April 2015 - Dillo-dev - lists.mailman3.com

Crash with ascii text scribbled across a data structure.

by james.from.wellington＠gmail.com

>From time to time, I get a most peculiar SEGV. Here's an example on fairly clean source up to c761c66 on a mac: Program received signal EXC_BAD_ACCESS, Could not access memory. Reason: 13 at address: 0x0000000000000000 0x00000001000064b7 in Fl_Widget::position (this=0x118148910, X=2003792928, Y=1869881479) at Fl_Widget.H:321 321 void position(int X,int Y) {resize(X,Y,w_,h_);} (gdb) bt #0 0x00000001000064b7 in Fl_Widget::position (this=0x118148910, X=2003792928, Y=1869881479) at Fl_Widget.H:321 #1 0x000000010008521a in dw::fltk::FltkViewport::updateCanvasWidgets (this=0x101208a10, dx=0, dy=-25) at fltkviewport.cc:502 #2 0x0000000100086492 in dw::fltk::FltkViewport::scrollTo (this=0x101208a10, x=0, y=347) at fltkviewport.cc:398 #3 0x0000000100084c10 in dw::fltk::FltkViewport::scroll (this=0x101208a10, dx=0, dy=-25) at fltkviewport.cc:413 #4 0x0000000100084c4a in dw::fltk::FltkViewport::vscrollbarChanged (this=0x101208a10) at fltkviewport.cc:156 #5 0x0000000100084c65 in dw::fltk::FltkViewport::vscrollbarCallback (vscrollbar=0x101208c90, viewportPtr=0x101208a10) at fltkviewport.cc:161 #6 0x0000000100100b13 in Fl_Widget::do_callback (this=0x101208c90, o=0x101208c90, arg=0x101208a10) at Fl_Widget.cxx:328 #7 0x000000010001a8ef in Fl_Widget::do_callback (this=0x101208c90) at Fl_Widget.H:840 #8 0x00000001001005c1 in Fl_Valuator::handle_drag (this=0x101208c90, v=347) at Fl_Valuator.cxx:95 #9 0x00000001000ec3bc in Fl_Scrollbar::handle (this=0x101208c90, event=19) at Fl_Scrollbar.cxx:142 #10 0x00000001000857d7 in dw::fltk::FltkViewport::handle (this=0x101208a10, event=19) at fltkviewport.cc:304 #11 0x00000001000ceb39 in send (o=0x101208a18, event=19) at Fl_Group.cxx:98 #12 0x00000001000cf338 in Fl_Group::handle (this=0x10053d760, event=19) at Fl_Group.cxx:254 #13 0x0000000100007aa2 in UI::handle (this=0x10053d760, event=19) at ui.cc:788 #14 0x00000001000ceb39 in send (o=0x10053d760, event=19) at Fl_Group.cxx:98 #15 0x00000001000cf338 in Fl_Group::handle (this=0x10053d6b0, event=19) at Fl_Group.cxx:254 #16 0x00000001000ceb39 in send (o=0x10053d6b0, event=19) at Fl_Group.cxx:98 #17 0x00000001000cf338 in Fl_Group::handle (this=0x10053cf60, event=19) at Fl_Group.cxx:254 #18 0x00000001000b7b94 in Fl_Window::handle (this=0x10053cf60, ev=19) at Fl.cxx:1643 #19 0x00000001000b67e0 in send (event=19, to=0x10053cf60, window=0x10053cf60) at Fl.cxx:1136 #20 0x00000001000b8ad1 in Fl::handle_ (e=19, window=0x10053cf60) at Fl.cxx:1444 #21 0x00000001000b8b7a in Fl::handle (e=19, window=0x10053cf60) at Fl.cxx:1216 #22 0x00000001000ab1b3 in cocoaMouseWheelHandler (theEvent=0x1181582b0) at Fl_cocoa.mm:943 #23 0x00000001000ab1e8 in -[FLView scrollWheel:] (self=0x101505a50, _cmd=0x7fff8965ff98, theEvent=0x1181582b0) at Fl_cocoa.mm:2250 #24 0x00007fff890580c7 in -[NSWindow sendEvent:] () #25 0x00007fff88f8cafa in -[NSApplication sendEvent:] () #26 0x00000001000aed74 in +[FLApplication sendEvent:] (self=0x1001b36d0, _cmd=0x7fff896552c0, theEvent=0x1181582b0) at Fl_cocoa.mm:1609 #27 0x00000001000ae989 in do_queued_events (time=1e+20) at Fl_cocoa.mm:749 #28 0x00000001000ae9bf in fl_wait (time=1e+20) at Fl_cocoa.mm:773 #29 0x00000001000aea86 in fl_mac_flush_and_wait (time_to_wait=1e+20) at Fl_cocoa.mm:792 #30 0x00000001000b9081 in Fl::wait (time_to_wait=1e+20) at Fl.cxx:528 #31 0x00000001000b910b in Fl::run () at Fl.cxx:589 #32 0x00000001000052a7 in main (argc=1, argv=0x7fff5fbff280) at dillo.cc:589 (gdb) print (char *)this $6 = 0x118148910 "lore ipsum the text out of the editor window in the email.\n\nMore lorem ibsum.\n\nRegards,\nJames." 0x118148910 is correctly allocated to dw::flak::ui::FltkLabelButtonResource, according to Apple's debugging malloc. I can't find, in the debugging malloc, the allocation for storing the text in the text-box. The worst cases are that someone is scribbling malloc control structures, or someone is scribbling addresses in a dillo data structure. I've seen something essentially the same where the scribble contained a URL. I'm contemplating relinking with electric fence. Regards, James.

April 22, 2015

website colors (Re: Secure: yes, Though Features Missing; and Bugs)

by eocene＠gmx.com

I wrote: > Miroslav wrote: > > On Sat, Apr 04, 2015 at 05:24:46AM +0000, eocene wrote: > > > A couple of years ago, I experimented with a different color scheme, but > > > the other members of the core team weren't enthusiastic about it, so I > > > never made the change. Example: http://www.dillo.org/test/dillo2.html > > > > > Pls, pls. dillo devs, allow that color scheme. > > > > If that were the default, I wouldn't have so much difficulties. And > > surely others!, as I wrote. Why do large associations, businesses, > > others, never, almost never have unreadable pages?, never experiment > > with colors --such as, worth mentioning, some smaller, always only > > smaller entities, make pages in revert, white on black: which is almost > > never good, unless you want to exclude readers... > > > > Pls. dillo devs, allow that color scheme. > > Thank you! > > Well, that's 2 votes for it... And we're at zero votes against, here...

April 12, 2015

website colors (Re: Secure: yes, Though Features Missing; and Bugs)

by eocene＠gmx.com

Miroslav wrote: > On Sat, Apr 04, 2015 at 05:24:46AM +0000, eocene wrote: > > A couple of years ago, I experimented with a different color scheme, but > > the other members of the core team weren't enthusiastic about it, so I > > never made the change. Example: http://www.dillo.org/test/dillo2.html > > > Pls, pls. dillo devs, allow that color scheme. > > If that were the default, I wouldn't have so much difficulties. And > surely others!, as I wrote. Why do large associations, businesses, > others, never, almost never have unreadable pages?, never experiment > with colors --such as, worth mentioning, some smaller, always only > smaller entities, make pages in revert, white on black: which is almost > never good, unless you want to exclude readers... > > Pls. dillo devs, allow that color scheme. > Thank you! Well, that's 2 votes for it...

April 5, 2015

Documentation fixes: autolinking

by onepoint＠starurchin.org

Sebastian Geerken wrote: > Sorry for the late response. > > On Sa, Apr 04, 2015, Jeremy Henty wrote: > > > > I wrote: > > > > > Jorge Arellano Cid wrote: > > > > > > > On Mon, Mar 16, 2015 at 03:28:43PM +0000, Jeremy Henty wrote: > > > > > > > > > > Three small patches that fix autolinking errors in the documentation. > > > > > The first two escape the autolinking syntax in a few places where it > > > > > was not intended to be a link. The third fixes a few places where a > > > > > link was intended but the autolink syntax was not quite right. > > > > > > > > > > http://starurchin.org/dillo/patch/doxygen-escape-hash-characters.html > > > > > http://starurchin.org/dillo/patch/doxygen-escape-double-colons.html > > > > > http://starurchin.org/dillo/patch/doxygen-fix-simple-name-refs.html > > > > > > > > > > Comments are welcome. > > > > > > > > > > If these are OK I can push them myself as I have write permission. > > > > > (It's actually easier for me to do it myself because Mercurial Queues > > > > > does not notice when the remote repository has a patch that is > > > > > identical to a patch in its queue and generates spurious errors.) > > > > > > > > +1. > > > > > > > > Just not sure about the last one. I don't know whether it > > > > changes the intended semantic for human readers. > > > > > > I think it actually corrects the semantics. An initial '::' means > > > resolve the name in the top namespace, not the current one, but the > > > class names used in the docs are relative to dw/lout. So I think the > > > patch does the right thing. > > > > Bump! Is it OK for me to push these or is more discussion needed? > > Haven't yet tested them, but they look OK; so push them if they don't > break anything. OK, it is done. > As for "::" vs. "\ref": I added "::" to have a reference at all. I'm > not sure about the exact rules, but Doxygen seems to regard anything > containing "::" as reference. > > If "\ref" works, it looks cleaner IMO. I agree. Regards, Jeremy Henty

April 4, 2015

Documentation fixes: autolinking

by sgeerken＠dillo.org

Sorry for the late response. On Sa, Apr 04, 2015, Jeremy Henty wrote: > > I wrote: > > > Jorge Arellano Cid wrote: > > > > > On Mon, Mar 16, 2015 at 03:28:43PM +0000, Jeremy Henty wrote: > > > > > > > > Three small patches that fix autolinking errors in the documentation. > > > > The first two escape the autolinking syntax in a few places where it > > > > was not intended to be a link. The third fixes a few places where a > > > > link was intended but the autolink syntax was not quite right. > > > > > > > > http://starurchin.org/dillo/patch/doxygen-escape-hash-characters.html > > > > http://starurchin.org/dillo/patch/doxygen-escape-double-colons.html > > > > http://starurchin.org/dillo/patch/doxygen-fix-simple-name-refs.html > > > > > > > > Comments are welcome. > > > > > > > > If these are OK I can push them myself as I have write permission. > > > > (It's actually easier for me to do it myself because Mercurial Queues > > > > does not notice when the remote repository has a patch that is > > > > identical to a patch in its queue and generates spurious errors.) > > > > > > +1. > > > > > > Just not sure about the last one. I don't know whether it > > > changes the intended semantic for human readers. > > > > I think it actually corrects the semantics. An initial '::' means > > resolve the name in the top namespace, not the current one, but the > > class names used in the docs are relative to dw/lout. So I think the > > patch does the right thing. > > Bump! Is it OK for me to push these or is more discussion needed? Haven't yet tested them, but they look OK; so push them if they don't break anything. As for "::" vs. "\ref": I added "::" to have a reference at all. I'm not sure about the exact rules, but Doxygen seems to regard anything containing "::" as reference. If "\ref" works, it looks cleaner IMO. Sebastian

April 4, 2015

Secure: yes, Though Features Missing; and Bugs

by miroslav.rovis1＠zg.ht.hr

On Sat, Apr 04, 2015 at 05:24:46AM +0000, eocene wrote: > Miroslav wrote: > > But, since I started about it here, let me explain. Look here, how my > > dillo brings pages to me currently: > > https://forums.gentoo.org/viewtopic-t-1004014.html#7724054 > > and currently, download and open the mkv files... (So people reading > > later, open the, hopefully, webm files. That's how it was, talking to > > you in the future.) > > I don't think XFT is enabled in your FLTK. That would explain why you > aren't getting DejaVu Sans and the rest... > > > ##################################################################### > > The Homepage is particularly too hard to read for, I guess at least a > > non-insignificant though small percentage (no, not less than a small > > percentage number, it's not insignificant!) of visitors. The background > > must be very light. As soon as it is too dark, the font is not easily > > readable. > > ##################################################################### > > A couple of years ago, I experimented with a different color scheme, but > the other members of the core team weren't enthusiastic about it, so I > never made the change. Example: http://www.dillo.org/test/dillo2.html > Pls, pls. dillo devs, allow that color scheme. If that were the default, I wouldn't have so much difficulties. And surely others!, as I wrote. Why do large associations, businesses, others, never, almost never have unreadable pages?, never experiment with colors --such as, worth mentioning, some smaller, always only smaller entities, make pages in revert, white on black: which is almost never good, unless you want to exclude readers... Pls. dillo devs, allow that color scheme. Thank you! > > _______________________________________________ > Dillo-dev mailing list > Dillo-dev at dillo.org > http://lists.dillo.org/cgi-bin/mailman/listinfo/dillo-dev -- Miroslav Rovis Zagreb, Croatia http://www.CroatiaFidelis.hr

April 4, 2015

Secure: yes, Though Features Missing; and Bugs

by eocene＠gmx.com

Miroslav wrote: > But, since I started about it here, let me explain. Look here, how my > dillo brings pages to me currently: > https://forums.gentoo.org/viewtopic-t-1004014.html#7724054 > and currently, download and open the mkv files... (So people reading > later, open the, hopefully, webm files. That's how it was, talking to > you in the future.) I don't think XFT is enabled in your FLTK. That would explain why you aren't getting DejaVu Sans and the rest... > ##################################################################### > The Homepage is particularly too hard to read for, I guess at least a > non-insignificant though small percentage (no, not less than a small > percentage number, it's not insignificant!) of visitors. The background > must be very light. As soon as it is too dark, the font is not easily > readable. > ##################################################################### A couple of years ago, I experimented with a different color scheme, but the other members of the core team weren't enthusiastic about it, so I never made the change. Example: http://www.dillo.org/test/dillo2.html

April 4, 2015

patch: provide a redirection-blocked page

by eocene＠gmx.com

I was in the mood to make the use of domainrc more pleasant. Having to check the terminal for redirection-denied MSGs is no fun, and sometimes I imagine that I'm the only person who uses it due to its...limited friendliness. Patch attached.

April 4, 2015

Documentation fixes: autolinking

by onepoint＠starurchin.org

I wrote: > Jorge Arellano Cid wrote: > > > On Mon, Mar 16, 2015 at 03:28:43PM +0000, Jeremy Henty wrote: > > > > > > Three small patches that fix autolinking errors in the documentation. > > > The first two escape the autolinking syntax in a few places where it > > > was not intended to be a link. The third fixes a few places where a > > > link was intended but the autolink syntax was not quite right. > > > > > > http://starurchin.org/dillo/patch/doxygen-escape-hash-characters.html > > > http://starurchin.org/dillo/patch/doxygen-escape-double-colons.html > > > http://starurchin.org/dillo/patch/doxygen-fix-simple-name-refs.html > > > > > > Comments are welcome. > > > > > > If these are OK I can push them myself as I have write permission. > > > (It's actually easier for me to do it myself because Mercurial Queues > > > does not notice when the remote repository has a patch that is > > > identical to a patch in its queue and generates spurious errors.) > > > > +1. > > > > Just not sure about the last one. I don't know whether it > > changes the intended semantic for human readers. > > I think it actually corrects the semantics. An initial '::' means > resolve the name in the top namespace, not the current one, but the > class names used in the docs are relative to dw/lout. So I think the > patch does the right thing. Bump! Is it OK for me to push these or is more discussion needed? Regards, Jeremy Henty

April 4, 2015

Secure: yes, Though Features Missing; and Bugs

by miroslav.rovis1＠zg.ht.hr

( I already wrote this entire message, and only then noticed eocene didn't send this to dillo-dev but only to me. But I now have to, All written. Can't correct that now. ) On Fri, Apr 03, 2015 at 08:14:36PM +0000, eocene wrote: > miroslav.rovis1 wrote: > > ukrainian at uabox ~ $ Domain: Default accept. > > By the way, if you're interested in making your browsing more secure, > you could modify your domainrc file not to have 'accept' as its default > rule. Let me first reply to this message. I really wish to start viewing dillo more comfortably first, so this will wait. But, since I started about it here, let me explain. Look here, how my dillo brings pages to me currently: https://forums.gentoo.org/viewtopic-t-1004014.html#7724054 and currently, download and open the mkv files... (So people reading later, open the, hopefully, webm files. That's how it was, talking to you in the future.) Those mkv files will, hopefully, soon be replaced with the HTML5 compliant webm files. So if opened those, that's how it shows currently. I was only able to use dillo after reverting my 1024x768 capable old monitor, to 800x600, in another display in X (:1.0 in, say, ffmpeg command line), else, I would have gone blind from effort to read. And I still have difficulty reading some of the pages. ##################################################################### The Homepage is particularly too hard to read for, I guess at least a non-insignificant though small percentage (no, not less than a small percentage number, it's not insignificant!) of visitors. The background must be very light. As soon as it is too dark, the font is not easily readable. ##################################################################### I'm not at poor-accessibility eyesight yet, I don't have problems with most of pages. I really don't. Pls. take notice of this claim of mine and value the outlined paragraph with respect to that claim! I hope this will help dillo homepage become more readable. So, first I really need the fonts, and will follow your kind advice, eocene, on the fonts, as soon as I find more free time. domainrc is great thing to try to sort too, but the other issue is urgent. Pls allow more time for me to reply to your other mail (I got two mails of kind advice from eocene. Can't go online (the SourceForge archive) to see if my bad provider and bad regime filtered some emails out from dillo-dev. Busy elsewhere. -- Miroslav Rovis Zagreb, Croatia http://www.CroatiaFidelis.hr

April 3, 2015