December 2007 - Dillo-dev - lists.mailman3.com

should new windows inherit panel size and fullscreen mode?

by Johannes.Hofmann＠gmx.de

Hello, I would like that new browser windows inherit the panel size and fullscreen mode from the current window. What do others think? Cheers, Johannes

Dec. 11, 2007

Re: Colspan and table percentages

by Johannes.Hofmann＠gmx.de

On Tue, Dec 11, 2007 at 10:16:01AM -0300, Jorge Arellano Cid wrote: > On Mon, Dec 10, 2007 at 06:22:36PM +0100, Johannes Hofmann wrote: > > Hi Jorge, > > > > just noticed that http://www.dillo.org/MList.html still looks weird > > with dillo-fltk: The main text is rendered just one word wide here. > > Yes, that's a special colspan case. > > Please test now! Works fine now. Thanks, Johannes > > -- > Cheers > Jorge.- > > _______________________________________________ > Dillo-dev mailing list > Dillo-dev(a)dillo.org > http://lists.auriga.wearlab.de/cgi-bin/mailman/listinfo/dillo-dev

Dec. 11, 2007

BUG: assertion failed: "no widget": in dw::core::DeepIterator::searchSideward

by Johannes.Hofmann＠gmx.de

On Tue, Dec 11, 2007 at 12:06:37PM +0000, Jeremy Henty wrote: > On Mon, Dec 10, 2007 at 05:44:08PM +0100, Johannes Hofmann wrote: > > > Just one thing. I would rather have tested the return value of > > getTopIterator() for NULL. > > Heh, that *was* my original implementation of isEmpty() . Then I > noticed the hasContents member so I used that since it shortened the > code. I believe hasContents could be eliminated since it is true when > and only when the stack is non-empty, though I haven't exhaustively > checked this. > > > I think the code is generally a bit optimistic here by assuming > > non-NULL return value from getTopIterator(). Even if this is now > > always the case with your fix, it may change again in the future. > > Yes it may. But I think much of the code assumes this, so if this > changes in the future then a lot of code will need to be reviewed, > preferably by someone who understands the design. > > > What do you think, if we add checks for NULL at those three points > > in the code where getTopIterator() is called, do we still need to > > check with isEmpty() to call resetLink() / resetSelection() in > > buttonPress()? > > Yes, otherwise you get a crash in DeepIterator::compareTo, which also > assumes that the stack is non-empty. (In fact the comments suggest > that it assumes the tops of the stacks of the two DeepIterator objects > point to the same widget.) When I found that crash I decided to give > up on rewriting lots of code to handle empty DeepIterators and instead > simply prevent empty DeepIterators appearing in Selection objects. > > Even if other code ought to handle empty DeepIterators better, I think > it's reasonable to argue that Selection objects should never use them. > After all, what could it possible mean for a selection to begin or end > at an empty DeepIterator? So I think my bugfix makes Selections do > the right thing, even if other code should be checking > getTopIterator() as you suggest. Yes, I fully agree with you. Selection is fixed. We may need to come back to DeepIterator later. Regards, Johannes > > Regards, > > Jeremy Henty > > _______________________________________________ > Dillo-dev mailing list > Dillo-dev(a)dillo.org > http://lists.auriga.wearlab.de/cgi-bin/mailman/listinfo/dillo-dev

Dec. 11, 2007

Re: Colspan and table percentages

by onepoint＠starurchin.org

On Tue, Dec 11, 2007 at 01:04:02PM +0000, Jeremy Henty (ie. me) wrote: > I've attached a simple example below. It depends on the table cell > having colspan="2". Not only does each word appear on a line by > itself OK, this is now fixed. > ... but no horizontal scrollbar appears even though one word extends > past the right of the page. This issue remains. Regards, Jeremy Henty <?xml version="1.0" encoding="UTF-8"?> <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"> <html> <head> <title> Dillo: display and layout strangeness in a table cell with colspan="2" </title> </head> <body> <table> <tr> <td colspan="2"> aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa bbbbb cccccc dddddddd eeeeee </td> </tr> </table> </body> </html> _______________________________________________ Dillo-dev mailing list Dillo-dev(a)dillo.org http://lists.auriga.wearlab.de/cgi-bin/mailman/listinfo/dillo-dev

Dec. 11, 2007

[patch] fix close button height in page source window

by jcid＠dillo.org

On Mon, Dec 10, 2007 at 06:15:56PM +0100, Johannes Hofmann wrote: > Hi, > > the height of the close button of the source code window changes > with the size of the window. The patch below fixes its size. Committed. > PS: Scrolling in the page source window is dead slow here. But > that's definately a fltk issue... Yes, and we still have the resort of showing the source inside a new dillo browser window (this would add find text, but lose the nice line numbers), with the gain of faster scrolling. Now, if we delegate viewsource to a dpi, it would be simple to add the line numbers. Just some thoughts. -- Cheers Jorge.-

Dec. 11, 2007

Re: Colspan and table percentages

by jcid＠dillo.org

On Mon, Dec 10, 2007 at 06:22:36PM +0100, Johannes Hofmann wrote: > Hi Jorge, > > just noticed that http://www.dillo.org/MList.html still looks weird > with dillo-fltk: The main text is rendered just one word wide here. Yes, that's a special colspan case. Please test now! -- Cheers Jorge.-

Dec. 11, 2007

Re: Colspan and table percentages

by onepoint＠starurchin.org

On Mon, Dec 10, 2007 at 06:22:36PM +0100, Johannes Hofmann wrote: > just noticed that http://www.dillo.org/MList.html still looks weird > with dillo-fltk: The main text is rendered just one word wide here. I see this quite a lot. I've attached a simple example below. It depends on the table cell having colspan="2". Not only does each word appear on a line by itself but no horizontal scrollbar appears even though one word extends past the right of the page. Regards, Jeremy Henty <?xml version="1.0" encoding="UTF-8"?> <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"> <html> <head> <title> Dillo: display and layout strangeness in a table cell with colspan="2" </title> </head> <body> <table> <tr> <td colspan="2"> aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa bbbbb cccccc dddddddd eeeeee </td> </tr> </table> </body> </html>

Dec. 11, 2007

BUG: assertion failed: "no widget": in dw::core::DeepIterator::searchSideward

by onepoint＠starurchin.org

On Mon, Dec 10, 2007 at 05:44:08PM +0100, Johannes Hofmann wrote: > Just one thing. I would rather have tested the return value of > getTopIterator() for NULL. Heh, that *was* my original implementation of isEmpty() . Then I noticed the hasContents member so I used that since it shortened the code. I believe hasContents could be eliminated since it is true when and only when the stack is non-empty, though I haven't exhaustively checked this. > I think the code is generally a bit optimistic here by assuming > non-NULL return value from getTopIterator(). Even if this is now > always the case with your fix, it may change again in the future. Yes it may. But I think much of the code assumes this, so if this changes in the future then a lot of code will need to be reviewed, preferably by someone who understands the design. > What do you think, if we add checks for NULL at those three points > in the code where getTopIterator() is called, do we still need to > check with isEmpty() to call resetLink() / resetSelection() in > buttonPress()? Yes, otherwise you get a crash in DeepIterator::compareTo, which also assumes that the stack is non-empty. (In fact the comments suggest that it assumes the tops of the stacks of the two DeepIterator objects point to the same widget.) When I found that crash I decided to give up on rewriting lots of code to handle empty DeepIterators and instead simply prevent empty DeepIterators appearing in Selection objects. Even if other code ought to handle empty DeepIterators better, I think it's reasonable to argue that Selection objects should never use them. After all, what could it possible mean for a selection to begin or end at an empty DeepIterator? So I think my bugfix makes Selections do the right thing, even if other code should be checking getTopIterator() as you suggest. Regards, Jeremy Henty

Dec. 11, 2007

Re: Colspan and table percentages

by Johannes.Hofmann＠gmx.de

Hi Jorge, just noticed that http://www.dillo.org/MList.html still looks weird with dillo-fltk: The main text is rendered just one word wide here. Cheers, Johannes On Sun, Dec 09, 2007 at 05:58:04PM -0300, Jorge Arellano Cid wrote: > On Sun, Dec 09, 2007 at 02:07:48PM -0300, Jorge Arellano Cid wrote: > > Hi! > > > > After a lot of work, the patch for better table rendering > > (regarding colspan and percentages) was good enough to be > > committed. Please enjoy and send feedback (or patches) ;-). > > > > - Implemented a weighted apportionment algorithm for table colspan. > > - Implemented percentage widths in tables (better rendering!). > > - Fixed an initialization bug in hruler. > > + Minor fix to table-wide percentage apportion. > > Now about:splash renders OK. > > -- > Cheers > Jorge.- > > _______________________________________________ > Dillo-dev mailing list > Dillo-dev(a)dillo.org > http://lists.auriga.wearlab.de/cgi-bin/mailman/listinfo/dillo-dev

Dec. 10, 2007

[patch] fix close button height in page source window

by Johannes.Hofmann＠gmx.de

Hi, the height of the close button of the source code window changes with the size of the window. The patch below fixes its size. Cheers, Johannes PS: Scrolling in the page source window is dead slow here. But that's definately a fltk issue... diff -r 043b53a25dc3 src/dialog.cc --- a/src/dialog.cc Mon Dec 10 15:42:03 2007 +0100 +++ b/src/dialog.cc Mon Dec 10 18:12:15 2007 +0100 @@ -123,7 +123,7 @@ void a_Dialog_text_window(const char *tx ReturnButton *b = new ReturnButton (0, wh-bh, ww, bh, "Close"); b->callback(window_close_cb, window); - window->resizable(window); + window->resizable(td); window->end(); window->show(); }

Dec. 10, 2007