- Dillo-dev - lists.mailman3.com

website FAQ update

by corvid＠lavabit.com

I made http://www.dillo.org/FAQ.corvid.html The old FAQ is still at http://www.dillo.org/FAQ.html for now. So if you have comments/find mistakes/whatever....

May 15, 2008

tabs - what are they?

by corvid＠lavabit.com

Johannes wrote: > while thinking about tabs I realized that I'm not really sure > what they are. > > * On the one hand they could be implemented just as a GUI alternative > to multiple bowser windows. > > * On the other hand they could be something like temporary bookmarks > that are easily accessible e.g. via buttons below the panel. That > way one would only need one layout and view. > The only issue here would be that one would expect a separate > navigation stack for each "tab". This mail http://lists.auriga.wearlab.de/pipermail/dillo-dev/2008-January/003685.html says the latter. PS I read your subject line and thought, "They're eight spaces, of course." :)

May 15, 2008

tabs - what are they?

by Johannes.Hofmann＠gmx.de

Hi, while thinking about tabs I realized that I'm not really sure what they are. * On the one hand they could be implemented just as a GUI alternative to multiple bowser windows. * On the other hand they could be something like temporary bookmarks that are easily accessible e.g. via buttons below the panel. That way one would only need one layout and view. The only issue here would be that one would expect a separate navigation stack for each "tab". What do you think? Cheers, Johannes

May 15, 2008

Dillo freeze

by jcid＠dillo.org

On Mon, May 12, 2008 at 09:29:11PM +0200, Johannes Hofmann wrote: > On Mon, May 05, 2008 at 05:41:32PM -0400, Jorge Arellano Cid wrote: > > Hi there, > > > > Dillo freezes on this page: > > > > http://www.kernel.org/pub/linux/kernel/v2.6/linux-2.6.24.6.tar.bz2.sign > > > > Dillo bug or FLTK2? > > Other than I initially thought it is a Dillo bug. The IOwatch for > reading is not removed properly. As a result the select call in the > fltk event loop keeps failing with EBADF. > Attached is a fix for the problem, but I'm not sure whether it is > correct. Jorge, can you please have a look at it? I'm reviewing it, and in the process found 3 more bugs related to the same problem, but not necessarily to the patch... Stay tuned. -- Cheers Jorge.-

May 15, 2008

patch: some more window deletion

by corvid＠lavabit.com

I thought I was going to have to remove the viewport from the UI because both UI and layout would want to delete it, but it appears that TopGroup gets informed of the deletion and everything turns out fine.

May 15, 2008

Future of dillo2

by 4winter＠informatik.uni-hamburg.de

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Diego wrote: >> With my recent patch (dpid-cleanup-and-DPIPATH.dillo2) one can use the >> attached script to run dillo in-tree which is nice if you just want to >> try it out but should also be useful for development purposes (easily >> manage different branches without the need to install them somewhere). > > The patch seems good to me, but i have not look it very deep or test it. Some more testing turned out that it's not enough to add another path to search for dpis. With no dpidrc fill_services_list fails: [dpid]: dpid.c:467: fill_services_list: Fatal error [dpid]: - Can't find the directory for dpis. This seems very non-plugin-like to me. What's the point in being able to have dpis in ~/.dillo/... but to actually use them you have to add them to your ~/.dillo/dpidrc. I mean it's okay up to this point, but consider you have one custom dpi in ~/.dillo/... : * in order to use it you have to create a custom dpidrc too * judging by a quick look at dpid.c/register_service I'd say that you have to copy the content of the system wide dpidrc file too since your custom one takes preference over the system wide one * now there's a package installed containing one more dpi. It might be doable to update the system wide dpidrc via a post-install script within the package (though thats error prone, you have to remove your entries on package removal etc.) * but what about the custom dpircs? You cannot have the post-install script modify those as well (well, you could, but naaa) > The idea is that each dpi can have its own files in its directory > (preferences, data, cache files ... ). I could go with preferences and data files, but caching feels like a no-no to me since on most installations the user is unable to write into those directories. 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 iD8DBQFIKgZ9oPmwNWhsaZYRAjL+AJ9SjhWBUNh5WcO51GkJLaqtDyjFEwCeMQkz jkTcvl8xIbP3EedZP8T72fc= =xAX5 -----END PGP SIGNATURE-----

May 13, 2008

Future of dillo2

by 4winter＠informatik.uni-hamburg.de

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Andreas Kemnade wrote: > Hmm, can you really fully test dillo in its build directory? > You won't have the dpi stuff and loose functionality. Perhaps that should be > noted too. With my recent patch (dpid-cleanup-and-DPIPATH.dillo2) one can use the attached script to run dillo in-tree which is nice if you just want to try it out but should also be useful for development purposes (easily manage different branches without the need to install them somewhere). One catch is the way dillo expects the dpis to be in their own directory the script solves that by creating a temporary directory, but this seems like a bad hack. What is the big idea behind this anyway? * dpis cant usually write to "their" own directory anyway * afaik this isn't a namespace thing either (since the directory is named like the executable without the .dpi extension) So unless there's a reason that I cannot see I would suggest to get rid of the X/X.dpi thing all together and stuff them all into one directory. 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 iD8DBQFIKdUToPmwNWhsaZYRAqtyAJwNUOKoOPhKxJGAKKRBl68kIQ5PCQCgpV3D VFbjZobjZbmkSbEFa8lyNAY= =ro4T -----END PGP SIGNATURE-----

May 13, 2008

patch: Re: [Dillo-dev] should zlib be required?

by jcid＠dillo.org

On Mon, May 12, 2008 at 10:24:19PM +0000, corvid wrote: > Joerg wrote: > > On Sat, May 10, 2008 at 12:17:37PM +0000, corvid wrote: > > > Should zlib be required? > > > > Given many sites allow compression for HTML pages, it is useful to have > > around anyway. I would just make it a requirement. > > Attached. > I'm not sure that the old_libs code is serving any purpose, > but, being an autoconf newbie, I left it alone. Autoconf is a complex dragon! old_libs was for not stuffing libz in LIBS when the test failed. That way it could continue without zlib and without png. Now, IMHO it looks orthogonal and clean to leave it there. Committed. (now let's pray for it to work everywhere... :-) -- Cheers Jorge.-

May 13, 2008

Future of dillo2

by jcid＠dillo.org

On Mon, May 12, 2008 at 10:30:14PM +0000, corvid wrote: > > On Mon, May 12, 2008 at 09:35:46PM +0000, corvid wrote: > > > > http://www.dillo.org/Plans.html > > > > > > Did you commit/make/whatever it? > > > I thought "What changed?" and looked at a cached version, > > > and didn't see any differences. > > > > Not much here AFAIR the form elements and the rewrite of tables > > have now a green ball (finished). > > > > The bulk of it is adding what "should" be added. :-) > > I see the problem now. > > You didn't update the alt tags :) Yes! Now they're updated, and the extra items added, some sorting made and it looks more like what we have. I believe if we can complete the high priority items and stabilize the code we'd have a release candidate in our hands. If Sebastian completes the floating objects in textblock, that'd be great for the release, because dillo2 would show much better rendering in DIV-based and TABLE-based pages Comments? -- Cheers Jorge.-

May 13, 2008

patch: Re: [Dillo-dev] should zlib be required?

by corvid＠lavabit.com

Joerg wrote: > On Sat, May 10, 2008 at 12:17:37PM +0000, corvid wrote: > > Should zlib be required? > > Given many sites allow compression for HTML pages, it is useful to have > around anyway. I would just make it a requirement. Attached. I'm not sure that the old_libs code is serving any purpose, but, being an autoconf newbie, I left it alone.

May 13, 2008

Future of dillo2

by jcid＠dillo.org

On Mon, May 12, 2008 at 11:40:41PM +0200, Diego . wrote: > 2008/5/12 Jorge Arellano Cid <jcid(a)dillo.org>: > > > On Mon, May 12, 2008 at 03:51:42PM -0400, Jorge Arellano Cid wrote: > > > [...] > > > I'll try to update some pages in the site as soon as I find > > > some time. > > > > Just updated: > > > > http://www.dillo.org/cvs.html > > > > with directions on how to get dillo running. Developers, > > packagers, and experienced user should be pointed here. > > > > Somebody please check the directions work. :-P > > > > I doubt about the "exit" command, but i have not test the directions. "exit" is used as "logout" to get out of root. > and: > > > > http://www.dillo.org/Plans.html > > > > There is a line not updated under the "Current Plans for dillo2" text: > "--updated February 2007" > > It can be confusing. Ooops. Fixed. -- Cheers Jorge.-

May 13, 2008

Future of dillo2

by jcid＠dillo.org

On Mon, May 12, 2008 at 03:51:42PM -0400, Jorge Arellano Cid wrote: > [...] > I'll try to update some pages in the site as soon as I find > some time. Just updated: http://www.dillo.org/cvs.html with directions on how to get dillo running. Developers, packagers, and experienced user should be pointed here. Somebody please check the directions work. :-P and: http://www.dillo.org/Plans.html with what's already done. Note that we have more features than what this page states. Some new features as: * multiple character sets (in both HTTP and META). * HTTP-1.1's chunked transfer support. * optional image loading. * experimental gzip decoder. * Implemented the file input control for forms. * Optimized scrolling. * 64-bit ready out of the box. should be added here. -- Cheers Jorge.-

May 12, 2008

Future of dillo2

by jcid＠dillo.org

On Mon, May 12, 2008 at 06:24:30PM +0100, Jeremy Henty wrote: > On Mon, May 12, 2008 at 12:20:45PM -0400, Jorge Arellano Cid wrote: > > On Mon, May 12, 2008 at 03:52:59PM +0000, corvid wrote: > > > The consequences of not informing the outside world about the progress > > > of dillo in any detail: http://wiki.linuxfromscratch.org/blfs/wiki/dillo2 > > > > Please feel free to write them to get that page back on track before > > somebody believes what was written. ;) > > I have a BLFS Trac account so I can probably do it myself. But what > should I change? At a glance: > > * We *do* support FTP and HTTPS. > * Change "can support cookies" to "supports cookies". Yes, but https is basic. If a page has multiple images, the dialog window may spam the user once per image. That's why https went disabled for normal users, and experienced guys/devs were left the chance to enable it. > * "Dillo always interprets web pages as if they had the ISO-8859-1 > encoding." True? It depends on what dillo they talk about. It looks like a dillo2 page, and in that case, it's FLTK2-based, it has nothing to do with GTK1, glib nor GTK2, just FLTK2. The code is now a mixture of C/C++ instead of pure C, etc... It's important to make a difference between dillo1 (C lang, GTK1/glib, __unmantained__, and dillo2 (active branch). If we talk about dillo2 then the statement: "Dillo always interprets web pages as if they had the ISO-8859-1 encoding." is absolutely false. > > * "useless for reading non-English web pages" Surely not true, though > maybe "useless for reading multibyte-encoded pages" is true. If we talk about dillo2, false again. Dillo2 uses utf-8 by default, and utf-8 is multibyte-encoded. On top of that we have all the work corvid put in the character decoding, so now dillo2 can decode a boat load of charsets into utf-8 using iconv. Non english pages render quite well provided you have the necessary fonts installed. There're a few exceptions as with those that render backwards or have collapsing glyphs. For instance, set: vw_fontname="DejaVu Sans" in ~/.dillo/dillorc2 and go to: http://www.gnu.org/home.fa.html Worth a thounsand words! then go to: http://www.gnu.org/ scroll to the bottom and you'll see plenty of non-english pages, that you can click into. If you go to: http://www.bds-bg.org/ you'll see windows-1251 charset handled properly by dillo. etc. > Anything else? Yes, we should compile and provide this information from our web site. Otherwise it will be extremely hard for a non-developer to get right. A good diff document may be got from reviewing the Changelog alone. Here we have the record of all the work we've put in dillo2; from time to time I find there things forgotten long ago! I'll try to update some pages in the site as soon as I find some time. That's the point: lack of time/manpower. All the help is appreciated. Another very important point to be stated is that the plugins and the dpi daemon need to be installed for most things to work. After the build this is done with: make install-strip. Bottom line: current dillo2 is for developers. After we make a release it will be user ready. HTH. -- Cheers Jorge.-

May 12, 2008

Dillo freeze

by Johannes.Hofmann＠gmx.de

On Mon, May 05, 2008 at 05:41:32PM -0400, Jorge Arellano Cid wrote: > Hi there, > > Dillo freezes on this page: > > http://www.kernel.org/pub/linux/kernel/v2.6/linux-2.6.24.6.tar.bz2.sign > > Dillo bug or FLTK2? Other than I initially thought it is a Dillo bug. The IOwatch for reading is not removed properly. As a result the select call in the fltk event loop keeps failing with EBADF. Attached is a fix for the problem, but I'm not sure whether it is correct. Jorge, can you please have a look at it? Cheers, Johannes

May 12, 2008

Future of dillo2

by onepoint＠starurchin.org

On Mon, May 12, 2008 at 08:08:38PM +0200, Andreas Kemnade wrote: > Hmm, can you really fully test dillo in its build directory? You > won't have the dpi stuff and loose functionality. Perhaps that > should be noted too. Good catch! We don't want BLFS telling users to run Dillo in ways we don't support. Personally I don't understand the author's fear of installing alpha software. Does Dillo's "make install" do anything other than install a few executables and config files? If not then what's the problem with installing it? Also, the instructions should specify " ./configure --prefix=/usr " as the BLFS convention is to install everything in /usr rather than /usr/local . Jeremy Henty

May 12, 2008

Future of dillo2

by akemnade＠informatik.uni-bremen.de

Hi, On Mon, 12 May 2008 18:24:30 +0100 Jeremy Henty <onepoint(a)starurchin.org> wrote: > On Mon, May 12, 2008 at 12:20:45PM -0400, Jorge Arellano Cid wrote: > > On Mon, May 12, 2008 at 03:52:59PM +0000, corvid wrote: > > > The consequences of not informing the outside world about the progress > > > of dillo in any detail: http://wiki.linuxfromscratch.org/blfs/wiki/dillo2 > > > > Please feel free to write them to get that page back on track before > > somebody believes what was written. ;) > > I have a BLFS Trac account so I can probably do it myself. But what > should I change? At a glance: > > * We *do* support FTP and HTTPS. > * Change "can support cookies" to "supports cookies". > > * "Dillo always interprets web pages as if they had the ISO-8859-1 > encoding." True? > > * "useless for reading non-English web pages" Surely not true, though > maybe "useless for reading multibyte-encoded pages" is true. > > Anything else? > Hmm, can you really fully test dillo in its build directory? You won't have the dpi stuff and loose functionality. Perhaps that should be noted too. Greetings Andreas Kemnade

May 12, 2008

Future of dillo2

by onepoint＠starurchin.org

On Mon, May 12, 2008 at 12:20:45PM -0400, Jorge Arellano Cid wrote: > On Mon, May 12, 2008 at 03:52:59PM +0000, corvid wrote: > > The consequences of not informing the outside world about the progress > > of dillo in any detail: http://wiki.linuxfromscratch.org/blfs/wiki/dillo2 > > Please feel free to write them to get that page back on track before > somebody believes what was written. ;) I have a BLFS Trac account so I can probably do it myself. But what should I change? At a glance: * We *do* support FTP and HTTPS. * Change "can support cookies" to "supports cookies". * "Dillo always interprets web pages as if they had the ISO-8859-1 encoding." True? * "useless for reading non-English web pages" Surely not true, though maybe "useless for reading multibyte-encoded pages" is true. Anything else? Regards, Jeremy Henty

May 12, 2008

Future of dillo2

by corvid＠lavabit.com

Jorge wrote: > On Mon, May 12, 2008 at 03:52:59PM +0000, corvid wrote: > > The consequences of not informing the outside world about the progress > > of dillo in any detail: http://wiki.linuxfromscratch.org/blfs/wiki/dillo2 > > Quite funny! :-) > > Got to have courage to write such nonsense! It just shows what impression dillo gives a person who is focused on other things and giving the README a skim. > Please feel free to write them to get that page back on track > before somebody believes what was written. ;) I will. > BTW, if you want to update the website, please do. I don't have access yet.

May 12, 2008

Future of dillo2

by jcid＠dillo.org

On Mon, May 12, 2008 at 03:52:59PM +0000, corvid wrote: > The consequences of not informing the outside world about the progress > of dillo in any detail: http://wiki.linuxfromscratch.org/blfs/wiki/dillo2 Quite funny! :-) Got to have courage to write such nonsense! Please feel free to write them to get that page back on track before somebody believes what was written. ;) More seriously, I'll update the CVS page with all the necessary intructions to build dillo2. Beware that dillo2 is aimed towards developers, not users, until its first release. That's the message we need to pass across. (I made the mistake with dillo1 and ended up with a perception even more strange than the page you cite). BTW, if you want to update the website, please do. Just let me write the 6 years page (after a frozen year, the project restarted at year 8). You can start before I write that page. -- Cheers Jorge.-

May 12, 2008

Future of dillo2

by corvid＠lavabit.com

The consequences of not informing the outside world about the progress of dillo in any detail: http://wiki.linuxfromscratch.org/blfs/wiki/dillo2 (dillo+blfs mentioned previously by Jeremy in http://lists.auriga.wearlab.de/pipermail/dillo-dev/2008-March/004009.html)

May 12, 2008

Daily builds and tests / 0install feeds / PIE

by 4winter＠informatik.uni-hamburg.de

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hey hey :) I setup two boxes (one x86 running Debian lenny, one x86_64 running Ubuntu Hardy) to build daily snapshots of dillo2 (and fltk2). Tarballs, logs and screenshots are located at https://winter.gotdns.org/~teythoon/build/dillo2/ [suggestions for more interesting screenshots are welcome, like some testcases...] I also created a 0install feed (see http://0install.net/) for those daily builds. If you are on a "not that outdated" linux distribution (think of the libraries) on either x86 or x86_64 you can start dillo by running $ 0launch https://winter.gotdns.org/~teythoon/build/0feeds/dillo-fltk.xml If there is some interest in this I will try to make the builds more general (linking against older/more general versions of the required libraries etc). 0install feeds require programs to be position independent. Not like - -fpic or -pie but they should work where ever they and their files lie within the virtual file system. This is mostly true for dillo. dpid is located via PATH but the location to the dpis is (somewhat) hardcoded into the dpirc. So I went into dpid.c/register_all and poked around. Attached is a patch that allows me to specify an additional location for dpis via the environment variable DPIPATH. I also reorganized the code somewhat to minimize code duplication and I removed usr_srv_num since it was unused. Please let me know what you think :) 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 iD8DBQFIJw6poPmwNWhsaZYRAj4hAJ4rmvSQPPnHiGFQMXaTAGw3wg4JOQCeM9M3 oT6YvhD/s1ncdMQqAMifxgA= =NxaA -----END PGP SIGNATURE-----

May 11, 2008

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

by jcid＠dillo.org

On Sat, May 10, 2008 at 09:11:27PM +0000, corvid wrote: > Jorge wrote: > > There's still a bug hiding, most probably inside the code > > setting the scroll position in dw2. When you have a page with > > repush (i.e. with charset specified in META element, e.g. > > www.dillo.org :), and you scroll the page, and then reload. It > > will be shown from the origin, even though a_UIcmd_set_scroll_xy > > is called from nav.c:321 with the right coords (enable the _MSG > > to check). > > Scrolled down a little and hit reload. > In Layout::adjustScrollPos(), scrollY of 60 became 0 because > canvasAscent == 5 and canvasDescent == 5. > Looking at the backtrace, > a_UIcmd_set_scroll_xy() from a_Nav_expect_done() from a_Web_dispatch_by_type() > would seem to be too early. AFAIR it was handled with an idle function in dw1 in dillo1. Now it's also handled with an idle function, and I seem to recall Sebastian told me something about anchors long ago, but I don't remember. It is tricky because you have to handle several cases, with some events. For instance: * When the page is still loading and you have an anchor to scroll to. Here you have to retry until the page loads enough to scroll to the anchor, and you should keep watching just in case an image placed before the anchor arrives later, and then adjust again. Now if the user decides to scroll with the mouse, before the position settles, then you have to abort the idle scrolling, and let him control it. All this seasoned with window resize, page changes and other events. :-) -- Cheers Jorge.-

May 11, 2008

Future of dillo2

by jcid＠dillo.org

On Sun, May 04, 2008 at 08:30:32PM +0200, Sebastian Geerken wrote: > Hi, Hi Sebastian! > just if you were courious: I'm still living and will continue working > on dillo. That's very good news. New developers are helping to push the project forward, and your presence is a very important asset for the team. > > On Sun, Apr 20, Jorge Arellano Cid wrote: > > [...] > > We could: > > > > * Attract new developers with a dillo2 release, > > leveraging the excellent documentation that we have. > > We have: > > - extensive documentations on the critical parts of dillo, expecially > on Dw, > - a prototype for CSS, and > - some more documents and prototypes on future ideas, including > scripting and graphical plugins. > > These should be copied on the web sever, and there should be a > prominent note on the web site linking to a new delevolers' section. > (Since attracting developers has high priority, this is what a visitor > should notice first.) Yes, I was planning on stating the quest for more developers in the splash page of the next release. BTW, it'd be great to also convert dillo2's plain text docs to doxygen format someday. > For those not listening to dillo-dev, it is important to know that > dillo is still an active project. I think we can sort the house and then make a release so the world knows. That's why having TABS and at least a CSS prototype matters. I don't want reviews to end with the typical line: "but no TABS and no CSS". ;-) If we have TABS (a thing I can spare some time to implement) and a CSS protoype published somewhere, the message will be more like: "it comes with: ..., antialiased fonsts, low-dependencies, UTF-8, TABS, and they already started with CSS." That sounds much better! > > * Planning for CSS, floating objects and TABS. > > (this are important features to a nice browsing experience) > > Just FYI: I've started a bit on floats. Jorge, I once told you that it > would take a week; actually it is a bit more complicated, nevertheless > it is not a big job like CSS. Speaking about floats, I was trying to make NBSP (non breaking spaces) work for pages like: http://fltk.org/newsgroups.php?gfltk.development+T (with a narrow window, the parenthesis are spared). but found that NBSP-handling wasn't coded in textblock yet: /* NOTE: Most code relies on that all values of nowrap are * equal for all words within one line. */ I planned to give it a try, but didn't know what was the model for handling it (a flag in the previous word, a flag in the previous and next word, an empty word, etc). Now that you're working in the textblock, can you give it a review? The attached patch is what the parser (html.cc) could look like to handle this. It uses WHITE_SPACE_NOWRAP in style, but I didn't know whether to use: addText(strdup(""), S_TOP(html)->style); or DW2TB(html->dw)->addSpace(S_TOP(html)->style); for NBSP whitespace. It should be simple to test with it. > > * Coordinate with FLTK2, DSL (Damn small linux) and other > > lightweight distros, Debian installer team, MINIX3, etc. > > Perhaps we can find there other "lightweight zealots". :-) Yes, and maybe someday people will start liking/using this lightweight browser. AMEN -- Cheers Jorge.-

May 11, 2008

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

by corvid＠lavabit.com

Jorge wrote: > There's still a bug hiding, most probably inside the code > setting the scroll position in dw2. When you have a page with > repush (i.e. with charset specified in META element, e.g. > www.dillo.org :), and you scroll the page, and then reload. It > will be shown from the origin, even though a_UIcmd_set_scroll_xy > is called from nav.c:321 with the right coords (enable the _MSG > to check). Scrolled down a little and hit reload. In Layout::adjustScrollPos(), scrollY of 60 became 0 because canvasAscent == 5 and canvasDescent == 5. Looking at the backtrace, a_UIcmd_set_scroll_xy() from a_Nav_expect_done() from a_Web_dispatch_by_type() would seem to be too early.

May 10, 2008

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

by jcid＠dillo.org

Hi, On Thu, May 08, 2008 at 09:25:59PM +0000, corvid wrote: > 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. Oh yes. I didn't want to dig deep into it, but after this there was no chance to hide. :) There was a semantic confussion with the E2EReload flag, which was split into URL_E2EQuery and URL_ReloadPage. Now we can tell the difference between an first-time E2EQuery and an E2E reload. Patch committed. There's still a bug hiding, most probably inside the code setting the scroll position in dw2. When you have a page with repush (i.e. with charset specified in META element, e.g. www.dillo.org :), and you scroll the page, and then reload. It will be shown from the origin, even though a_UIcmd_set_scroll_xy is called from nav.c:321 with the right coords (enable the _MSG to check). -- Cheers Jorge.-

May 10, 2008

[patch]: dillo2 configure.in update for latest GNU m4

by jcid＠dillo.org

On Sat, May 10, 2008 at 08:37:36AM +0100, Jeremy Henty wrote: > On Sat, May 10, 2008 at 07:45:27AM +0100, Jeremy Henty wrote: > > > ... "AC_SUBST(src doc)" complains that "src doc" is not a valid > > shell tag. ... Anyway, replacing "AC_SUBST(src doc)" with > > "AC_SUBST(src)" and "AC_SUBST(doc)" seems to fix things. Patch > > attached. > > On second thoughts, it appears that "AC_SUBST(src doc)" does nothing > because the *.in files don't contain "@src doc@", "@src@" or "@doc@" > anywhere. Maybe we should just delete it. Here's an alternative > patch that does that. Committed. -- Cheers Jorge.-

May 10, 2008

should zlib be required?

by joerg.sonnenberger＠web.de

On Sat, May 10, 2008 at 12:17:37PM +0000, corvid wrote: > Should zlib be required? Given many sites allow compression for HTML pages, it is useful to have around anyway. I would just make it a requirement. Joerg

May 10, 2008

should zlib be required?

by corvid＠lavabit.com

Should zlib be required? I notice that testing for zlib is currently just a part of deciding whether to enable png. I either need to make configure.in require it, or make zlib optional with some ifdefs. As for which, though, I don't have much of an opinion.

May 10, 2008

[patch]: dillo2 configure.in update for latest GNU m4

by onepoint＠starurchin.org

On Sat, May 10, 2008 at 07:45:27AM +0100, Jeremy Henty wrote: > ... "AC_SUBST(src doc)" complains that "src doc" is not a valid > shell tag. ... Anyway, replacing "AC_SUBST(src doc)" with > "AC_SUBST(src)" and "AC_SUBST(doc)" seems to fix things. Patch > attached. On second thoughts, it appears that "AC_SUBST(src doc)" does nothing because the *.in files don't contain "@src doc@", "@src@" or "@doc@" anywhere. Maybe we should just delete it. Here's an alternative patch that does that. Jeremy Henty

May 10, 2008

[patch]: dillo2 configure.in update for latest GNU m4

by onepoint＠starurchin.org

I just updated my auto* tools including GNU m4 and the dillo2 build broke because "AC_SUBST(src doc)" complains that "src doc" is not a valid shell tag. I guess that the macro used to split at spaces and iterate over the result whereas the new version no longer does so. Anyway, replacing "AC_SUBST(src doc)" with "AC_SUBST(src)" and "AC_SUBST(doc)" seems to fix things. Patch attached. Please comment. Jeremy Henty

May 10, 2008

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

CVS glitch

by jcid＠dillo.org

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. Normally I update more or less as soon as > I see Jorge's "committed" message and the patches appear, but this > time I saw his acknowledgement late yesterday afternoon and the > patches still weren't there by midnight. They turned up this morning > so it was not a permanent failure. > > Did this affect anyone else? There was a HW failure in the remote transaction ordering and scheduling device of the CVS system. (i.e. me! :) > I've never seen it before! Good to know it works most of the time! -- Cheers Jorge.-

May 7, 2008

CVS glitch

by onepoint＠starurchin.org

Just so you know, for some reason it took a long time for the latest patches to show up in CVS. Normally I update more or less as soon as I see Jorge's "committed" message and the patches appear, but this time I saw his acknowledgement late yesterday afternoon and the patches still weren't there by midnight. They turned up this morning so it was not a permanent failure. Did this affect anyone else? I've never seen it before! Jeremy Henty

May 7, 2008

Render to image

by kfa＠gmx.net

Hi @, First of all, thanks guys for doing this great project :) Wondered if I could ask for a few pointers before digging my way through the source to see if I could do the following: I'm thinking of using Dillo as a rendering engine inside of another project. All I need is having it render offscreen to an image object of some kind, producing a click map for the links would be an added advantage. I know how to do this with Mozilla but that is such a heavyweight beast. I don't need any network code as well, all input material (html and media files) woud be provided by the enclosing application. Does anyone have a hint on where to start looking, which files / classes / libs out of the dillo package I would need for this? Thanks, Klaus

May 6, 2008

[patch]: Refactor {FltkListResource, FltkOptionMenuResource}::sizeRequest()

by jcid＠dillo.org

On Tue, May 06, 2008 at 12:25:44PM +0100, Jeremy Henty wrote: > > > A simple refactoring: FltkListResource and FltkOptionMenuResource > > use the same code to determine the maximum string width. > > And here's the patch! D'oh! Committed. -- Cheers Jorge.-

May 6, 2008

[bugfix]: FltkSelectionResource leaks memory

by jcid＠dillo.org

On Tue, May 06, 2008 at 11:47:21AM +0100, Jeremy Henty wrote: > > corvid pointed this out to me. Patch attached. Committed. -- Cheers Jorge.-

May 6, 2008

Future of dillo2

by Johannes.Hofmann＠gmx.de

Hi Sebastian, On Sun, May 04, 2008 at 08:30:32PM +0200, Sebastian Geerken wrote: > Hi, > > just if you were courious: I'm still living and will continue working > on dillo. That's fantastic! > > On Sun, Apr 20, Jorge Arellano Cid wrote: > > [...] > > We could: > > > > * Attract new developers with a dillo2 release, > > leveraging the excellent documentation that we have. > > We have: > > - extensive documentations on the critical parts of dillo, expecially > on Dw, > - a prototype for CSS, and > - some more documents and prototypes on future ideas, including > scripting and graphical plugins. > > These should be copied on the web sever, and there should be a > prominent note on the web site linking to a new delevolers' section. > (Since attracting developers has high priority, this is what a visitor > should notice first.) > > For those not listening to dillo-dev, it is important to know that > dillo is still an active project. > > > * Planning for CSS, floating objects and TABS. > > (this are important features to a nice browsing experience) > > Just FYI: I've started a bit on floats. Jorge, I once told you that it > would take a week; actually it is a bit more complicated, nevertheless > it is not a big job like CSS. > > > * Coordinate with FLTK2, DSL (Damn small linux) and other > > lightweight distros, Debian installer team, MINIX3, etc. > > Perhaps we can find there other "lightweight zealots". :-) > > Sebastian > > P.S. I'll soon review the changes regarding redrawing and related. > When working on floats, I noticed that the redraw optimization after > resizes does not work fully correctly, there are situations where only > a explicit redraw (e.g. hide and show window) makes size changes > visible. I've seen this too but can't reproduce it consistently. Cheers, Johannes > > P.P.S. In the doxygen documentation, there are two lists, > html/bug.html and html/todo.html (generated from the \bug and \todo > tags), with open issues. Furthermore, some problems are described in > html/fltk-problems.html. > > > _______________________________________________ > Dillo-dev mailing list > Dillo-dev(a)dillo.org > http://lists.auriga.wearlab.de/cgi-bin/mailman/listinfo/dillo-dev

May 6, 2008

Future of dillo2

by sgeerken＠dillo.org

Hi, just if you were courious: I'm still living and will continue working on dillo. On Sun, Apr 20, Jorge Arellano Cid wrote: > [...] > We could: > > * Attract new developers with a dillo2 release, > leveraging the excellent documentation that we have. We have: - extensive documentations on the critical parts of dillo, expecially on Dw, - a prototype for CSS, and - some more documents and prototypes on future ideas, including scripting and graphical plugins. These should be copied on the web sever, and there should be a prominent note on the web site linking to a new delevolers' section. (Since attracting developers has high priority, this is what a visitor should notice first.) For those not listening to dillo-dev, it is important to know that dillo is still an active project. > * Planning for CSS, floating objects and TABS. > (this are important features to a nice browsing experience) Just FYI: I've started a bit on floats. Jorge, I once told you that it would take a week; actually it is a bit more complicated, nevertheless it is not a big job like CSS. > * Coordinate with FLTK2, DSL (Damn small linux) and other > lightweight distros, Debian installer team, MINIX3, etc. Perhaps we can find there other "lightweight zealots". :-) Sebastian P.S. I'll soon review the changes regarding redrawing and related. When working on floats, I noticed that the redraw optimization after resizes does not work fully correctly, there are situations where only a explicit redraw (e.g. hide and show window) makes size changes visible. P.P.S. In the doxygen documentation, there are two lists, html/bug.html and html/todo.html (generated from the \bug and \todo tags), with open issues. Furthermore, some problems are described in html/fltk-problems.html.

May 6, 2008

Test

by sgeerken＠dillo.org

Hi, this is just a test message. Sebastian

May 6, 2008

Dillo freeze

by Johannes.Hofmann＠gmx.de

Hi, On Mon, May 05, 2008 at 05:41:32PM -0400, Jorge Arellano Cid wrote: > Hi there, > > Dillo freezes on this page: > > http://www.kernel.org/pub/linux/kernel/v2.6/linux-2.6.24.6.tar.bz2.sign > > Dillo bug or FLTK2? Sounds a bit like http://fltk.org/newsgroups.php?s91+gfltk.bugs+v92+T0+Qfile_chooser but that's supposed to be fixed: http://fltk.org/newsgroups.php?s95+gfltk.bugs+v101+T0+Qfile_chooser Will check that. Cheers, Johannes

May 6, 2008

ListResources: version 3 / scrolling bug / resizing bug

by 4winter＠informatik.uni-hamburg.de

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Jorge Arellano Cid wrote: >> I haven't looked into whether the added functionality works (didn't know >> where to look), but it compiles cleanly and runs just fine. > > Please check the attached page. > You should be able to select any set from the list, and to confirm > it works by pressing the test button and looking at the generated URL. Yes, that works nicely. But I noticed that I couldn't scroll the list content neither with my mouse wheel emulation nor with a real scroll wheel mouse. Playing around with dillo trying to scroll the content horizontally (which btw does not work with mouse buttons 6 and 7) I noticed that when I look at a large page (think /.), scroll down all the way to the bottom of the page and then resize the window making it wider you can see the content being re-rendered and moving towards the top of the window "escaping" the viewport. This should be easy to fix if you know where to look (which I don't at the moment...). 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 iD8DBQFIIEFLoPmwNWhsaZYRAuh4AJ4tTH93CQQT40PLwGm9hxfOb6O1tgCeIyD1 sDi8Q/ZWL5PSoVQt1DNiR2Q= =YVGC -----END PGP SIGNATURE-----

May 6, 2008