- Dillo-dev - lists.mailman3.com

Garbage chars added

by corvid＠lavabit.com

Tomas wrote: > "corvid" <corvid(a)lavabit.com> wrote: > > > Tomas wrote: > > > What's causing the addition of garbage symbols, like here (comments > > > by tr (6th & 8th comment(most visible in 8th))): > > > http://www.nezinau.lt/microsoft-lazerines-peles-sugeba-nustebinti#comments > > > Backspace? > > > > Looks like it. > > > > I pasted a random chunk of text ending in a non-ASCII character into > > a textarea with a form submitting as UTF-8. > > "mink?" became "mink%C5%A1" in the submitted data, which I believe is > > right. Then I pressed backspace in the textarea, leaving "mink", and > > submitted again. Now it submits "mink%A1", which is wrong. > > Input type=text doesn't seem to have the problem. > > Probably something wrong in fltk's TextEditor, then. I'll see if I can > > find where it's missing some utf8-awareness. > > > > Hey you were quick to figure that out :) In fltk/src/TextEditor.cxx, you could try replacing the kf_backspace with: int TextEditor::kf_backspace(int, TextEditor* e) { int oldpos = e->insert_position(); if (!e->buffer()->selected() && e->move_left()) e->buffer()->select(e->insert_position(), oldpos); kill_selection(e); e->show_insert_position(); e->maybe_do_callback(); return 1; } Looks like their kf_delete is broken as well...

Oct. 23, 2008

More love for Dillo2!

by jcid＠dillo.org

On Wed, Oct 22, 2008 at 11:00:11PM +0100, Jeremy Henty wrote: > > Linux Today linked to this: > > TuxArena: Dillo 2.0 Gets Tabbed Browsing > > http://tuxarena.blogspot.com/2008/10/dillo-20-gets-tabbed-browsing.html Good! Link added in "Press" section. -- Cheers Jorge.-

Oct. 23, 2008

Garbage chars added

by sandshrew＠gmail.com

On Thu, 23 Oct 2008 14:43:09 +0000 "corvid" <corvid(a)lavabit.com> wrote: > Tomas wrote: > > What's causing the addition of garbage symbols, like here (comments > > by tr (6th & 8th comment(most visible in 8th))): > > http://www.nezinau.lt/microsoft-lazerines-peles-sugeba-nustebinti#comments > > Backspace? > > Looks like it. > > I pasted a random chunk of text ending in a non-ASCII character into > a textarea with a form submitting as UTF-8. > "mink?" became "mink%C5%A1" in the submitted data, which I believe is > right. Then I pressed backspace in the textarea, leaving "mink", and > submitted again. Now it submits "mink%A1", which is wrong. > Input type=text doesn't seem to have the problem. > Probably something wrong in fltk's TextEditor, then. I'll see if I can > find where it's missing some utf8-awareness. > Hey you were quick to figure that out :) -- Brain: an apparatus with which we think we think. - A. Bierce

Oct. 23, 2008

Garbage chars added

by corvid＠lavabit.com

Tomas wrote: > What's causing the addition of garbage symbols, like here (comments by > tr (6th & 8th comment(most visible in 8th))): > http://www.nezinau.lt/microsoft-lazerines-peles-sugeba-nustebinti#comments > Backspace? Looks like it. I pasted a random chunk of text ending in a non-ASCII character into a textarea with a form submitting as UTF-8. "mink?" became "mink%C5%A1" in the submitted data, which I believe is right. Then I pressed backspace in the textarea, leaving "mink", and submitted again. Now it submits "mink%A1", which is wrong. Input type=text doesn't seem to have the problem. Probably something wrong in fltk's TextEditor, then. I'll see if I can find where it's missing some utf8-awareness.

Oct. 23, 2008

css

by Johannes.Hofmann＠gmx.de

On Wed, Oct 22, 2008 at 10:33:44PM +0000, corvid wrote: > Johannes wrote: > > This approach has some issues. In contrast to > > http://www.dillo.org/CSS.html it would only be possible to apply > > style information that is available when starting parsing of a > > document. However we could start rendering with what style > > information is available and reload once all css files have been > > loaded. > > So is the idea that your method is more lightweight and a > repush() once the CSS arrives looks the same as applying > the CSS on the fly? That's the idea. > > If this is deemed superior, there might be things we could rip out of > dw that were written with the other scheme in mind. > No, I don't think this is superior. However, we need to be able to apply available styles while parsing anyway. So I would start coding that. And maybe this together with the repush thing gives acceptable behaviour already. Cheers, Johannes

Oct. 23, 2008

Garbage chars added

by francis＠daoine.org

On Thu, Oct 23, 2008 at 03:52:56PM +0300, Tomas R wrote: > On Thu, 23 Oct 2008 13:43:04 +0100 > Francis Daly <francis(a)daoine.org> wrote: > > On Thu, Oct 23, 2008 at 02:34:10PM +0300, Tomas R wrote: Hi there, > > > What's causing the addition of garbage symbols, like here (comments > > > by tr (6th & 8th comment(most visible in 8th))): > > > http://www.nezinau.lt/microsoft-lazerines-peles-sugeba-nustebinti#comments > > > Backspace? > > If you're asking "did dillo put it there previously when uploading a > > comment", that's something I can't answer easily -- perhaps capturing > > the network traffic when uploading will show whether it is in dillo's > > output; if so, more investigation will be needed to find why. > > I am asking WHY dillo put them there. The comments containing > garbage symbols were submitted by me with dillo2 and I think I used > Backspace where the garbage chars are added (check the page with some > full blown browser). I've taken a local copy of the html page, modified the form so that it points to a local "echo all input" cgi script, and I do not see any extra characters added. I've pasted in some of the (accented) text from the main page; I've deleted and backspaced and added my own text within dillo; and when I press "Tinka!", I see nothing unusual in the bytes returned by my "echo" script. I presume that means there's nothing unusual in the bytes sent by dillo. Right now, I can't reproduce the problem you're reporting. Can you provide a reproduction recipe? Preferably one that can be run using a local web server, so that we won't be adding unnecessary test comments to someone's live site. The garbage characters on the page aren't all the same character. Presumably either dillo is adding in the rubbish, or is misinterpreting the bytes written in the comment, or else the receiving side is somehow breaking things. When we can see the bytes being sent by dillo, we'll know which half (sender/receiver) is broken. Good luck, f -- Francis Daly francis(a)daoine.org

Oct. 23, 2008

Garbage chars added

by francis＠daoine.org

On Thu, Oct 23, 2008 at 02:34:10PM +0300, Tomas R wrote: > What's causing the addition of garbage symbols, like here (comments by > tr (6th & 8th comment(most visible in 8th))): > http://www.nezinau.lt/microsoft-lazerines-peles-sugeba-nustebinti#comments > Backspace? The page is claimed to be UTF-8, but it contains 6 invalid byte sequences (according to wc). So, there's garbage in the HTML which is interpreted as garbage. Dillo shows it because it is there. If you're asking "did dillo put it there previously when uploading a comment", that's something I can't answer easily -- perhaps capturing the network traffic when uploading will show whether it is in dillo's output; if so, more investigation will be needed to find why. Good luck, f -- Francis Daly francis(a)daoine.org

Oct. 23, 2008

Garbage chars added

by sandshrew＠gmail.com

What's causing the addition of garbage symbols, like here (comments by tr (6th & 8th comment(most visible in 8th))): http://www.nezinau.lt/microsoft-lazerines-peles-sugeba-nustebinti#comments Backspace? -- Brain: an apparatus with which we think we think. - A. Bierce

Oct. 23, 2008

pasting URL

by thomas-forum＠orgis.org

Am Wed, 22 Oct 2008 22:46:52 +0100 schrieb higuita <higuita7(a)yahoo.co.uk>: > you can already paste a URL to the X icon in the urlbar and it > will load... Indeed. Good to know. > the midle click on the body is reserved to dragging the page Now that's funny. I have my middle button configured to emulate scroll wheels when pressed+dragging (EmulateWheel option in Xorg config). This is kindof convenient in all around X11, especiall with a trackpoint or trackball... The funny part is that this Xorg feature hides the middle click + dragging action from dillo -- it gets vertical and horizontal scroll events instead, while the single middle click that could paste the URL (let through by Xorg, because of EmulateWheelclickToo) is blocked from that by dillo, which prepares for movement of the mouse -- which is impossible to get in this config. So an unreachable feature of dillo (dragging) makes a different expected feature (middle-click pasting into window) unreachable. I envision making this configurable in dillo (deactivate dragging to activate pasting)... or recreate the EmulateWheelclickToo behaviour to use the isolated click for the differing action. But that being said -- this is a really minor issue compared to CSS support. Alrighty then, Thomas. PS: As a side note: dillo does not handle the horizontal scroll events:-( This is a more major issue than middle-click pasting, 'cause I got really used to 2D scrolling. I hope FTLK2 doesn't just swallow these.

Oct. 23, 2008

pasting URL

by onepoint＠starurchin.org

On Wed, Oct 22, 2008 at 10:46:52PM +0100, higuita wrote: > you can already paste a URL to the X icon in the urlbar and it > will load... So it does! That's neat! Regards, Jeremy Henty [not the OP, BTW]

Oct. 23, 2008