers (and a free bonus Content-Type msg change)
by place＠gobigwest.com Jan. 12, 2008

Jan. 12, 2008

Finally got around to it. Since it's another one of those things I am probably never going to encounter in the outside world, I tested by sticking whitespace into headers sent by file.c. PS I see that this guy has a rant about how the headers can attain an arbitrary degree of horribleness if you really want to go by the letter of the rfc: http://www.and.org/texts/server-http

2 2

Re: patch: FltkViewBase::getClippingView source-copying error
by jcid＠dillo.org Jan. 12, 2008

Jan. 12, 2008

On Tue, Jan 08, 2008 at 04:32:50PM +0000, place wrote: > I finally found out why alt tags had been so very flaky for me all along. Good, it was also biting me from a long time! Committed. -- Cheers Jorge.-

1 0

Feature request
by Beartooth＠swva.net Jan. 8, 2008

Jan. 8, 2008

With other browsers, if I copy a URL that begins with www (i.e., no http://), I will get to the site; with some, I don't even need the www -- I can type simply "loc.gov" for instance, to be taken to the Library of Congress. This is a great help, not only with sites whose address one knows by heart, but also with email, in which people omit the front part(s) of sites routinely. It would be an especially welcome feature for those of us who keep Dillo set as our default browser. I have no idea how those other browsers do it, nor why Dillo does not. Is there a reason? -- Beartooth Staffwright, PhD, Neo-Redneck Linux Convert Fedora 8; Alpine 0.99999, Pan 0.132; Privoxy 3.0.6; Dillo 0.8.6, Galeon 2.0.3, Epiphany 2.20, Opera 9.24, Firefox 2.0 Remember I know precious little of what I am talking about.

4 5

patch: circles and discs
by place＠gobigwest.com Jan. 8, 2008

Jan. 8, 2008

I almost regret this because I liked the little squares best.

2 1

patch: http warnings
by place＠gobigwest.com Jan. 8, 2008

Jan. 8, 2008

HTTP/1.1 has warning headers to tell you things like "110 Response is stale" or "214 Transformation applied". There should probably be some mechanism to actually notify the user properly -- nobody's ever going to notice an MSG_HTTP string. As was the case with the "100 Continue"s some time ago, I am not encountering genuine ones to test against. I just tested by letting it treat date headers as warnings...

2 1

select crash
by place＠gobigwest.com Jan. 8, 2008

Jan. 8, 2008

dillo crashed on a form where an option didn't have a value (clicked on the Go button at the bottom of http://www.syllable-hardware.info/hardware.php?list.1 ) A glance at the spec suggested that what I wanted in that case was the contents of the tag. So... Trying the change below at least makes it not crash, but I don't know the code, so I'm not calling this a patch. diff -pur dillo2/src/html.cc dillo2-cur/src/html.cc --- dillo2/src/html.cc 2008-01-02 20:20:31.000000000 +0000 +++ dillo2-cur/src/html.cc 2008-01-08 01:30:08.000000000 +0000 @@ -3808,7 +3808,8 @@ static void Html_submit_form2(DilloHtml if (sel_res->isSelected(i)) { DilloHtmlOption *option = input->select->options->get (i); - Html_append_input(DataStr, input->name, option->value); + char *val = option->value ? option->value : option->content; + Html_append_input(DataStr, input->name, val); } } break;

3 2

Re: patch: crash when .dillo/ is absent (cookierc)
by jcid＠dillo.org Jan. 8, 2008

Jan. 8, 2008

On Tue, Jan 08, 2008 at 06:26:07AM +0000, place wrote: > Poking around on mobileread.com and I saw where the guy who put dillo > on the iliad mentioned that the fltk dillo segfaults when there's no > .dillo dir. Well, we can't have _that_ happening. > > Plus, if somebody's on a crippled system, they might desperately > need info from the web to get themselves out of whatever they just > accidentally did to their computer... > Committed! -- Cheers Jorge.-

1 0

patch: color shading
by place＠gobigwest.com Jan. 3, 2008

Jan. 3, 2008

This should make shading better...

2 4

[patch] segfault when closing window while it is loading
by Johannes.Hofmann＠gmx.de Jan. 2, 2008

Jan. 2, 2008

Happy new year! Below is a small patch to fix a crash that happens when I open a new window (e.g. with the middle mouse button) and close it immediately while it is still loading the page. Johannes --- a/src/uicmd.cc Mon Dec 31 22:49:02 2007 +0100 +++ b/src/uicmd.cc Tue Jan 01 12:38:14 2008 +0100 @@ -125,6 +125,7 @@ void a_UIcmd_close_bw(void *vbw) Layout *layout = (Layout*)bw->render_layout; MSG("a_UIcmd_close_bw\n"); + a_Bw_stop_clients(bw, BW_Root + BW_Img + Bw_Force); ui->destroy(); delete(layout); a_Bw_free(bw);

3 4

Re: Colspan and table percentages
by place＠gobigwest.com Jan. 2, 2008

Jan. 2, 2008

I've noticed that sometimes people will make a table with two columns for the purpose of indenting their text, specifying only the width of the first column as some percentage. When they do this, dillo squishes everything to the right. Tiny example: <html><head><title></title></head><body> <table width=500 border=3> <tr> <td width="50%">Is explicitly 50%</td> <td>should be 50%</td> </tr> </table> </body></html>

2 1

hook up the limit_text_width pref
by place＠gobigwest.com Jan. 2, 2008

Jan. 2, 2008

I'm not sure under what circumstances this pref. actually changes anything, but here it is anyway. (I didn't notice a difference when I enabled it in my old dillo either...)

2 1

patch: start_page and fullwindow_start prefs
by place＠gobigwest.com Jan. 2, 2008

Jan. 2, 2008

2 1

patch: the position part of the geometry pref.
by place＠gobigwest.com Jan. 2, 2008

Jan. 2, 2008

(That is, the X and Y of "geometry=WxH+X+Y") It appears that fltk just _guesses_ what the border size might be rather than actually determining the value. So when I run this with no window mgr, it puts the window where it would be if there were a 4-pixel-wide border and 21-pixel-high title bar. Same with twm, which is suboptimal because twm doesn't use those dimensions. As for fvwm2, which is the one I actually use -- and does appear to have those dimensions -- somehow the window ends up eight pixels too far down. But I'm regarding this as Not My Problem.

2 1

[PATCH] image load graphics improvements
by onepoint＠starurchin.org Jan. 2, 2008

Jan. 2, 2008

This patch tweaks the image load graphics: the spacing is nicer and the 'G' in "IMG" is more legible. You can preview the changes at http://starurchin.org/dillo/image_load/ . Regards, Jeremy Henty

2 1

[PATCH] initialise Layout::scrollIdleNotInterrupted
by onepoint＠starurchin.org Jan. 2, 2008

Jan. 2, 2008

This is another valgrind pacifier; it squashes an error "Conditional jump or move depends on uninitialised value(s)" at "dw::core::Layout::updateAnchor() (layout.cc:474)" from "valgrind dw2-cur/test/dw-lists". Regards, Jeremy Henty

2 1

Re: "de-exporting" prefs stuff
by jcid＠dillo.org Jan. 2, 2008

Jan. 2, 2008

On Tue, Jan 01, 2008 at 01:25:01AM +0000, place wrote: > Inspired by the 'de-exporting' of all of the gunk that had been in html.hh > a couple of months ago. (plus a few new defines and a 'static const') Committed. -- Cheers Jorge.-

1 0

[patch] fix hanging stop button
by Johannes.Hofmann＠gmx.de Jan. 2, 2008

Jan. 2, 2008

Hello, when pressing the stop button, it get's deactivated properly, but seems to get stuck in the pressed position. The patch below fixes the problem here. For the back- and forward- buttons it's not really necessary, because the url change seems to cause the necessary redraw anyway. However I added it there too for consistency. Regards, Johannes --- a/src/ui.cc Fri Dec 28 18:33:50 2007 +0100 +++ b/src/ui.cc Sat Dec 29 10:36:50 2007 +0100 @@ -21,6 +21,7 @@ #include <fltk/Input.h> #include <fltk/Output.h> #include <fltk/run.h> +#include <fltk/damage.h> #include <fltk/BarGroup.h> #include <fltk/PackedGroup.h> #include <fltk/xpmImage.h> @@ -996,12 +997,15 @@ void UI::button_set_sens(UIButton btn, i switch (btn) { case UI_BACK: (sens) ? Back->activate() : Back->deactivate(); + Back->redraw(DAMAGE_HIGHLIGHT); break; case UI_FORW: (sens) ? Forw->activate() : Forw->deactivate(); + Forw->redraw(DAMAGE_HIGHLIGHT); break; case UI_STOP: (sens) ? Stop->activate() : Stop->deactivate(); + Stop->redraw(DAMAGE_HIGHLIGHT); break; default: break;

2 1

[patch] make TopGroup a PackedGroup
by Johannes.Hofmann＠gmx.de Jan. 2, 2008

Jan. 2, 2008

Hello, as I didn't get any feedback from Sebastian yet on my proposal to make the scrollbars part of UI instead of FltkViewport, I decided to split my initial patch. Below is a version that still uses a PackedGroup in UI, but leaves the scrollbars in FltkViewport. Here comes the details: * Make scrollbars really children of FltkViewport, as it was probabely intended. * Use widget local coordinates in FltkViewport::adjustScrollbarsAndGadgets. * Adjust draw() and handle() so that scrollbars work properly. * As a consequence replace() now works properly, so remove the workaround. * Make TopGroup a PackedGroup and let it do the layout. * Simplify fullscreen mode. This also fixes annoying problems with the statusbar when new windows are created in fullscreen mode. Cheers, Johannes

3 8

Best wishes for this year 2008
by jcid＠dillo.org Jan. 1, 2008

Jan. 1, 2008

Hi guys, I want to start the year expressing my gratitude to all of you who have contributed your time and effort to make the Dillo project continue forward. The working atmosphere is very nice and that's a very important asset! Our ability to keep working steadily and to attract more developers is probably the cornerstone that can make Dillo --as a project-- thrive and evolve. The manpower behind a project is what pushes it forward. Once again, thanks to you all, and receive my best wishes. -- Cheers Jorge.-

2 1

[PATCH] SELECT form controls
by onepoint＠starurchin.org Dec. 31, 2007

Dec. 31, 2007

This patch adds (partial) support for SELECT controls in forms. Please note: * SELECTED and DISABLED are supported. * If no item is selected in the HTML source then the first item is selected by default. This means the code must delay adding any items until the SELECT is closed, since the DW API forces us to choose whether to select an item when creating it and won't let us change that later. * MULTI and SIZE are not supported: we render all SELECT controls as option menus that allow only one choice. I believe more work needs to be done on the DW ListResource before we can support multiple selections. * I added IN_OPTION to the HTML parsing flags. The code is also more careful than before to set/unset IN_FORM, IN_TEXTAREA, IN_SELECT and IN_OPTION. This may be a little redundant but I wanted to be sure that HTML tag soup can't make the flags inconsistent. I doubt the overhead of the extra bit-flipping is significant. Regards, Jeremy Henty

4 6

Re: patch: here are some consts for pixmaps.h
by jcid＠dillo.org Dec. 31, 2007

Dec. 31, 2007

Committed. -- Cheers Jorge.-

1 0

Re: patch: removing extraneous header inclusion
by jcid＠dillo.org Dec. 31, 2007

Dec. 31, 2007

On Sat, Dec 29, 2007 at 12:54:05AM +0000, place wrote: > Removed some where I couldn't see that they were being used for anything in > that file. > In one case, I added a prefs.h include (I had deleted a msg.h include, > and prefs.h is included in msg.h). > Committed. -- Cheers Jorge.-

1 0

Re: patch: here are a few more statics for dpi
by jcid＠dillo.org Dec. 31, 2007

Dec. 31, 2007

Committed. -- Cheers Jorge.-

1 0

patch: just hooking up panel preferences (panel_size, small_icons)
by place＠gobigwest.com Dec. 29, 2007

Dec. 29, 2007

Interestingly, the default was and is Tiny if panel_size is not mentioned in the rc, the default was and is Small in dillorc2, the initial value had been Medium in ui.cc, and the default was and is Large if the panel_size value is not understood.

3 2

[PATCH] dw::fltk::ui::FltkOptionMenuResource::isSelected()
by onepoint＠starurchin.org Dec. 28, 2007

Dec. 28, 2007

This patch provides a correct implementation of dw::fltk::ui::FltkOptionMenuResource::isSelected(), replacing the current stub. It is necessary for dillo2 to support SELECT controls in forms. Notes: * We implement Johannes' suggestion of modifying createNewMenu, using a single callback for each menu and calling item() to get the selected item. This gives a very clean implementation that doesn't add any new data members or virtual methods to the base classes. * We use FLTK widgets' user_data() to attach dw objects to FLTK widgets so that the FLTK callback can retrieve those dw objects from its arguments. This works very nicely and saves code since we don't have to add new data members to existing classes. * To avoid making identical changes everywhere an Item is created we refactored and created a new Item::createNewWidget() method. * For consistency of style we also refactored and created Item::createNewGroupWidget() and FltkSelectionResource::createNewItem(). (Despite their simplicity these refactorings actually account for most of the patch.) I should finish polishing the dillo2 SELECT patch soon. Then I'll be able to set comment thresholds on Slashdot! (These days that's the thing that most often forces me to quit dillo2 and start another browser, so this is a big itch to scratch.) Regards, Jeremy Henty

2 1

Q: Form SELECT support: what to do in the DW code?
by onepoint＠starurchin.org Dec. 26, 2007

Dec. 26, 2007

I'm working on supporting SELECT controls in forms. I've got a working first draft of the changes to the Dillo parser, but it's of no use yet because the Dillo menu widgets are incomplete; the menus report every item as selected and the items don't yet do anything when you select them (unless I've missed something). So I'm wondering how to fix this. We need a way to connect Items to callbacks. We need some abstraction because ListResources and OptionMenuResources will do things differently. The most straightforward thing to do is add a virtual function to SelectionResource that returns the callback to attach to the FLTK items we create. Is that OK, or do we need something more abstract? Assuming the Item will be passed to the callback, we really need a way to get back from the Item to the ListResource. The easiest way is to simply add a new data member to Itens that points to the ListResource. Again, is that OK, or should we be more abstract and not hardwire that implentation into the base class? I'm a fan of abstraction (I used to be a mathematician) but I'm inclined not to go too abstract unless there's good reason, so I think the suggestions above are good enough. Thoughts? If we do the above then fixing OptionMenuResource should be easy; add a data member to record the last selected Item, connect Items to a callback that updates that data member and implement isSelected() to compare the item being queried with the selected item. Is it worth submitting a patch along these lines, or should I do it differently? Regards, Jeremy Henty

2 2

[PATCH] FLTK Resources should call ::fltk::setfont before ::fltk::getwidth
by onepoint＠starurchin.org Dec. 23, 2007

Dec. 23, 2007

If you point Dillo at <URL:http://starurchin.org/dillo/bugs/form_setfont.html> you'll see that the textareas are rendered with different widths despite having the same number of columns. The problem is that FltkResource objects don't call ::fltk::setfont before calling ::fltk::getwidth to calculate their width, so they use the font that was set by the previous header elements, even though those elements were properly closed in the HTML source. This patch adds the missing ::fltk::setfont calls to Fltk{Entry,Button,Multiline}Resource . I've looked at all other uses of ::fltk::getwidth. The ones in dw2-cur/dw/fltkplatform.cc are properly protected by calls to ::fltk::setfont , but the ones in dw2-cur/dw/fltkpreview.cc and dillo2-cur/src/dialog.cc are not. I can't tell if they need to be protected, perhaps the current FLTK font is always correct when they are called? Thoughts? Regards, Jeremy Henty

2 5

[PATCH] FltkViewport fixes and TextEditor for FltkMultiLineResource
by onepoint＠starurchin.org Dec. 22, 2007

Dec. 22, 2007

Finally (I hope)! This patch changes FltkMultiLineResource to use the FLTK TextEditor. It includes fixes for the double free problem in FltkMultiLineResource that Johannes found, Johannes' fixes for the Viewport drawing code and the patch for the Viewport scrolling bug (which was previously posted separately so please ignore that post). Regards, Jeremy Henty diff -pru -- dw2-ref/dw/fltkui.cc dw2-cur/dw/fltkui.cc --- dw2-ref/dw/fltkui.cc 2007-12-19 12:26:59.000000000 +0000 +++ dw2-cur/dw/fltkui.cc 2007-12-22 12:48:23.000000000 +0000 @@ -30,7 +30,7 @@ #include <fltk/Group.h> #include <fltk/Input.h> #include <fltk/SecretInput.h> -#include <fltk/MultiLineInput.h> +#include <fltk/TextEditor.h> #include <fltk/RadioButton.h> #include <fltk/CheckButton.h> #include <fltk/Choice.h> @@ -533,7 +533,7 @@ FltkMultiLineTextResource::FltkMultiLine FltkSpecificResource <dw::core::ui::MultiLineTextResource> (platform), numCols(cols), numRows(rows) { - initText = NULL; + buffer = new ::fltk::TextBuffer; editable = false; init (platform); @@ -541,27 +541,23 @@ FltkMultiLineTextResource::FltkMultiLine FltkMultiLineTextResource::~FltkMultiLineTextResource () { - if (initText) - delete initText; + for (Iterator <ViewAndWidget> it = viewsAndWidgets->iterator (); + it.hasNext(); ) { + ViewAndWidget *viewAndWidget = it.getNext (); + ((::fltk::TextEditor *) viewAndWidget->widget)->buffer (0); + } + delete buffer; } ::fltk::Widget *FltkMultiLineTextResource::createNewWidget (core::Allocation *allocation) { - ::fltk::MultiLineInput *input = - new ::fltk::MultiLineInput (allocation->x, allocation->y, - allocation->width, - allocation->ascent + allocation->descent); - - if (viewsAndWidgets->isEmpty ()) { - // First widget created, attach the set text. - if (initText) - input->value (initText); - } else - input->value - (((::fltk::MultiLineInput*)viewsAndWidgets->getFirst()->widget)->value ()); - - return input; + ::fltk::TextEditor *text = + new ::fltk::TextEditor (allocation->x, allocation->y, + allocation->width, + allocation->ascent + allocation->descent); + text->buffer (buffer); + return text; } void FltkMultiLineTextResource::sizeRequest (core::Requisition *requisition) @@ -586,23 +582,12 @@ void FltkMultiLineTextResource::sizeRequ const char *FltkMultiLineTextResource::getText () { - if (viewsAndWidgets->isEmpty ()) - return initText; - else - return ((::fltk::MultiLineInput*)viewsAndWidgets->getFirst()->widget)->value (); + return buffer->text (); } void FltkMultiLineTextResource::setText (const char *text) { - if (initText) - delete initText; - initText = strdup (text); - - for (Iterator <ViewAndWidget> it = viewsAndWidgets->iterator (); - it.hasNext(); ) { - ViewAndWidget *viewAndWidget = it.getNext (); - ((::fltk::MultiLineInput*)viewAndWidget->widget)->value (initText); - } + buffer->text (text); } bool FltkMultiLineTextResource::isEditable () diff -pru -- dw2-ref/dw/fltkui.hh dw2-cur/dw/fltkui.hh --- dw2-ref/dw/fltkui.hh 2007-12-19 12:26:59.000000000 +0000 +++ dw2-cur/dw/fltkui.hh 2007-12-22 12:48:23.000000000 +0000 @@ -7,6 +7,7 @@ #include <fltk/Button.h> #include <fltk/Menu.h> +#include <fltk/TextBuffer.h> namespace dw { namespace fltk { @@ -317,7 +318,7 @@ class FltkMultiLineTextResource: public FltkSpecificResource <dw::core::ui::MultiLineTextResource> { private: - const char *initText; + ::fltk::TextBuffer *buffer; bool editable; int numCols, numRows; diff -pru -- dw2-ref/dw/fltkviewbase.cc dw2-cur/dw/fltkviewbase.cc --- dw2-ref/dw/fltkviewbase.cc 2007-12-06 17:29:59.000000000 +0000 +++ dw2-cur/dw/fltkviewbase.cc 2007-12-22 12:48:23.000000000 +0000 @@ -53,9 +53,7 @@ void FltkViewBase::draw () { int d = damage(); - if (d == DAMAGE_CHILD) { - Group::draw (); - } else if (d == DAMAGE_VALUE) { + if (d == DAMAGE_VALUE) { container::typed::Iterator <core::Rectangle> it; ::fltk::Rectangle view (0, 0, w (), h ()); @@ -76,7 +74,16 @@ void FltkViewBase::draw () pop_clip (); } } - } else { + + d &= ~DAMAGE_VALUE; + } + + if (d & DAMAGE_CHILD) { + Group::draw (); + d &= ~DAMAGE_CHILD; + } + + if (d) { ::fltk::Rectangle rect (x(), y(), w(), h()); /* fltk-clipping does not use widget coordinates */ @@ -98,6 +105,7 @@ void FltkViewBase::draw () drawRegion.clear (); } + core::ButtonState getDwButtonState () { int s1 = event_state (); @@ -113,7 +121,6 @@ core::ButtonState getDwButtonState () return (core::ButtonState)s2; } - int FltkViewBase::handle (int event) { bool processed; @@ -349,6 +356,23 @@ FltkWidgetView::~FltkWidgetView () delete canvasWidgets; } +void FltkWidgetView::layout () { + /** + * pass layout to child widgets. This is needed for complex fltk + * widgets as TextEditor. + * We can't use Group::layout() as that would rearrange the widgets. + */ + for(Iterator <TypedPointer < ::fltk::Widget> > it + = canvasWidgets->iterator (); + it.hasNext (); ) { + ::fltk::Widget *widget = it.getNext()->getTypedValue (); + + if (widget->layout_damage ()) { + widget->layout (); + } + } +} + void FltkWidgetView::drawText (core::style::Font *font, core::style::Color *color, core::style::Color::Shading shading, diff -pru -- dw2-ref/dw/fltkviewbase.hh dw2-cur/dw/fltkviewbase.hh --- dw2-ref/dw/fltkviewbase.hh 2007-12-06 17:29:59.000000000 +0000 +++ dw2-cur/dw/fltkviewbase.hh 2007-12-22 12:48:23.000000000 +0000 @@ -74,6 +74,8 @@ public: FltkWidgetView (int x, int y, int w, int h, const char *label = 0); ~FltkWidgetView (); + void layout(); + void drawText (core::style::Font *font, core::style::Color *color, core::style::Color::Shading shading, diff -pru -- dw2-ref/dw/fltkviewport.cc dw2-cur/dw/fltkviewport.cc --- dw2-ref/dw/fltkviewport.cc 2007-11-18 07:36:53.000000000 +0000 +++ dw2-cur/dw/fltkviewport.cc 2007-12-22 12:48:23.000000000 +0000 @@ -46,7 +46,7 @@ FltkViewport::FltkViewport (int x, int y vscrollbar->set_vertical(); vscrollbar->callback (vscrollbarCallback, this); - scrollX = scrollY = 0; + scrollX = scrollY = scrollDX = scrollDY = 0; dragScrolling = 0; gadgetOrientation[0] = GADGET_HORIZONTAL; @@ -154,6 +154,8 @@ void FltkViewport::layout () { theLayout->viewportSizeChanged (this, w(), h()); adjustScrollbarsAndGadgetsAllocation (); + + FltkWidgetView::layout (); } void FltkViewport::draw_area (void *data, const Rectangle& cr )

2 1

[PATCH] initialise FltkViewport::{scrollDX,scrollDY}
by onepoint＠starurchin.org Dec. 22, 2007

Dec. 22, 2007

This patch initialises FltkViewport::{scrollDX,scrollDY}. This removes a valgrind warning about a conditional jump depending on uninitialised data. It also fixes the mysterious sporadic disappearing page bug that was afflicting the development of the TextEditor patch. This makes me very happy, and confirms yet again that valgrind is awesome! Regards, Jeremy Henty diff -pru -- dw2-ref/dw/fltkviewport.cc dw2-cur/dw/fltkviewport.cc --- dw2-ref/dw/fltkviewport.cc 2007-12-22 01:23:39.000000000 +0000 +++ dw2-cur/dw/fltkviewport.cc 2007-12-22 01:46:11.000000000 +0000 @@ -46,7 +46,7 @@ FltkViewport::FltkViewport (int x, int y vscrollbar->set_vertical(); vscrollbar->callback (vscrollbarCallback, this); - scrollX = scrollY = 0; + scrollX = scrollY = scrollDX = scrollDY = 0; dragScrolling = 0; gadgetOrientation[0] = GADGET_HORIZONTAL;

2 2

[PATCH] TEXTAREA controls submit their contents
by onepoint＠starurchin.org Dec. 22, 2007

Dec. 22, 2007

This patch makes TEXTAREA controls submit their contents when their form is activated. Note that I reworked the handling of the embed and widget members of the form structure for TEXTAREA controls. This was because I was getting crashes and corruption when the form attempted to submit itself. I think the various coercions between (void *), (Widget *) and (Embed *) were going wrong. Changing things so that widget == embed is also more consistent with the way the other embedded form controls are handled. Regards, Jeremy Henty diff -pru -- dillo2-ref/src/html.cc dillo2-cur/src/html.cc --- dillo2-ref/src/html.cc 2007-12-16 22:40:25.000000000 +0000 +++ dillo2-cur/src/html.cc 2007-12-17 01:59:29.000000000 +0000 @@ -3792,6 +3792,11 @@ static void Html_submit_form2(DilloHtml entryres = (EntryResource*)((Embed*)input->widget)->getResource(); Html_append_input(DataStr, input->name, entryres->getText()); break; + case DILLO_HTML_INPUT_TEXTAREA: + MultiLineTextResource *textres; + textres = (MultiLineTextResource*)((Embed*)input->widget)->getResource(); + Html_append_input(DataStr, input->name, textres->getText()); + break; case DILLO_HTML_INPUT_CHECKBOX: case DILLO_HTML_INPUT_RADIO: ToggleButtonResource *cb_r; @@ -4240,7 +4245,7 @@ static void Html_tag_close_textarea(Dill form = html->forms->getRef (html->forms->size() - 1); form->inputs->get(form->inputs->size() - 1).init_str = str; widget = (Widget*)(form->inputs->get(form->inputs->size() - 1).widget); - ((MultiLineTextResource *)widget)->setText(str); + ((MultiLineTextResource *)((Embed *)widget)->getResource ())->setText(str); html->InFlags &= ~IN_TEXTAREA; } @@ -4292,12 +4297,10 @@ static void Html_tag_open_textarea(Dillo Widget *widget; Embed *embed; - widget = (Widget*)textres; - embed = new Embed(textres); - EntryResource *entryres = (EntryResource*)embed->getResource(); + widget = embed = new Embed(textres); /* Readonly or not? */ if (Html_get_attr(html, tag, tagsize, "readonly")) - entryres->setEditable(false); + textres->setEditable(false); Html_add_input(form, DILLO_HTML_INPUT_TEXTAREA, widget, embed, name, NULL, NULL, false);

3 20

Fixed a bug in the textblock's wrapping algorithm.
by jcid＠dillo.org Dec. 20, 2007

Dec. 20, 2007

A small bug fix in the textblock's wrapping algorithm was committed. It fixes some border cases like the attached page. -- Cheers Jorge.-

1 0

[patch] resize progress indicators dynamically
by Johannes.Hofmann＠gmx.de Dec. 19, 2007

Dec. 19, 2007

Hello, this patch resizes the progress indicators as needed. I don't like the constant "+ 4". Any ideas? Regards, Johannes diff -r 0a8a112b8e48 src/ui.cc --- a/src/ui.cc Mon Dec 17 14:07:38 2007 +0100 +++ b/src/ui.cc Mon Dec 17 19:33:53 2007 +0100 @@ -409,18 +409,24 @@ PackedGroup *UI::make_location() */ PackedGroup *UI::make_progress_bars(int wide, int thin_up) { + int width, height; + ProgBox = new PackedGroup(0,0,0,0); ProgBox->begin(); // Images - IProg = new InvisibleBox(0,0,pr_w,0, + IProg = new InvisibleBox(0,0,0,0, wide ? "Images\n0 of 0" : "0 of 0"); IProg->box(thin_up ? THIN_UP_BOX : EMBOSSED_BOX); IProg->labelcolor(GRAY10); + IProg->measure_label(width, height); + IProg->resize(width + 4, height + 4); // Page - PProg = new InvisibleBox(0,0,pr_w,0, + PProg = new InvisibleBox(0,0,0,0, wide ? "Page\n0.0KB" : "0.0KB"); PProg->box(thin_up ? THIN_UP_BOX : EMBOSSED_BOX); PProg->labelcolor(GRAY10); + PProg->measure_label(width, height); + PProg->resize(width + 4, height + 4); ProgBox->type(PackedGroup::ALL_CHILDREN_VERTICAL); ProgBox->end(); @@ -471,24 +477,24 @@ Group *UI::make_panel(int ww) if (PanelSize == P_tiny) { if (Small_Icons) - xpos = 0, bw = 22, bh = 22, fh = 0, lh = 22, lbl = 0, pr_w = 45; + xpos = 0, bw = 22, bh = 22, fh = 0, lh = 22, lbl = 0; else - xpos = 0, bw = 28, bh = 28, fh = 0, lh = 28, lbl = 0, pr_w = 45; + xpos = 0, bw = 28, bh = 28, fh = 0, lh = 28, lbl = 0; } else if (PanelSize == P_small) { if (Small_Icons) - xpos = 0, bw = 20, bh = 20, fh = 0, lh = 20, lbl = 0, pr_w = 45; + xpos = 0, bw = 20, bh = 20, fh = 0, lh = 20, lbl = 0; else - xpos = 0, bw = 28, bh = 28, fh = 0, lh = 28, lbl = 0, pr_w = 45; + xpos = 0, bw = 28, bh = 28, fh = 0, lh = 28, lbl = 0; } else if (PanelSize == P_medium) { if (Small_Icons) - xpos = 0, bw = 42, bh = 36, fh = 0, lh = 22, lbl = 1, pr_w = 60; + xpos = 0, bw = 42, bh = 36, fh = 0, lh = 22, lbl = 1; else - xpos = 0, bw = 45, bh = 45, fh = 0, lh = 28, lbl = 1, pr_w = 60; + xpos = 0, bw = 45, bh = 45, fh = 0, lh = 28, lbl = 1; } else { // P_large if (Small_Icons) - xpos = 0, bw = 42, bh = 36, fh = 22, lh = 22, lbl = 1, pr_w = 60; + xpos = 0, bw = 42, bh = 36, fh = 22, lh = 22, lbl = 1; else - xpos = 0, bw = 45, bh = 45, fh = 28, lh = 28, lbl = 1, pr_w = 60; + xpos = 0, bw = 45, bh = 45, fh = 28, lh = 28, lbl = 1; } if (PanelSize == P_tiny) { @@ -767,6 +773,7 @@ void UI::set_page_prog(size_t nbytes, in void UI::set_page_prog(size_t nbytes, int cmd) { char str[32]; + int width, height; if (cmd == 0) { PProg->deactivate(); @@ -779,6 +786,8 @@ void UI::set_page_prog(size_t nbytes, in str[0] = '\0'; } PProg->copy_label(str); + PProg->measure_label(width, height); + PProg->resize(width + 4, height + 4); PProg->redraw_label(); } } @@ -790,6 +799,7 @@ void UI::set_img_prog(int n_img, int t_i void UI::set_img_prog(int n_img, int t_img, int cmd) { char str[32]; + int width, height; if (cmd == 0) { IProg->deactivate(); @@ -802,6 +812,8 @@ void UI::set_img_prog(int n_img, int t_i str[0] = '\0'; } IProg->copy_label(str); + IProg->measure_label(width, height); + IProg->resize(width + 4, height + 4); IProg->redraw_label(); } } diff -r 0a8a112b8e48 src/ui.hh --- a/src/ui.hh Mon Dec 17 14:07:38 2007 +0100 +++ b/src/ui.hh Mon Dec 17 19:33:53 2007 +0100 @@ -48,7 +48,7 @@ class UI : public fltk::Window { int MainIdx; // Panel customization variables int PanelSize, CuteColor, Small_Icons; - int xpos, bw, bh, fh, lh, lbl, pr_w; + int xpos, bw, bh, fh, lh, lbl; // TODO: Hack for fullscreen mode int Panel_h, Status_h;

2 3

[PATCH] FltkMultilineTextResource: <Enter> is normal
by onepoint＠starurchin.org Dec. 19, 2007

Dec. 19, 2007

This patch removes the special behaviour of the <Enter> key for the FltkMultiLineTextResource. In fact it removes the callback altogether, the resource simply holds text and has no activity. Regards, Jeremy Henty diff -pru -- dw2-ref/dw/fltkui.cc dw2-cur/dw/fltkui.cc --- dw2-ref/dw/fltkui.cc 2007-12-17 23:39:52.000000000 +0000 +++ dw2-cur/dw/fltkui.cc 2007-12-18 16:05:59.000000000 +0000 @@ -552,8 +552,6 @@ FltkMultiLineTextResource::~FltkMultiLin new ::fltk::MultiLineInput (allocation->x, allocation->y, allocation->width, allocation->ascent + allocation->descent); - input->callback (widgetCallback, this); - input->when (::fltk::WHEN_ENTER_KEY_ALWAYS); if (viewsAndWidgets->isEmpty ()) { // First widget created, attach the set text. @@ -586,21 +584,6 @@ void FltkMultiLineTextResource::sizeRequ } } -void FltkMultiLineTextResource::widgetCallback (::fltk::Widget *widget, - void *data) -{ - /* The (::fltk::event_key() == ::fltk::ReturnKey) test - * is necessary because WHEN_ENTER_KEY also includes - * other events we're not interested in. For instance pressing - * The Back or Forward, buttons, or the first click on a rendered - * page. BUG: this must be investigated and reported to FLTK2 team - */ - printf ("when = %d\n", widget->when ()); - if ((widget->when () & ::fltk::WHEN_ENTER_KEY_ALWAYS) && - (::fltk::event_key() == ::fltk::ReturnKey)) - ((FltkMultiLineTextResource*)data)->emitActivate (); -} - const char *FltkMultiLineTextResource::getText () { if (viewsAndWidgets->isEmpty ()) diff -pru -- dw2-ref/dw/fltkui.hh dw2-cur/dw/fltkui.hh --- dw2-ref/dw/fltkui.hh 2007-12-17 16:00:34.000000000 +0000 +++ dw2-cur/dw/fltkui.hh 2007-12-18 16:06:11.000000000 +0000 @@ -321,8 +321,6 @@ private: bool editable; int numCols, numRows; - static void widgetCallback (::fltk::Widget *widget, void *data); - protected: ::fltk::Widget *createNewWidget (core::Allocation *allocation);

2 1