May 2008 - Dillo-dev - lists.mailman3.com

does RGBA work for you?

by corvid＠lavabit.com

Johannes wrote: > On Fri, May 09, 2008 at 02:23:09PM +0000, corvid wrote: > > (I'd seen that there was talk on the fltk forums recently about RGBA and images > > ending up with the wrong dimensions, but that's not my problem.) > > > > If you uncomment RGBA in test/dw_images_scaled.cc, does dragging the window > > around cause part of it to turn pinkish? > > (I'll attach a diff, even though it's not exactly difficult.) > > No, seems to work fine here. The rgab stuff in test/image from the > fltk tarball also seems to work. > I have xft enabled in my fltk build. Perhaps that makes a > difference? I have xft back on (I'd grown too accustomed to antialiasing to do without it). I'd forgotten about fltk/test. Nothing seems to be misbehaving there. Hmm. Maybe digging through the test source will enlighten me. I can hope, anyway...

May 9, 2008

does RGBA work for you?

by Johannes.Hofmann＠gmx.de

On Fri, May 09, 2008 at 02:23:09PM +0000, corvid wrote: > (I'd seen that there was talk on the fltk forums recently about RGBA and images > ending up with the wrong dimensions, but that's not my problem.) > > If you uncomment RGBA in test/dw_images_scaled.cc, does dragging the window > around cause part of it to turn pinkish? > (I'll attach a diff, even though it's not exactly difficult.) No, seems to work fine here. The rgab stuff in test/image from the fltk tarball also seems to work. I have xft enabled in my fltk build. Perhaps that makes a difference? Cheers, Johannes

May 9, 2008

does RGBA work for you?

by corvid＠lavabit.com

(I'd seen that there was talk on the fltk forums recently about RGBA and images ending up with the wrong dimensions, but that's not my problem.) If you uncomment RGBA in test/dw_images_scaled.cc, does dragging the window around cause part of it to turn pinkish? (I'll attach a diff, even though it's not exactly difficult.) I had started by hacking up the dillo side to send RGBA, and it worked fairly well in general for images where every pixel is either completely opaque or completely transparent. Gradients didn't seem to work too well, and often these pink/red regions were showing up. So far it seems as if the value of green relative to the value of alpha is what can make it particularly lose its mind, but that's probably wrong. You know how first impressions are... I followed it down to fltk/src/x11/Image.cxx, by which point it had been transformed to ARGB which looked about right to me. Then there's lots of xrender and transforms and things I don't know anything about. I tried some setcolor and fillrect in FltkImgbuf::draw(), but it didn't seem to matter, so I guess that's not my problem... If the question is "Why bother?", dillo2/src/png.c currently just mixes in whatever the background color is the first time we see an image on the page.

May 9, 2008

[BUG]: anchors break when a page includes '<meta http-equiv="Content-Type" ...'

by corvid＠lavabit.com

Jorge wrote: > > [Johannes wrote:] > > I noticed one regression with the new anchor code: > > The "Continue Reading Older Posts" link on www.boingboing.net > > (e.g. http://www.boingboing.net/2008/04/03/#entry-44240) no longer > > scrolls to the right position. > > There were mixed issues here. > > I just committed a patch that modifies/fixes anchor handling > with repush (on the last weeks anchors didn't work for me even > inside the bookmarks page). > > Now it works, and more like Firefox. This is, now reloading a > page with fragment, scrolls to the fragment position, but going > back and forward remembers the last user-scrolled position; quite > comfortable. > > The page you mention, works OK when it's a local file, but not > from the network. This looks like a bug hiding in the offset > calculation or scroll adjustment while the page loads images. > > In brief: with the patch the situation improves a lot, and > there's still a bug to fix. Things seem flaky now with redirection. Here's what happens if I 1 type in dillo.org 2 type in minix3.org $ dillo dillo_dns_init: Here we go! (threaded) Disabling cookies. Nav_open_url: new url='about:blank' Nav_open_url: new url='http://dillo.org' Dns_server [0]: dillo.org is 0x8171648 Connecting to 134.102.206.165 Nav_open_url: new url='http://www.dillo.org/' Dns_server [0]: www.dillo.org is 0x8194228 Connecting to 134.102.206.165 META Content-Type changes charset to: iso-8859-1 Nav_open_url: new url='http://www.dillo.org/' META Content-Type gave charset as: iso-8859-1 a_Nav_expect_done: repush! Nav_open_url: new url='http://minix3.org' Dns_server [0]: minix3.org is 0x8147220 Connecting to 130.37.20.20 Nav_open_url: new url='http://www.minix3.org/' Dns_server [0]: www.minix3.org is 0x8198090 Connecting to 130.37.20.20 a_Nav_expect_done: e2ereload! META Content-Type changes charset to: iso-8859-1 Nav_open_url: new url='http://www.dillo.org/' META Content-Type gave charset as: iso-8859-1 a_Nav_expect_done: repush! I did not press Back.

May 8, 2008

[BUG]: anchors break when a page includes '<meta http-equiv="Content-Type" ...'

by jcid＠dillo.org

Hi, > [Johannes wrote:] > I noticed one regression with the new anchor code: > The "Continue Reading Older Posts" link on www.boingboing.net > (e.g. http://www.boingboing.net/2008/04/03/#entry-44240) no longer > scrolls to the right position. There were mixed issues here. I just committed a patch that modifies/fixes anchor handling with repush (on the last weeks anchors didn't work for me even inside the bookmarks page). Now it works, and more like Firefox. This is, now reloading a page with fragment, scrolls to the fragment position, but going back and forward remembers the last user-scrolled position; quite comfortable. The page you mention, works OK when it's a local file, but not from the network. This looks like a bug hiding in the offset calculation or scroll adjustment while the page loads images. In brief: with the patch the situation improves a lot, and there's still a bug to fix. -- Cheers Jorge.-

May 8, 2008

Render to image

by kfa＠gmx.net

Justus Winter wrote: > Klaus F. Abel wrote: >> I know how to do this with Mozilla but that is such a heavyweight beast. > I was looking for something like this once but was unable to find > anything useful. Could you give me some hints? Here is an article about how to create thumbnail views in a sidebar: http://weblogs.mozillazine.org/roc/archives/2005/05/rendering_web_p.html > > Cheers, Klaus

May 8, 2008

Render to image

by 4winter＠informatik.uni-hamburg.de

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Klaus F. Abel wrote: > All I need is having it render offscreen to an image object of some kind, +1 on this one. This could also be useful to track regressions by rendering test cases and comparing the resulting images. > I know how to do this with Mozilla but that is such a heavyweight beast. I was looking for something like this once but was unable to find anything useful. Could you give me some hints? Cheers, Justus - -- gpg key fingerprint: C82D 382A AB38 1A54 5290 19D6 A0F9 B035 686C 6996 -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.6 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iD8DBQFIIikJoPmwNWhsaZYRAvkwAKCm5k+jxFr9ztmTfJj9TFOWVCdIaACfQCZA ewJ/ynlt35sfqXLGVWVsWh8= =sI/R -----END PGP SIGNATURE-----

May 8, 2008

[patch] fix redraw issue on resize

by jcid＠dillo.org

On Wed, May 07, 2008 at 07:31:00PM +0200, Johannes Hofmann wrote: > Hi, > > attached patch fixes an obvious problem with the redraw optimization > that results in artifacts when resizing the window. > There may be other issues left. Committed. -- Cheers Jorge.-

May 7, 2008

[patch] fix redraw issue on resize

by Johannes.Hofmann＠gmx.de

Hi, attached patch fixes an obvious problem with the redraw optimization that results in artifacts when resizing the window. There may be other issues left. Cheers, Johannes

May 7, 2008

CVS glitch

by onepoint＠starurchin.org

On Wed, May 07, 2008 at 09:01:35AM -0400, Jorge Arellano Cid wrote: > On Wed, May 07, 2008 at 08:56:26AM +0100, Jeremy Henty wrote: > > > > Just so you know, for some reason it took a long time for the > > latest patches to show up in CVS. > > There was a HW failure in the remote transaction ordering and > scheduling device of the CVS system. (i.e. me! :) Ah, I hope you rebooted quickly! :) Jeremy Henty

May 7, 2008